Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

May 2021

 Research methodology 6

Contents  Background of the research 7

 Scope of the research 8

2. Summary of key messages 10

3. Overview of IDP software products 12  Understanding enterprise grade IDP solutions 13

 OCR vs. IDP 14

 Drivers of IDP Solution 15

 Types of IDP solution 16

 Partner ecosystem 17

4. IDP Product PEAK Matrix® characteristics 18  PEAK Matrix positions – summary 19

 For more information on this and other research PEAK Matrix framework 20 published by Everest Group, please contact us:  Everest Group PEAK Matrix for IDP 21

Anil Vijayan, Vice President  Characteristics of Leaders, Major Contenders, and Aspirants 24

Ashwin Gopakumar, Practice Director  Technology vendors’ capability summary dashboard 27 Senior Analyst Samikshya Meher, 5. IDP market – competitive landscape 32 Shiven Mittal, Senior Analyst Utkarsh Shahdeo, Senior Analyst

 Leaders 39 Contents – ABBYY 40 – AntWorks 42

– Automation Anywhere 44

– IBM 46

– Kofax 48

– WorkFusion 50

 Major Contenders 52

– BIS 53

– Celaton 55

– Datamatics 57

– EdgeVerve 59

– Evolution AI 61

– HCL Technologies 63

– Hypatos 65

– Hyperscience 67

– Indico 69

– Infrrd 71

– JIFFY.ai 73

– Parascript 77 Contents – Rossum 79 – Singularity Systems 81

– UST SmartOps 83

 Aspirants 85

– GuardX 86

– i3systems 88

– qBotica 90

– SortSpoke 92

– TAIGER 94

7. IDP product capability trends and predictions 96

8. Appendix 102 Glossary 103

 Research methodology

 Background of the research

 Focus of the research 01  Summary of key messages

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 5 Our research methodology is based on four pillars of strength to produce actionable and insightful research for the industry

01 02 03 04

Robust definitions Primary sources Diverse set of Fact-based research and frameworks of information market touchpoints Data-driven analysis Function-specific Annual RFIs, vendor Ongoing interactions with expert pyramids, Total Value briefings & buyer across key perspectives, Equation (TVE), interviews, and stakeholders, input from trend-analysis across PEAK Matrix®, and web-based surveys a mix of perspectives market adoption, market maturity and interests, supports contracting, and both data analysis and vendors thought leadership

Proprietary database on Intelligent Document Processing (IDP) capabilities of 27 technology vendors Repository of existing research in IDP Dedicated team for IDP research Executive-level relationships with buyers, service providers, technology providers, and industry associations

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 6 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Background of the research

Background of the research Everest Group defines Intelligent Document Processing (IDP) as any software product or solution that captures data from documents (e.g., email, text, PDF, and scanned documents), categorizes, and extracts relevant data for further processing using AI technologies such as computer vision, OCR, Natural Language Processing (NLP), and machine/deep learning. These solutions are typically non-invasive and can be integrated with internal applications, systems, and other automation platforms. IDP products find a wide variety of use cases from different business functions and verticals. Adoption of IDP solutions can not only help enterprises achieve cost savings, but also improve their workforce productivity and employee & customer experience. These products are also rapidly evolving in the sophistication of their capabilities, features, and functionalities. In this study, we assess IDP software products that leverage AI/cognitive capabilities and are available for independent licensing. They are offered either as platforms that allow enterprises to deploy out-of-the-box solutions using pre-built modules, or as custom solutions to buyers with the intent of classifying and extracting data from documents.

In this study, we analyze the IDP technology landscape across various dimensions:  Everest Group’s PEAK Matrix® evaluation, a comparative assessment of 27 leading IDP technology vendors  Overview of IDP software products  Competitive landscape of the IDP technology vendor market  Everest Group’s remarks on key strengths and limitations for each IDP technology vendor  IDP product capability trends and predictions

Scope of this report: Geography Products Technology vendors Global Intelligent Document Processing 27 leading IDP (IDP) technology vendors

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 7 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 The IDP supplier landscape consists of multiple players that play varying roles

NOT EXHAUSTIVE

Focus of this research IDP landscape

IDP Independent Software Vendors IT-BPS service providers (ISV)

Technology vendors that offer IDP solutions as a Service providers that provide IDP solutions in their stand-alone product/solution; typically available for services portfolio – may or may not be available as independent licensing stand-alone products/solutions

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 8 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Everest Group’s SOT research is based on multiple sources of proprietary information

Proprietary database of 27 IDP technology vendors Vendors assessed  The database tracks the vendors’ offering/capabilities for: – Document processing and software learning features – Product-related training and support services – Interoperability, monitoring, and improvement features – Availability and adoption of commercial model(s) – Deployment and hosting options – IT governance and security – Partnerships with service providers and other technology vendors Proprietary operational information database of technology vendors (updated annually)  The database tracks the following operational information for each vendor: – Revenue and number of FTEs – Portfolio coverage in terms of industry, – Number of clients geography, process areas, and buyer size – FTE split by different lines of business Demonstrations and interactions with technology vendors and other industry stakeholders  Detailed demos for a comprehensive product view and executive-level discussions with IDP vendors that cover: – Current state of the market – Opportunities and challenges – Vision and strategy – Emerging areas of investment – Annual performance and outlook Buyer reference interviews, ongoing buyer surveys, and interactions  Interviews with technology vendors’ reference clients and enterprise IDP buyers to get the buyer perspective around: – Drivers and objectives for adopting IDP – Apprehensions and challenges – Assessment of vendors’ performance – Emerging priorities / buying criteria – Outcomes achieved – Lessons learnt and best practices

The source of all content is Everest Group unless otherwise specified Confidentiality: Everest Group takes its confidentiality pledge very seriously. Any contract-specific information collected will only be presented back to the industry in an aggregated fashion

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 10 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Summary of key messages

 Everest Group classifies IDP technology vendors on the Everest Group Products Everest Group Intelligent Document Processing (IDP) Products PEAK Matrix® Assessment 2021 PEAK Matrix® into the three categories of Leaders, Major Contenders, and Aspirants: Leaders Major Contenders Aspirants Star Performers High – Leaders: ABBYY, AntWorks, Automation Anywhere, Kofax, IBM, and Leaders WorkFusion WorkFusion Major Contenders Hyperscience HCL Automation Anywhere – Major Contenders: BIS, Celaton, Datamatics, EdgeVerve, Evolution AI, HCL Rossum Technologies Kofax Infrrd ABBYY Technologies, Hypatos, Hyperscience, Indico, Infrrd, JIFFY.ai, Nividous, AntWorks IBM Datamatics Hypatos Parascript, Rossum, Singularity Systems, and UST SmartOps Parascript Singularity Systems – Aspirants: GuardX, i3systems, qBotica, SortSpoke, and TAIGER JIFFY.ai Nividous BIS  Datamatics, HCL Technologies, Hyperscience, Infrrd, and Rossum have Celaton Indico UST SmartOps

Market impact EdgeVerve demonstrated the strongest year-over-year movement on both market impact and SortSpoke GuardX vision & capability dimensions and emerged as “2021 IDP Market Star Evolution AI Performers” Aspirants qBotica i3systems (Measures impact created in the market)  ABBYY, Automation Anywhere, IBM, Kofax, and WorkFusion have the highest TAIGER Low license revenue from IDP products Low High Vision & capability (Measures ability to deliver products successfully)

Note: Star Performers are selected based on a relative comparison of vendors’ performance along both the market impact and vision & capability dimensions between our previous and current PEAK Matrix® assessment. Those vendors with the greatest year-over-year improvement are designated as Star Performers

 Understanding enterprise grade IDP solutions

 OCR vs. IDP

 Drivers of IDP solutions 03  Types of IDP solutions  Partner ecosystem

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 12 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Understanding enterprise grade IDP solutions IDP software solutions blend the power of AI technologies to efficiently process all types of documents and feed the output into downstream applications

An enterprise-grade IDP solution performs the following actions:  Pre-processing: Performs image pre-processing to increase the quality of the scanned document and uses OCR/computer vision technology to capture data  Classification: Indexes & classifies the documents into categories using text mining & ML/DL capabilities  Extraction: Extracts relevant data leveraging NLP and ML/DL capabilities for further processing  Post-processing: Validates the extracted data with the help of pre-defined taxonomies, data dictionary, and business validation rules

IDP solution

Pre-processing Classification Extraction Post-processing Incoming documents

Structured data Image Data capture Indexing & Extraction of Data validation pre-processing classification relevant data

Internal / external Auto-crop, noise Computer vision Text mining & machine / NLP and machine Powering business validation reduction, etc. and OCR deep learning /deep learning technologies rules etc.

Structured data fed to downstream Human-in-the-loop applications (ERP, CRM, SAP legacy for verification and systems, etc.) via APIs or RPA correction

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 13 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 OCR vs. IDP IDP solutions are capable of processing documents with greater accuracy and are more resilient to changes in document templates than traditional OCR

Conventional OCR/template-based solution IDP solution

It may use OCR to convert images of documents to a digital OCR converts images of documents into machine-encoded format, but extracts specific information using machine learning text and extracts specific fields based on templates and/or deep learning

It uses rule-based or template-based extraction. User needs The extraction does not depend on the template but content. User to train the system for each template type needs to do minimal (if any) training for minor template changes

Every converted document needs to be manually reviewed, Once the system is trained, Straight Through Processing (STP) unless the input documents are standard (in quality, positional can be enabled. The percentage of STP achieved can vary elements, etc.)

Cannot process unstructured documents such as contracts With the help of Natural Language Processing (NLP) capabilities, and emails the system can process complex unstructured documents and can also create summaries

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 14 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Drivers of IDP solutions Enterprises look to adopt IDP solutions to unlock insights from semi-structured/unstructured data to improve operational efficiency and strategic outcomes

1 Reduce the overall cost of processing huge volumes of data

2 Increase accuracy and speed

Drivers of IDP adoption

Potential synergy that can be achieved through integrating 3 IDP products with RPA and BPM solutions

4 Enterprises’ intent to improve strategic business outcomes beyond cost reduction (e.g., improved customer experience)

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 15 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Types of IDP solutions The IDP software market primarily comprises two types of solutions: package-based and platform-based solutions

Types of Package-based solutions solutions in IDP Platform-based solutions software market

Refers to IDP software solutions offered as package or closed Denotes the IDP software solutions offered as platforms, which AI-savvy solutions, where the vendors or implementation partners will enterprise users can use to build and deploy specific use cases predominantly oversee the customization, configuration, and themselves, or with support from the vendor or implementation partners deployment of the solution Pros: Pros:  Enterprises can build use cases on their own  Dedicated resources and solutioning experts can assist  Allows for experimentation with choosing the best fit models and greater enterprises in deploying the product across complex use cases degree of control  Consistency in terms of accuracy and performance Cons: Cons: Given the current state of maturity and scarcity of skilled resources, most Inability to build new use cases or tinker with existing ones enterprises end up having to use external support for new use cases, thereby diluting the promise of a platform solution

Currently, both models are viable in the market. Given the scarcity of skilled resources today, most platform solutions end up acting as package-based solutions, except in the case of mature enterprises with dedicated data science talent.

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 16 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Partner ecosystem The partner ecosystem consists of multiple players that support and leverage each other

NOT EXHAUSTIVE

Partner ecosystem

RPA vendors Specialist system integrators

IDP and RPA vendors partner with each other and Specialist system integrators offer technical and provide their clients a wholesome automation operation support to service providers, tech vendors, solution or buyers to implement IDP solutions

Source: Everest Group (2021)

 PEAK Matrix® positions – summary

 PEAK Matrix® framework

 Everest Group PEAK Matrix® for IDP 04  Characteristics of Leaders, Major Contenders, and Aspirants  Technology vendors’ capability summary dashboard

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 18 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 IDP Products PEAK Matrix® 2021 – summary

 The Products PEAK Matrix® is a proprietary framework used to assess the market impact and overall vision & capability of technology vendors  Based on Everest Group’s comprehensive evaluation framework, the Products PEAK Matrix ®, we segment 27 technology vendors into three categories:

Leaders Major Contenders

Aspirants Star Performers

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 19 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Everest Group PEAK Matrix® is a proprietary framework for assessment of market impact and vision & capability

Everest Group PEAK Matrix

High

Leaders

Major Contenders Market impact Market Aspirants Measures impact created impact inMeasures the market) (

Low Low High Vision & capability (Measures ability to deliver products successfully)

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 20 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 IDP Products PEAK Matrix® evaluation dimensions

Measures impact created in the market – captured through three subdimensions

Market adoption Leaders Number of clients, revenue base, and YoY growth Major Contenders

Portfolio mix Diversity of client base across industries, geographies, business functions,

and enterprise size class impact Market Aspirants Value delivered Value delivered to the client based on customer feedback and other measures Vision & capability

Measures ability to deliver products successfully. This is captured through five subdimensions

Vision and strategy Document processing capability Monitoring and improvement Implementation and support Commercial model Vision for the client and itself; future Software learning, extraction & Performance tracking, operational Hosting options, training, Progressiveness, flexibility, and client roadmap and strategy classification, unstructured document analytics, reporting, and integration maintenance, partnerships with adoption of available commercial processing, interoperability, and with third-party BI tools resellers / system integrators, and models security and compliance complementary technology vendors

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 21 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Everest Group confers the Star Performers title on vendors that demonstrate the most improvement over time on the PEAK Matrix®

Methodology Everest Group selects Star Performers based on the relative YoY improvement on the PEAK Matrix

Year 1

In order to assess advances on market impact, we evaluate each vendor’s performance across a number Year 0 of parameters including:  Yearly YoY revenue growth  # of new clients  Improvement in portfolio mix Market impact Market  Improvement in value delivered

Vision & capability

In order to assess advances on vision and capability, We identify the vendors whose improvement ranks in the top we evaluate each vendor’s performance across a number quartile and award the Star Performer rating to those of parameters including: vendors with:  Innovation  The maximum number of top-quartile performance  Improvements in product features and functionalities improvements across all of the above parameters  Expansion of product associated consulting, training, AND support, and maintenance capabilities  At least one area of top-quartile improvement performance  Technology/domain specific investments in both market success and capability advancement

The Star Performers title relates to YoY performance for a given vendor and does not reflect the overall market leadership position, which is identified as Leader, Major Contender, or Aspirant.

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 22 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Everest Group PEAK Matrix® Intelligent Document Processing (IDP) Products PEAK Matrix® Assessment 2021

Everest Group Intelligent Document Processing (IDP) Products PEAK Matrix® Assessment 2021

High Leaders Leaders WorkFusion Major Contenders Major Contenders Hyperscience HCL Automation Anywhere Aspirants Technologies Rossum Kofax Star Performers Infrrd ABBYY AntWorks IBM Datamatics Hypatos Parascript Singularity Systems JIFFY.ai Nividous BIS Celaton Indico Market impact Market UST SmartOps EdgeVerve SortSpoke GuardX Evolution AI Aspirants qBotica

(Measures impact created impact in(Measures the market) i3systems TAIGER Low Low High Vision & capability (Measures ability to deliver products successfully)

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 23 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 IDP Products PEAK Matrix® characteristics

Leaders: ABBYY, AntWorks, Automation Anywhere, IBM, Kofax, and WorkFusion  Leaders continue to create new opportunities for the IDP market by investing in R&D to enhance AI capabilities and pre-packaged out-of-the-box solutions. They are also focusing on localized markets and have expanded support for processing documents in Asian and Middle Eastern regional languages such as Chinese, Korean, and Arabic  Responding to the rising demand for a holistic intelligent automation platform, Leaders are integrating their IDP platform with complementary technologies such as RPA, process mining, and IVA by forging partnerships or developing in-house capabilities

Major Contenders: BIS, Celaton, Datamatics, EdgeVerve, Evolution AI, HCL Technologies, Hypatos, Hyperscience, Indico, Infrrd, JIFFY.ai, Nividous, Parascript, Rossum, Singularity Systems, and UST SmartOps  The majority of Major Contenders capitalized on the spike in demand for IDP solutions, created due to COVID-19, by providing pre-packaged out-of-the-box solutions. They focused on reducing the implementation time and total cost of ownership for enterprises by introducing a SaaS offering of their platform  Major Contenders are pushing toward increasing their revenue from indirect sales channels. As part of their GTM, they have been forging partnerships with service providers and SIs for reselling and implementing the product

Aspirants: GuardX, i3systems, qBotica, SortSpoke, and TAIGER  The majority of aspirants are focusing on a particular industry, providing vertical-specific solutions to cater to enterprise needs. They are developing more specialized point solutions to address specific use cases  Aspirants are laying greater focus on processing unstructured documents and developing NLP capabilities to differentiate themselves from the leading and established vendors in the market

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 24 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Everest Group has identified Datamatics, HCL Technologies, Hyperscience, Infrrd, and Rossum as the 2021 Star Performers (page 1 of 2)

IDP vendors Star Distinguishing features Distinguishing features of capability Change in PEAK Matrix® Performers of market success in 2021 advancements in 2021 positioning for IDP product

 TruCap+ showed strong YoY revenue growth in  Developed an out-of-the-box analytics Strengthened its Major 2020, growing at a higher-than-average market rate dashboard to track software performance such Contenders positioning  Improved its overall client satisfaction score as accuracy and time for data extraction  Also experienced significant client growth,  Enhanced customer support by constituting an increasing its share of clients in Continental Europe online support forum for users, which is also reflected in its client satisfaction score

 Exacto achieved an above-market growth in  Expanded data extraction capabilities to also Strengthened its Major revenue in 2020, thus significantly improving its detect and extract barcodes and QR codes Contenders positioning market share  Developed a microservices-based architecture  More than doubled its number of clients, with individual components for digitization, expanding its footprint in emerging geographies extraction, classification, and configuration; of APAC, MEA, and LATAM enabled support for containerized deployment

 Achieved nearly 300% YoY revenue growth in 2020  Enhanced NLP capabilities for processing free- Strengthened its Major  More than doubled its client base, strengthening its flowing text in unstructured documents Contenders positioning foothold in the government and public sector  Improved document classification capabilities  Diversified the platform’s deployments in business by introducing the ability to group different areas such as F&A and healthcare industry-specific pages in a document  Introduced SLA-based work queue prioritization of documents Source: Everest Group (2021)

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 25 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Everest Group has identified Datamatics, HCL Technologies, Hyperscience, Infrrd, and Rossum as the 2021 Star Performers (page 2 of 2)

IDP vendors Star Distinguishing features Distinguishing features of capability Change in PEAK Matrix® Performers of market success in 2021 advancements in 2021 positioning for IDP product

 Achieved above-market average YoY growth in  Improved the platform’s UI to give users Strengthened its Major terms of revenue and clients in 2020 greater control for adding new document types Contenders positioning  Diversified its portfolio of clients by acquiring new and training the ML model clients in small and SMB enterprise segments  Introduced page and document-level classification models for classifying a mixed documents and applying the appropriate extraction model

 Realized more than 200% YoY revenue growth  Enhanced the AI engine to allow customization Strengthened its Major in 2020, thereby significantly increasing its and training for extraction of data Contenders positioning market share  Improved user interface to incorporate human-  More than doubled its client base, expanding to in-the-loop feedback for training model high-tech & telecom, travel and logistics, and  Got ISO-27001 and HIPAA compliance to professional services sectors adhere to security standards

Source: Everest Group (2021)

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 26 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Summary dashboard | Market impact and vision & capability assessment of vendors for IDP 2021 Leaders

Measure of capability: High Low

Market impact Vision & capability Document Vision and processing Monitoring and Implementation Commercial Technology vendor Market adoption Portfolio mix Value delivered Overall strategy capability improvement and support model Overall

ABBYY

AntWorks

Automation Anywhere

IBM

Kofax

WorkFusion

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 27 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Summary dashboard | Market impact and vision & capability assessment of vendors for IDP 2021 Major Contenders (page 1 of 2)

Measure of capability: High Low

BIS

Celaton

Datamatics

EdgeVerve

Evolution AI

HCL Technologies

Hypatos

Hyperscience

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 28 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Summary dashboard | Market impact and vision & capability assessment of vendors for IDP 2021 Major Contenders (page 2 of 2)

Measure of capability: High Low

Indico

Infrrd

JIFFY.ai

Nividous

Parascript

Rossum

Singularity Systems

UST SmartOps

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 29 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Summary dashboard | Market impact and vision & capability assessment of vendors for IDP 2021 Aspirants

Measure of capability: High Low

GuardX i3systems qBotica

SortSpoke

TAIGER

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 30 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Summary dashboard | Market impact and vision & capability assessment of vendors for IDP 2021 Star Performers

Measure of capability: High Low

Datamatics

HCL Technologies

Hyperscience

Infrrd

Rossum

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 32 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 IDP competitive landscape – summary

ABBYY, Automation Anywhere, IBM, Kofax, and WorkFusion are the top vendors in terms of IDP license revenue. Datamatics, Hyperscience, and Rossum have witnessed the highest growth rate in the market

ABBYY, Automation Anywhere, IBM, Kofax, and Parascript are leaders in the number of clients in the IDP market. ABBYY, Automation Anywhere, HCL Technologies, Parascript, and Rossum have acquired the highest number of clients

WorkFusion has the highest market share in BFSI and CPG & retail verticals, with ABBYY and Hyperscience dominating in the government and public sector. Kofax and Automation Anywhere are among the leaders in healthcare and manufacturing sectors, respectively

Kofax has emerged as the leader in Continental Europe, followed by ABBYY. WorkFusion leads in the North American market. Emerging markets of LATAM, APAC, and Middle Eastern regions are dominated by Automation Anywhere and IBM

Among business processes, the highest market share comes from F&A processes, closely followed by banking industry-specific processes. WorkFusion and Automation Anywhere lead the market share in banking industry-specific and F&A use cases, respectively

Overall, the IDP market saw a growth rate of around 25-30% in 2020, with acquisition of clients across small, mid-sized, and large enterprises globally

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 33 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Industry diversity IBM and Kofax are among the leading players across major industries; ABBYY and Automation Anywhere are other vendors with high market share across industries Top technology vendors across major industries by revenue Top five technology vendors (arranged alphabetically)

Banking & capital markets Healthcare & pharma Insurance

Manufacturing High-tech & telecom Government and public sector

Note: Some assessments may exclude technology vendors’ inputs, and are therefore based on Everest Group estimates, which leverage our proprietary Transaction Intelligence (TI) database, technology vendors’ ongoing coverage, public disclosures, and interaction with buyers Source: Everest Group (2021)

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 34 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Business functions ABBYY and Kofax are the leading players across major business processes, followed by Automation Anywhere, IBM, and WorkFusion Top technology vendors across major processes by revenue Top five technology vendors (arranged alphabetically)

Finance & accounting Banking industry-specific Insurance industry-specific Healthcare industry-specific

Mailroom Procurement HR Contact center

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 35 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Major geographies Automation Anywhere, ABBYY, IBM, and Kofax have strong presence across major geographies. WorkFusion is among the leaders in North America and Middle East markets Top technology vendors across major geographies by revenue Top five technology vendors (arranged alphabetically) United Kingdom Continental Europe

North America

Asia Pacific Latin America Middle East & Africa

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 36 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Vendors are gradually shifting from a perpetual licensing model to subscription-based licensing; usage-based pricing too is becoming more prevalent, compared to fixed-capacity

Measure of pricing model offered by vendors1 76%-100% 51%-75% 26-50% 1-25% None

Periodic subscription

Subscription-based Fixed capacity-based Flexible usage-based Flexible usage-based Technology vendor Perpetual licensing licensing (per document) (per page) Per process-based

Leaders

Major Contenders

Aspirants

 IDP vendors, in general, have started shifting away from perpetual licensing to annual/monthly subscription-based licensing models. Flexible per document usage-based pricing is the most prevalent pricing model  Vendors are offering document-based pricing as compared to charging on the number of pages processed, in response to the downward pricing pressure from market. Document-based pricing lowers the total cost of processing, as it usually does not consider blank pages within a document. This has also led to the emergence of an enterprise-wide licensing model  As the market matures and more organizations look to scale up, we expect market adoption of output-based pricing to increase  Vendors are also offering a commitment to minimum Straight Through Processing (STP) rate as part of their SLA with enterprises

1 Based on licensing models offered by 27 vendors Source: Everest Group (2021)

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 37 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 Leaders offer solutions across all major processes, while Major Contenders and Aspirants are mainly focused on processes specific to F&A and BFSI sector

Everest Group’s view of the spread of respective vendor use cases across processes

Others1 11% 16% 19% 3% 1% Procurement 5% 1%  Restrictions and the shift to a touchless/digital economy have further HR 4% 2% amplified the demand for IDP solutions, particularly in BFSI and 7% Healthcare industry-specific 9% 26% healthcare sector. Industry-specific processes have been a key focus area for the majority of vendors Insurance industry-specific 13% 19%  Leaders and Major Contenders are also expanding to horizontal use 5% cases, with over 30% of their revenue being generated from document processing in horizontal business areas Finance and 21%  Aspirants are primarily targeting industry-specific use cases in the BFSI Accounting (F&A) 27% sector. They are providing specialized pre-packaged solutions to minimize implementation and deployment time 52%  Deployment of IDP for processing unstructured documents such as Banking & capital markets 32% legal contracts, leases, and summarizing long-form text is also industry-specific 24% witnessing a gradual rise across industries

Leaders Major Contenders Aspirants

1 Others include web-based, e-commerce, self-service, mailroom, document management, IT services, pharma industry-specific, and utilities industry-specific processes Note: Some assessments may exclude technology vendors’ inputs, and are therefore based on Everest Group estimates, which leverage our proprietary Transaction Intelligence (TI) database, technology vendors’ ongoing coverage, public disclosures, and interaction with buyers Source: Everest Group (2021)

 Leaders – ABBYY – AntWorks 06 – Automation Anywhere – IBM – Kofax – WorkFusion

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 39 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

ABBYY (page 1 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Market impact Vision & capability Document Vision and processing Monitoring and Implementation Commercial Market adoption Portfolio mix Value delivered Overall strategy capability improvement and support model Overall

Strengths Limitations

 ABBYY has a vision to empower organizations to access insights from their business documents  While the FlexiCapture product can extract data points from unstructured text documents through and processes through its technology solutions. Its portfolio of offerings includes proprietary OCR NLP, it presently lacks NLG capabilities to summarize text from long-form documents such as engine FineReader, FlexiCapture & Vantage for IDP, and ABBYY Timeline for process mining contracts and legal documents

 ABBYY FlexiCapture, the flagship IDP product of ABBYY, provides out-of-the-box pre-trained  ABBYY Timeline, its process mining solution, is leveraged for providing process insights to users. models for processing invoices, contracts, lease agreements, purchase orders, receipts, and IDs However, integrating the task mining / Desktop Process Mining (DPM) capabilities of the solution

 ABBYY, through its large global clientele, has experience of serving a large number of enterprises to FlexiCapture solution is part of the roadmap across a majority of key verticals including BFSI, government and public sector, healthcare, CPG &  Pre-built connectors for leading ERP providers such as SAP, Oracle, and Microsoft Dynamics are retail, high-tech & telecom, and manufacturing currently not available, and users need to write custom export scripts or integrate through a web

 It provides FlexiLayout studio to create layouts of the documents to be processed. Users can define services API. This could lead to a longer time to implement and set up the product for enterprises or modify fields to be extracted through the FlexiLayout interface. Further, business rules for  While Vantage 2.0 is developed on a cloud-native architecture and can be deployed in a validation can also be configured through custom business scripting containerized form, ABBYY FlexiCapture is not containerized

 In addition to core extraction capabilities, ABBYY FlexiCapture also has NLP competencies for capturing data from unstructured documents such as loans, leases, contracts, and agreements

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 40 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

ABBYY (page 2 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 ABBYY FlexiCapture supports over 200 languages for document processing and extraction. It can  ABBYY’s client portfolio for IDP is slightly skewed toward mid-size and larger enterprises with also process handwritten documents in around 130 languages. User interface is available in 13 revenue greater than US$1 billion, with relatively lower presence in the small and SMB enterprise languages to serve the needs of a global customer base segments

 It offers an out-of-the-box administration and monitoring console for users to monitor processes,  While ABBYY has deployed its IDP solution for a majority of key business processes such as F&A, employee productivity, and system workload. Users can also create custom reports based on their procurement, contact-center, and BFSI industry-specific, its capabilities are somewhat untested for requirements use cases in healthcare provider and utility vertical-specific business processes

 ABBYY has a wide network of partners consisting of resellers and implementation and training  A free community version of the IDP product is part of the roadmap for Vantage release and only partners, which can be leveraged by enterprises during the various phases of deployment. It also free trial licenses are offered for the FlexiCapture product, which could limit the user base and partners with a variety of vendors for complementary technologies such as RPA, cloud hosting, support community of the product

BPM, and analytics  Clients have indicated a need for improvement in the initial product training provided to users  It provides an online training and certification portal for developers, analysts, and business users during the implementation phase. They also highlighted that better training support can improve along with remote instructor-led training programs. It also has an active online support forum, which the ease of use of the software, enhancing employee experience can be leveraged for general queries around the product

 Referenced clients have indicated core data extraction capabilities in multiple languages, integration with complementary capabilities, and customer support as key areas of strength

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 41 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

AntWorks (page 1 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 AntWorks envisions to provide an AI-based multi-lingual IDP platform for capturing data from  While the BFSI vertical has been a key focus area for its IDP platform, it has relatively less structured, semi-structured, and unstructured documents. It also offers RPA and task mining / DPM experience of serving enterprises in other key industries such as high-tech & telecom, technologies to augment its IDP capabilities. Its IDP platform – Cognitive Machine Reading (CMR) – manufacturing, and CPG & retail

leverages image recognition techniques to extract data from documents  The platform’s ability to serve enterprises in certain business process areas such as mailrooms  It has leveraged fractal science-based proprietary algorithms for AI/ML and can process documents and vertical-specific use cases for healthcare provider, pharma, and utilities industries is still in Japanese, Thai, and Korean in addition to Latin languages. The platform can also detect stamps, somewhat untested

logos, check boxes, and signatures, which can be verified as well against a specimen sample  The majority of AntWorks’ clients are based out of North America, APAC, and MEA, and it has  It allows users to add and configure new fields and business rules for validation through the user relatively low presence in other geographies such as Continental Europe and LATAM

interface. Users can hover and click on the data point within the document to fill the missing  The platform lacks the ability to work on mobile devices. Developing an application for uploading extracted information and processing documents from mobile is part of the roadmap  It has made significant improvements in its analytics and reporting capabilities and provides insights  The CMR platform lacks processing capabilities for documents in Chinese, Arabic, and Cyrillic around accuracy rates, STP, and human workforce analytics. It also provides pre-built connectors languages, which could inhibit its adoption by some global enterprises with third-party BI tools such as Tableau and Power BI

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 42 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

AntWorks (page 2 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 It provides pre-packaged solutions for some of the common document processing use cases  It does not provide pre-built connectors for leading information systems such as SAP, Oracle, and including checks, invoices, bill of lading, financial statements, mortgages, and email classification Salesforce and has to be integrated via API calls. This could lead to longer implementation and

 The platform also offers NLP capabilities and can perform sentiment analysis for blogs, tweets, integration time for enterprises emails, and other text documents to generate insights  The ability to auto-redact sensitive information in documents is not available as part of the CMR

 AntWorks has rearchitected the product based on microservices architecture to enable partners to platform. This could restrict the usage for enterprises seeking to process highly sensitive or customize it based on customer requirements. It also supports containerized deployments for confidential documents flexibility  AntWorks does not offer its IDP platform in a SaaS model, and currently it is part of its roadmap.

 AntWorks has constituted an online training portal offering training and certification courses for Small and SMB segment enterprises, seeking lower TCO and quick implementation time, may find developers and business users. It also offers training via classroom mode in India, the US, Japan, this as a drawback and Australia  Referenced clients have highlighted a need to improve the platform user interface to make it more

 Referenced clients appreciate its AI/ML models for extraction and the learning abilities of the intuitive and easier-to-use. Revamping the UI is part of the roadmap for AntWorks and is expected platform. They have also lauded its customer support services and the speed of execution of the to be released in 2021 platform  Clients have expressed the need for better visibility and clearer communication of the product roadmap. Ability to integrate with complementary technologies such as RPA of other vendors has also been indicated as key limitation

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 43 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Automation Anywhere (page 1 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 Automation Anywhere has a vision to provide a one-stop-shop intelligent automation platform to  A majority of deployments for IQ Bot are focused on F&A, HR, BFSI, and healthcare-payer enable broader process automation through RPA, IDP, process discovery, and analytics industry-specific use cases. Its ability to serve enterprises in other key process areas such as capabilities. IQ Bot is its proprietary IDP platform that integrates with its in-house RPA and analytics procurement, contact center, mailroom, and healthcare provider industry-specific processes is still offerings somewhat untested

 It experienced a healthy YoY growth in terms of revenue and number of clients in 2020, and  While Automation Anywhere supports processing of images from mobile devices, it does not alert continues to maintain a strong presence across all major geographies users in case the image uploaded is below the threshold quality. Clients have expressed a

 Automation Anywhere has experience in serving enterprises across most of the major industries requirement for the functionality to notify users in case the uploaded images are of low-quality, including BFSI, manufacturing, high-tech & telecom, CPG & retail, healthcare, and public sector below threshold level   IQ Bot platform offers pre-trained out-of-the-box models for over 100 use cases to simplify the Clients have indicated a need to improve NLP capabilities and include functionalities to identify deployment process. Some of the key pre-built use cases include auto-extraction for invoices, and compare intent in documents across multiple languages. It presently does not natively provide purchase orders, insurance claims, Explanation of Benefits (EOB), and financial statements capabilities to compare documents based on intent for cases such as comparison of clause in standard contract vis-à-vis individual contract  It supports data extraction for nearly 190 languages and the platform UI is available in 10 different languages to serve a global customer base  IQ Bot provides key analytics insights and reporting around STP rates at a process-level. However, tracking of STP rates at a batch-level is presently part of the roadmap

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 44 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Automation Anywhere (page 2 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 IQ Bot has been architected as a web-based, cloud-native platform facilitating continuous integration  IQ Bot presently provides out-of-the-box integration only with Automation Anywhere’s in-house and deployment RPA capabilities. While it is a leading provider of RPA technology, this could pose a challenge for

 IQ Bot provides pre-defined dashboards for measuring STP rates and field-level accuracy. It also enterprises looking to integrate IQ Bot with other RPA platforms apart from Automation Anywhere integrates with Automation Anywhere’s analytics platform, Bot Insight, which allows users to  Clients expect the IQ Bot platform to have a lower processing time per page. They have indicated customize dashboards and generate insights around key business metrics the high processing time per page as a hindrance to scaling up the platform

 Automation Anywhere has a wide network of implementation partners and resellers, which can be  Ability to auto-redact or mask/blur sensitive information on documents is presently not available leveraged by enterprises for support during on-premise deployments and is part of the roadmap for the platform. This could be a concern for enterprises processing

 It also offers a free community edition of the IQ Bot platform, which is pre-trained for invoice documents containing highly confidential data processing. A free trial version is also available for prospective customers to assess the suitability of  Referenced clients have also suggested greater flexibility for end-users in selecting the ML model the platform for its usage for training documents and defining additional metrics for customized dashboarding and reporting

 Clients of IQ Bot have lauded the interoperability capabilities of the platform with complementary technologies such as RPA, product training and support provided by the vendor, and the ease of use

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 45 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

IBM (page 1 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 IBM envisions to provide a suite of intelligent automation technologies as part of its Cloud Pak  IBM’s IDP platform provides relatively few out-of-the-box pre-built use cases, as compared with offerings. IDP capabilities are available as part of IBM Cloud Pak for Automation peers, for processing documents. Including new pre-trained models for some of the common use

 It has a wide network of implementation and training partners globally across 190+ countries. It also cases, such as processing utility bills and tax forms, is part of the roadmap has a robust technology partner ecosystem to complement its technology offerings and enable  The platform presently cannot automatically generate training data batch for manual review or smooth integration alert users to retrain the data. Updating the extraction model based on human-in-the-loop data

 It supports data ingestion in a variety of document formats including TXT, CSV, HTML, RTF, XLS, verification and corrections is part of the roadmap as of October 2020 PDFs, and images. The platform can also process handwritten documents and detect objects such  The platform supports most of the common languages for document processing including English, as stamps, logos, signatures, and embedded photos in documents Latin languages, Arabic, Chinese, and Japanese. However, it presently cannot process

 The platform provides a mix of in-house and third-party OCR engines to enable users to select the documents in some regional languages such as Korean and Hindi most suitable OCR engine based on document/language processing requirements  It allows users to add and configure business validation rules from the platform; however, the

 In addition to data extraction and document classification, it also delivers AI capabilities through ability to add/configure business validation rules using external data such as a lookup from an integration with IBM Watson. NLP, sentiment analysis, and summarization of long-form text in external database or system can only be done by admin users documents are offered through IBM Watson libraries

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 46 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

IBM (page 2 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 IBM Business Automation Insights module allows users to capture and track events and offers  The analytics and reporting module does not provide pre-defined benchmark metrics for visualization and insights. It provides insights into manual worker performance by type of document comparison of performances. Although, users can configure their own metrics for benchmarking, or task, STP rates, as well as customized reporting capabilities providing it out-of-the-box can help in identifying best practices in the industry

 The platform provides the ability to auto-redact sensitive information on documents, customized  A free community version of the platform is not available, which could inhibit its adoption among role-based access to the system, and also encryption of data in transition and at rest smaller enterprises who would like to test the platform first before deploying it in production for

 It also logs key user actions such as login, task steps, and the documents edited by users. These larger use cases activities are time-stamped and can be used for auditing purposes  IBM’s commercial model is not tied to outcome-based pricing and does not commit to providing a

 It offers flexibility to be deployed on-premise and on public/private cloud. A SaaS version of the minimum level of accuracy in SLAs, which some other vendors have started offering for their IDP product is available, which is hosted on IBM cloud solution

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 47 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Kofax (page 1 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 Kofax provides a broad-based intelligent automation solution encompassing IDP, RPA, process  The majority of Kofax clients are based out of North America, Europe, and APAC, and it has orchestration, process discovery, and analytics. It envisions to assist enterprises in their digital served relatively fewer enterprises in emerging geographies such as MEA vis-à-vis other transformation journey through its technology offerings geographies

 It brings a rich experience of serving a large number of enterprise clients globally through its IDP  While its IDP platform has been deployed across a majority of key horizontal and vertical-specific solution across a majority of key verticals including BFSI, government & public sector, healthcare & business processes, the platform’s capability to serve enterprises in some of the lesser prevalent pharma, and manufacturing processes such as healthcare-provider industry-specific use cases is still somewhat untested

 Its IDP platform, Kofax TotalAgility, has been deployed across various horizontal processes such as  The out-of-the-box use cases are primarily focused on invoice and claims processing, and it F&A, HR, procurement, and contact center, and for vertical-specific use cases for BFSI, healthcare presently does not offer pre-built use cases for processing some of the unstructured documents payer, and pharma industries such as contracts, legal agreements, and mortgage documents

 TotalAgility offers pre-built out-of-the-box solutions for invoice and claims processing. The platform  The platform has the ability to detect and capture signatures; however, the ability to detect and can be pre-trained for professional services use cases by Kofax or through its large network of extract some other object items such as logos and stamps is presently part of the roadmap certified partners

 In addition to the desktop-based platform, Kofax also offers an application for processing documents on mobile devices. The app can alert users if a low-quality image is captured and comes with pre- built solutions for ID verification, check deposit, and bank cards processing

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 48 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Kofax (page 2 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 Kofax can process and extract data from documents in more than 200 languages including Arabic  The commercial model of Kofax’s IDP platform is based on usage-based pricing on the number of and Asian languages such as Chinese, Japanese, and Korean. The platform’s UI is available in pages/documents processed. It currently does not tie its pricing to outcomes achieved such as seven different languages to meet the needs of a global clientele accuracy levels or STP rates

 In addition to data extraction from documents, the platform also has NLP capabilities and can  Referenced clients have also indicated that there is scope to make the licensing cost and perform sentiment analysis, entity extraction, and summarization of long-form text and documents commercial model of the platform more progressive and flexible

 Kofax offers a built-in analytics and reporting capability to provide insights around field accuracy,  Clients also expect Kofax to provide better training support at the time of purchase of the product. confidence levels for classification and extraction, human workforce efficiency, and other cost and Business users have indicated that it requires training and in-depth knowledge of the platform to operational efficiency metrics unlock the maximum value out of it

 Kofax has a wide network of resellers and technology partners, implementation, and training  Enhancements in debugging capabilities and ease of integration with complementary technology partners products of other vendors are the two key areas where clients would like Kofax to focus

 Clients have lauded its customer support, timely responsiveness, and incorporation of customer feedback into product roadmap

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 49 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

WorkFusion (page 1 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 WorkFusion has a vision to provide a unified automation platform through its Intelligent Automation  WorkFusion experienced below market average YoY growth in terms of IDP revenue and clients in Cloud (IAC) offering that combines RPA, IDP, analytics, AI/ML, and BPM capabilities to automate 2020. It also has a low presence outside North America

processes  Its client portfolio for IDP is skewed toward the large buyer segment, thereby making it relatively  It is further investing in developing pre-built industry-specific use cases for IDP. It also plans to less experienced in serving mid-size and small enterprises

empower business users to leverage the platform in a self-service manner by expanding its  While WorkFusion has a strong presence in the BFSI sector and also has client base in healthcare educational content and enablement program and retail industries, its ability to serve enterprises in some other key verticals such as  WorkFusion provides around 200 pre-built solutions for use cases such as negative news search, manufacturing, high-tech & telecom, and government & public sector is somewhat untested

identity verification, and invoice processing. Customers can access these solutions through the  WorkFusion’s web-based modules such as Control Tower, WorkSpace, and analytics can be WorkFusion Use Case Navigator online portal accessed via mobile devices through mobile-web. It does not offer a dedicated mobile application  WorkFusion’s Automation Studio enables users to define and edit automation steps. Users can also for capturing and processing documents

define and add fields through the manual task editor and configure business rules validation through  It presently supports processing of handwritten documents only in the English language. This UI in Control Tower module of the platform could be a deterrent for enterprises looking to deploy IDP solutions for processing handwritten  It offers NLP capabilities including sentiment analysis and classification capabilities for online news documents in other languages articles

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 50 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

WorkFusion (page 2 of 2) Everest Group assessment – Leader

Measure of capability: High Low

Strengths Limitations

 It supports document processing and data extraction in more than 30 different languages including  WorkFusion presently has very few technology partners for complementary technologies such as Asian and Middle Eastern regional languages such as Chinese, Japanese, Korean, and Arabic. The RPA, BPM, and IVA. While it has in-house RPA and BPM capabilities, lack of pre-built connectors UI of the platform is also available in five different languages to third-party solutions may lead to longer implementation and integration time for enterprises

 WorkFusion offers out-of-the-box customizable analytics and reporting capabilities. It delivers  Clients have highlighted a need to reduce the implementation time of the platform. They also insights around field-level accuracy rate and confidence levels along with batch-level STP rates and expect better management of infrastructure requirements of enterprises where the software is human workforce analytics. It also includes pre-built connectors for third-party BI platforms such as deployed on-premise

Tableau and ELK  Clients have mentioned the pricing of the product as slightly on the higher side and more suitable  The IAC platform is also available as a SaaS offering for enterprises seeking lower TCO. It supports for larger enterprise-wide deployments. It may be perceived as expensive for enterprises looking a distributed deployment model based on microservices architecture, allowing customers to deploy for small-scale deployments

on-premise and on public/private cloud  Business users perceived the platform to be slightly complex to use due to the vastness of product  It has constituted an online training academy to offer training programs and certifications to users. capabilities. They felt that proper training is required to set up the product and make it operational

The online portal also has separate modules catering to the needs of business users and  Referenced clients have highlighted dedicated customer support, especially for geographies developers outside North America, as one of the key limitations for WorkFusion  Clients appreciate WorkFusion’s broader automation platform capabilities, customer-centric approach to product development, and the online training academy

 Major Contenders – BIS – Hypatos – Parascript – Celaton – Hyperscience – Rossum 05 – Datamatics – Indico – Singularity Systems – EdgeVerve – Infrrd – UST SmartOps – Evolution AI – JIFFY.ai – HCL Technologies – Nividous

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 52 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

BIS (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 BIS’ IDP platform, Grooper, captures, classifies, and extracts data from unstructured documents in  BIS currently serves clients based out of North America; its experience in serving clients in other various input formats including PDF, TXT, JPG, XLS, and CSV files geographies such as Continental Europe, the UK, and APAC is untested

 It can process different data types including free-flowing text, handwritten text, bar codes, logos,  Its clientele is skewed toward government & public services, energy & utilities, professional stamps, signatures, and checkboxes services, and high-tech & telecom industries. Its experience of serving clients in other key verticals

 Grooper provides flexibility to enterprise users to add, modify, and edit fields to be extracted. Users such as insurance, healthcare & pharma, and travel & logistics is relatively limited can also add new document types for extraction  BIS relies heavily on a direct sales channel and has scope for further expanding its partner

 The platform indicates confidence levels of the data extracted from documents. Further, enterprise network to serve its clients better users can configure and set confidence threshold for straight through processing of documents  BIS provides few pre-trained models for invoice processing; however, it does not offer pre-

 Grooper allows users to configure and add business rules for validation of extracted fields. The packaged solutions for common use cases such as KYC documents, employee onboarding, and software can also look into external database for validations claims processing   In addition to data extraction and document classification capabilities, Grooper also provides NLP While the solution is capable of processing documents in multiple Latin-based, Greek, and Cyrillic capabilities for extracting data from free-flowing text with the help of regular expressions languages, the ability to process documents in Arabic and Asian languages such as Chinese and Japanese is very limited. Moreover, currently the user interface is available only in English

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 53 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

BIS (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Grooper comes with built-in integration with multiple OCR including Transym 4, Transym 5, Google  The software lacks advanced capabilities such as text summarization and the ability to compare Tesseract, Azure OCR, ABBYY, and Prime OCR engines for digitization. It also performs image different free-flowing text within unstructured documents

pre-processing to improve the accuracy of the OCR it integrates with  The solution lacks out-of-the-box integration with leading BPM solutions as well as third-party BI  BIS offers training through both classroom programs and online training portal. The portal provides platforms, while the integration with RPA tools is in the roadmap, limiting its current ability to self-paced instructor-led training modules and separate certification courses for business users and seamlessly integrate with existing systems

developers  Presently, BIS does not provide native mobile application for its Grooper IDP platform. This limits  BIS provides flexible deployment options as it can be deployed on desktop/laptop, cloud, and on- the utility of the platform for enterprises

premise. It is also delivered as a SaaS offering  Grooper provides reports for basic performance metrics such as batch performance, processing  Ability to detect and redact PII, flag documents containing PII, and restrict access to documents and time, and tasks processed. However, it lacks visual dashboards for a unified view of analytics batches based on Access Control Lists (ACLs) strengthen its security features around worker performance and benchmark metrics

 Clients appreciate the scalability of the product, its ability to process complex documents, and the  BIS currently offers usage-based and process-based pricing and does not yet offer progressive overall product support offered by BIS commercial models such as outcome-based pricing

 Clients expect a more detailed product documentation, better quality assurance of new releases, and faster resolution of bug fixes in the product

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 54 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Celaton (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Celaton aims to provide IDP capabilities through its cloud-based inSTREAM™ platform, with major  Celaton’s clientele is based out of the UK, Continental Europe, and North America. Its experience focus on Accounts Payable (AP) automation. It leverages AI/ML, OCR, ICR, NLP, and cognitive of serving enterprises in other geographies such as LATAM, MEA, and APAC is relatively limited

computing to deliver IDP capabilities to its clients  It is relatively less experienced in serving clients in major industries including BFSI and healthcare  The inSTREAM™ platform allows document ingestion in a variety of formats (such as TXT, CSV, & pharma

PDF, JPG, PNG, and XLS) and extracts information from various data types including free-flowing  Celaton primarily caters to business processes in F&A, HR, and contact center and has limited text, bar codes, logos, and signatures experience in procurement and other industry-specific processes  It comes with pre-built models for use cases including invoice processing, sales order processing,  It is currently capable of extracting documents in Latin-based languages and lacks the ability to freight bill and audit for logistics business, and delay repayment claims extract documents in Middle Eastern and Asian languages such as Arabic, Hebrew, Chinese,  inSTREAM can be trained using past data as well as supervised learning, wherein it highlights Japanese, Korean, and Cyrillic languages

errors and exceptions for operators to make corrections through its point-and-click interface  The inSTREAM platform lacks the ability to extract data from complex data formats such as  The platform performs image pre-processing to enhance the quality of documents. It can also alert nested tables and free-flowing handwritten documents

users in case the image quality is below minimum acceptable threshold  Currently, it does not facilitate enterprise users to define, add, or modify fields to be extracted.  Its NLP capabilities allows the platform to perform sentiment analysis and search through a Business users also cannot add, configure, or manage validation rules in the platform. This is repository of scanned documents managed by Celaton resources at the back-end

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 55 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Celaton (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Using NLP, inSTREAM can generate suggested responses such as a template that needs to be  While it can integrate with third-party applications via APIs, SOAP and REST web services, and sent to the supplier for incorrect or non-compliant invoices based on the issue identified through CSV and XML file transfer, it lacks pre-built connectors with leading enterprise applications (such document extraction and validation as SAP, Microsoft, and Oracle), RPA products, and BPM tools

 The reporting capability includes metrics on batch-level STP rates, process-level SLAs, field-level  The analytics dashboard is in the roadmap and the performance data is currently available through accuracy, as well as manual worker performance reports created for customers through the inSTREAM Web Portal

 It allows role-based access control to the system with a provision to add unlimited roles. It also  The platform is currently offered only as a SaaS offering, limiting its flexibility for users looking for allows auditing and time-stamping of every possible action to ensure better governance and on-premise or private cloud-based deployments

compliance  Celaton only offers usage-based pricing model; hence, the solution is not suitable for clients  The platform is built on microservices architecture and supports containerization that allows easy looking for progressive pricing constructs such as outcome-based models

scaling of the platform  Clients expect greater visibility into the product roadmap and a higher sense of ownership for  Clients appreciate Celaton’s service support and operational teams’ responsiveness to client needs. implementation and delivery from Celaton. Some clients also expect improvement in the accuracy They also appreciate the master data validation capabilities of the platform of the solution

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 56 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Datamatics (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 TruCap+ is a component of Datamatics’ Intelligent Automation platform that combines its RPA  Datamatics’ portfolio primarily includes clients from BFSI, healthcare & pharmaceuticals, travel & platform TruBot, IDP solution TruCap+, and AI platform TruAI to deliver an overarching automation logistics, and manufacturing sectors; hence, its ability to serve clients in other sectors such as solution CPG & retail, public sector, and high-tech is untested

 TruCap+ comes with pre-trained models for various document types, which can be used out-of-the-  Clients have deployed TruCap+ largely for F&A and banking and healthcare payer industry- box. These include invoices, claims, letter of credit, check processing, salary slips, and vehicle specific processes. Therefore, the platform’s capabilities in areas such as HR, contact center, and registration, among others mailroom are relatively unproven

 It has an in-house OCR but can also integrate with third-party OCR engines. It uses a multiple OCR  While it has experience in serving clients in North America and APAC, its experience in approach to improve accuracy, wherein multiple OCR engines extract an information element, and Continental Europe, LATAM, and MEA is relatively low. Also, it has limited experience serving upon comparison, the most accurate one is passed through for processing small enterprise and SMB clients (revenue < US$1 billion)   The platform uses a combination of NLP and neural networks for extraction, correction, and The platform does not have the ability to identify/classify documents or pages within the same validation of data from documents, including semi-structured and large format unstructured document into different document types automatically, and this is currently done manually by the documents user  The platform presently does not have NLG capabilities for text summarization. It also does not  The platform can identify and extract various data types such as tables, free-flowing text, have the ability to search through a repository of scanned documents based on entity or intent handwritten text, barcodes, logos, and stamps. It can also detect signatures

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 57 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Datamatics (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 It provides role-based access to the system, which can be segregated between development, test,  The platform only supports Latin languages for extraction and comes only with an English user and production environments. It can also time-stamp and audit every document stage interface. Lack of support for other regional languages from an extraction and user interface

 A native dashboard allows users to monitor operations and quality-related metrics such as standpoint limits its use documents processed, STP rate, manual interventions, and queue status. The platform can also be  While the platform can integrate with Datamatics’ RPA product “TruBot” and other applications via integrated with Datamatics’ proprietary BI product “TruBI” APIs, it currently does not have pre-built connectors for other third-party enterprise applications,

 It provides flexible deployment options such as physical or virtual desktop, on-premise, and on cloud legacy information systems, RPA tools, or BPM tools   It offers a variety of training methods including direct training by the vendor, via partners, classroom Currently, it offers volume-based pricing linked to the number of pages processed along with a training, as well as through online training courses. It also provides certification courses with licensing fee. Lack of other progressive output-based pricing models and flexibility in the separate courses for different roles commercial model may be a deterrent for certain enterprises   There is an embedded help tool in the platform as well as an active online user community It is currently not offered in a SaaS model, which could limit its adoption by small enterprises and SMBs  Clients are overall satisfied with the product and have appreciated the technical expertise of the vendor, the ability to understand business processes, and a partnership approach  Clients have indicated that it can improve the time taken to implement the solution by reducing the number of pre-deployment iterations and pre-empting likely issues

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 58 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

EdgeVerve (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 EdgeVerve is a wholly owned subsidiary of Infosys that develops intelligent automation solutions for  EdgeVerve’s client base is primarily based out of North America, Europe, and APAC, with limited RPA, IDP, and DPM / task mining. It envisions to empower enterprises to accelerate their digital presence in LATAM and MEA

transformation initiatives through its technology offerings  Nia DocAI has been deployed primarily for BFSI industry-specific use cases and for processing  Nia DocAI is its AI and NLP-based IDP platform. It comes with pre-built solutions for processing contracts and other legal documents. Therefore, the robustness of the solution for some other insurance claims receipts, checks, LIBOR contracts, and master-service agreements. The platform major business areas such as F&A, HR, contact center, and mailroom is somewhat untested

also has pre-trained AI models for field-value extraction from standard documents such as invoices  A majority of EdgeVerve’s IDP clients belong to high-tech & telecom, BFSI, and healthcare and  It supports ingestion of documents in different formats including TXT, PDF, JPG/JPEG, PNG, and pharma sectors, and it has limited experience of serving enterprises in other verticals such as multi-page TIFF files public sector, manufacturing, and CPG & retail

 Objects such as stamps, logos, barcodes, and signatures can be detected through the platform.  While the platform can detect and extract signatures, the ability to verify signatures against a Custom models for detecting and extracting other objects can also be created using Nia DocAI’s specimen sample is currently not available

training and deployment workbench  EdgeVerve lacks dedicated IDP application and support for mobile platforms. Therefore, it cannot  In addition to data extraction and document classification, it also has NLP capabilities and can be accessed through mobile devices, which could impede adoption of the platform among perform intent classification, phrase detection and matching, and sentiment analysis. It can also enterprises for mobile use cases perform abstractive and extractive summarization of text

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 59 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

EdgeVerve (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Nia DocAI provides a data validation and data resolution UI and API that can be used by admin  Nia DocAI can digitize documents in nearly 50 languages; however, support for document users to select existing validation rules or define new rules and apply them to fields. Users also processing has been tested for only English. The platform’s accuracy is yet to be proven for have the option to add or configure new fields for extraction languages other than English

 The platform provides an out-of-the-box analytics dashboard to measure and display performance  While Nia DocAI tightly integrates with RPA and DPM offerings of EdgeVerve, it lacks partnerships metrics around accuracy of the extracted fields, throughput time, and workforce analytics such as and pre-built connectors with RPA tools of other leading vendors, which could lead to longer average number of corrections made by the reviewer per document implementation and integration time for enterprises

 Nia DocAI has pre-built connectors for some of the leading enterprise applications such as SAP,  Nia DocAI lacks features to auto-redact sensitive information or mask/blur confidential data in Oracle, and Microsoft Dynamics. It tightly integrates with EdgeVerve’s in-house RPA AssistEdge documents. This could inhibit the adoption of the solution among enterprises processing RPA and its DPM offering AssistEdge Discover documents with highly sensitive information

 EdgeVerve offers training and certification courses through both online and classroom mode. Online  EdgeVerve does not offer a free community version of the software and also does not have an courses provide self-learning and instructor-led training modules online community forum for users. This could lower the reach of the software and may lead to longer issue resolution time

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 60 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Evolution AI (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Evolution AI aims to support enterprise IDP needs through its managed services offering, Evolution  Evolution AI mainly focuses on the BFSI market and, hence, its ability to serve enterprises from Catalyze, in addition to its software offerings, Evolution Transcribe and Evolution NLP. It has also other industry verticals remains untested

released a self-service product, funded by UK Research and Innovation, to increase IDP adoption  While it has experience of serving clients in North America, Europe, and the UK, it lacks presence among SMEs in emerging geographies such as APAC, MEA, and LATAM  The solution leverages its proprietary OCR, NLP, computer vision algorithms, and neural network-  It currently caters to large enterprises (revenue > US$5 billion) and SMBs (revenue < US$50 based models for classification and extraction of fields and tables from electronic documents in million); its experience of serving mid-sized enterprises is limited various formats including emails, images, website, and PDF documents in more than 40 languages  The solution can identify signatures; however, it lacks the ability to extract other complex data  It comes with pre-trained models on common use cases prevalent in the BFSI sector including types such as bar codes, logos, and stamps invoice, quarterly reports, shareholder lists, balance sheets, profit and loss statements, trade  It lacks the flexibility to allow enterprise users to add, configure, and manage business rules using finance, and consumer price index reports the platform; this is currently performed by Evolution AI staff  Its point-and-click interface enables enterprise users to train new models for extraction of fields as  While the solution provides pre-built connectors to widely used RPA platforms, integration with well as tables; users can also configure workflows, create teams, and edit roles of the users within BPM tools and BI platforms is available only through REST APIs the platform to ensure quality checks

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 61 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Evolution AI (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 The self-learning capability of the software enables annotators to train and fine-tune the model as  Currently, the system offers reports to track error rates and throughput details. However, it does they perform corrections during production. Users can perform validations through an online not provide visualizations for enterprise users. Enterprise clients also expect better dashboarding interface, which is accessible on any mobile device and analytics capabilities to monitor software and workforce performance

 The software can perform various image pre-processing actions on low-quality images, without the  Evolution AI currently offers a usage-based pricing model; hence, the commercial model may need of operator involvement. It also notifies the user in case the image quality drops beneath a seem less flexible for enterprises looking for outcome-based progressive models

defined threshold  While it leverages partnerships with RPA vendors for smoother integration, there is scope to  The company has recently released Evolution Metrics, which provides a new interface layer for further strengthen its partnership with resellers and system integrators to cater to a broader post-processing, enabling enterprise users to perform validations on the extracted data enterprise audience

 The software provides NLP capabilities such as intent detection, sentiment analysis, as well as text  Clients expressed the need for higher flexibility to enable business users to customize the solution search within long-form documents such as agreements without depending on the vendor

 Evolution AI offers flexible hosting options including SaaS, private cloud, and on-premise deployment

 Clients appreciate the robustness of the ML models and support provided to resolve issues; clients also appreciate the emphasis on providing enhanced user experience

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 62 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

HCL Technologies (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 EXACTO™ is an AI-based information extraction product designed specifically for reading and  HCL’s portfolio includes clients from BFSI, manufacturing, and healthcare & pharmaceuticals classifying handwritten and typed fax/image-based documents. It uses ML techniques to extract sectors. Its ability to serve clients in other verticals such as CPG & retail, public sector, high-tech & information from semi-structured and unstructured documents telecom is untested

 It comes with pre-trained models for various document types including annual reports, invoices,  It primarily caters to clients in North America, Europe, and APAC regions, with relatively fewer claims, KYC and ID proofs, checks, home loan forms, contracts, and bank transaction statements, clients in the emerging markets of LATAM and MEA

among others  While it supports Chinese and Latin languages, it currently does not support other Asian  Clients have deployed the platform for a variety of process areas such as F&A, procurement, HR, languages such as Japanese and Korean, which limits is ability to expand in these markets

and mailroom, as well as various industry-specific use cases such as BFSI and healthcare &  While the platform can integrate with information systems via APIs, it currently does not have pre- pharmaceuticals built connectors for common enterprise applications such as SAP, Oracle, and other legacy  It has a broad clientele across small, mid-size, and large enterprise buyer segments information systems. This could lead to a longer implementation and integration time for

 It provides flexible hosting options on desktop, server/on-premise, or on cloud platforms such as enterprises AWS and Azure. It is also directly offered by the vendor in a SaaS model

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 63 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

HCL Technologies (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 The platform can input documents across most formats such as PDF, JPG/JPEG, XLS, and CSV. It  While it provides classroom training and has an embedded help tool, it does not have any online can extract various data types such as tables, free-flowing text, handwritten text, bar codes, logos, training portal or certification courses, which could restrict the accessibility of the trainings

and stamps. It can also detect and verify signatures against a specimen sample  Clients have highlighted the scope to improve data extraction from poor-quality images  Confidential data such as credit card details can be masked/redacted by the system, which enables  Clients indicated that scaling of the product to multiple languages and the ability to translate non- the platform to be deployed for use cases involving sensitive information. It also allows role-based English information to English will help support global implementation access to the system and generates audit trails for clear usage tracking  Clients have also highlighted planning of the testing phase and time-to-market as potential areas  It has out-of-the-box integration with various third-party RPA and BPM tools of improvement  It offers flexible commercial models including progressive outcome-based pricing, where the client is charged only for the documents where accuracy matches with the agreed upon values

 Clients have appreciated its focus on creating business value and taking the relationship beyond the contract. They recognize the strong relationship between the development team and the business unit in an offshore model as one of its strengths

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 64 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Hypatos (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Hypatos focuses on providing broad-based document processing automation including content  Hypatos mostly serves clients in Continental Europe, APAC, and North America. It is yet to validation, master data matching, and attribute enrichment, in addition to document classification establish its footprint in other geographies including the UK, MEA, and LATAM

and extraction capabilities, for back-office processes  It has primarily deployed IDP solutions in F&A and HR processes and has limited experience in  Hypatos offers a variety of pre-trained models for common use cases such as invoice-to-pay other key functions and industry-specific processes

processing, order processing, travel report processing, and P&C and health claims. It also provides  The solution lacks advanced capabilities such as the ability to recognize and validate signatures, custom models based on client requirements compare unstructured documents, and conduct sentiment analysis – which may be a deterrent for  The software can ingest documents in various formats such as TXT, JPG, PDF, PNG, XLS, XML, clients looking to handle complex unstructured documents such as agreements

JSON, and CSV files to extract data from free-flowing text as well as handwritten text, including  Validation rules integration, master data matching, and validation look-ups can only be configured extraction of bar codes, logos, and stamps by a technical user in the back-end. Addition of this feature in the user interface for business users  The Hypatos Studio interface allows operators to annotate document images to create training data without any technical background is currently in the roadmap

for new document type models or apply an existing base model and validate results to create  Similarly, user interface to allow business users to configure confidence level / accuracy threshold training data for model training for classification and extraction is also in the roadmap  It integrates with generally available OCR engines and applies various pre- and post-processing techniques to ensure better extraction accuracy

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 65 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Hypatos (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 The platform calculates a risk score (based on statistical confidence, heuristics, and value-at-risk)  While it has partnerships with RPA vendors, the solution lacks pre-built integrations with leading for validation of extracted results and determining straight through processing. It highlights fields enterprise applications such as Microsoft, SAP, and Oracle, and BPM tools – which may be a with low-risk score, missing information, and business rules violation in the review GUI for deterrent for clients looking for quick integration with their existing business applications and corrections by business users automation ecosystem

 Hypatos offers a comprehensive monitoring and analytics service, built on Grafana platform, that  The user interface is available only in English and German, rendering it unsuitable for clients includes system-related metrics (such as up-time and CPU usage), processing metrics (such as looking for UI in regional languages

pages processed and average processing time), and business KPIs (such as accuracy and STP  The solution does not lend itself well for clients looking for embedded help tool, online user rate per document) community and support forum, or online training support for business users  The platform can be hosted on private/public/hybrid cloud and on-premise, providing deployment  Hypatos has limited partnerships with service providers and system integrators, which might be a flexibility to enterprises deterrent to serving a broader set of enterprises  Clients are highly satisfied with the performance of the machine learning algorithms and flexibility &  It mainly offers subscription-based fixed capacity and usage-based pricing constructs and lacks ease of deployment of the solution. They appreciate the ease of use of the software and experience in offering progressive pricing models such as outcome-based models proactiveness in understanding business requirements and continuously updating the solution  Clients expect Hypatos to develop direct interfaces with leading BPM and workflow tools, enabling easier development and integration

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 66 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Hyperscience (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Hyperscience has a vision to provide AI-enabled document processing software focused on  The majority of clients are based out of North America, and it has low presence across some assisting enterprise employees in data entry tasks from different types of documents such as mature markets such as Europe and emerging economies such as APAC and MEA

handwritten, printed, and images into enterprise IT systems  BFSI, government and public sector, and professional services have been the key focus industries  It provides an automated training platform for the ML model, which runs in the background of the for Hyperscience, and it has limited experience in serving enterprises in other verticals such as software. Users can enhance the accuracy of the model through human-in-the-loop review and high-tech and telecom, manufacturing, and CPG & retail

feedback on extracted fields  Its IDP platform has been predominantly deployed for F&A, BPO, and BFSI industry-specific use  Hyperscience experienced strong YoY growth in terms of revenue and clients in 2020, driven cases; the robustness of the platform for business areas such as HR, procurement, contact center, primarily by BFSI and government and public sector in the North American market and other vertical-specific processes is somewhat untested

 Its IDP platform comes with pre-built models to identify intent and process documents with printed  Clientele of Hyperscience is skewed toward large enterprises and it has relatively lower and handwritten text including cursive text to extract fields. It also offers out-of-the-box models for experience of serving enterprises in mid-size, small, and SMB segments

document types such as invoices, paystubs, and checks  It presently supports document processing in only four languages – English, French, Spanish, and  Documents can be ingested in a variety of formats including TXT, PDF, JPG/JPEG, PNG, XLS, and German. Therefore, it may not be the best option for enterprises looking to process documents in DOCX. It can also detect objects such as bar codes and does signature extraction and verification Asian and Middle Eastern languages

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 67 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Hyperscience (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Hyperscience provides an out-of-the-box analytics and reporting capability. It allows users to set  While the platform has NLP capabilities and can also classify emails and documents into different their own SLA accuracy metrics for tracking software performance. It provides insights around field- categories, it presently lacks NLG capabilities and may not be well-suited for summarizing long- level accuracy, STP rates, and human workforce performance form text in documents

 The platform has built-in capability to auto-redact sensitive information when new layouts are being  Hyperscience partners with some of the leading RPA vendors for integrating its IDP platform with created based on populated documents. To further match enterprise standards for security, it offers RPA technology; however, it lacks pre-built connectors to enterprise IT systems such as SAP, data encryption and role-based access control Oracle, and Microsoft Dynamics, which could lead to longer set-up and integration time

 The product architecture is developed on microservices and also supports multi-tenant  It presently does not provide any online portal for training or certification of users. Additionally, it deployments. The application can also be deployed as a containerized solution using Docker also has relatively few service provider partners, which could be a constraint in accessing training

 Referenced clients have lauded customer support and the clarity of communication of the marketing resources for enterprises team. They have also highlighted the promptness of the team in listening to customer feedback and  Clients have indicated the commercial model and cost of the product as an inhibitor for adoption, incorporating changes especially for enterprises in the small and SMB segments. SaaS version of the software is also

 Clients appreciate the platform’s user interface, the accuracy of document processing, and the presently part of the roadmap, which could lower the TCO for enterprises overall product roadmap and enhancements made by the company  Clients have also voiced a need for product documentation and constituting an online community and support forum to empower users to solve product queries

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 68 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Indico (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Indico aims to provide strong AI/ML models for document extraction and also facilitates a human-in-  Indico currently serves clients in North America and lacks presence in other key geographies such the-loop validation capability for high accuracy through its IDP platform as Continental Europe, the UK, and APAC

 In addition to data extraction from semi-structured and unstructured documents, it is capable of  Its portfolio is heavily skewed toward clients in BFSI, manufacturing, and professional services processing data from multimedia formats such as images and videos sectors. Its ability to serve enterprises from other industry verticals is untested

 Enterprise users can build custom machine learning models on the platform and orchestrate  Indico does not provide pre-packaged solutions and only offers pre-trained models that users can multiple models in a workflow through a browser-based point-and-click interface for classification leverage for building custom models

and extraction  Currently, the platform supports English language; support for other Latin-based languages,  The platform comprises broadly three modules: the Teach module enables users to upload training Chinese, and Japanese is in the roadmap

samples and label them; the Review module allows users to define workflow queues and provides  While the platform provides integration support via webservices, it lacks pre-built connectors for human-in-the-loop interface for review and validation during production; the Explain module enables leading enterprise applications such as SAP, Oracle, and Microsoft as well as third-party BPM users to understand and improve performance of the developed custom models tools  The Review interface displays both the extracted data and document and highlights exceptions in  The reporting feature provides statistics on metrics such as the time taken per task, STP at the the corresponding fields for easy correction and validation field level, and exception rates. However, the platform lacks visualization and analytics dashboard for tracking and monitoring the software as well as workforce performance

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 69 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Indico (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Indico offers out-of-the-box integrations with leading RPA solutions such as Automation Anywhere,  The platform lacks flexibility to allow enterprise users to add and configure business rules for Blue Prism, and UiPath to enable easier deployment within client systems validation of extracted field and it is currently in the product roadmap

 The platform can be deployed on client server as well as on private and public cloud, based on  At present, it does not provide NLG features for text summarization in unstructured documents client data and security requirements; the solution is also delivered as a SaaS offering for and long-form text such as legal contracts and news articles

enterprises seeking a lower TCO  Enterprise users require some basic data science knowledge to build custom models and leverage  It provides SSO/SAML integration that enables information security, user authorization, and role the full potential of the platform

management. Users can also build redaction models to automatically redact confidential and  Indico provides common pricing models such as usage-based, fixed capacity-based, and process- sensitive information in documents based models; however, it does not offer progressive models such as outcome-based pricing  Clients appreciate the accuracy of extraction for unstructured documents, the ability to cater to structure

customer needs, and its speed of innovation  Clients highlighted that training and documentation for business users can be improved for better  Clients have also expressed a high level of satisfaction with cognitive features of the software such support; clients also indicated lack of clarity on the roll-out plan for new features as a limitation as machine learning models, computer vision, and NLP

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 70 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Infrrd (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Infrrd offers a low-code AI-based IDP platform, which allows business users to build and train  Infrrd’s clientele is primarily based in North America and its experience of serving enterprises in models across semi-structured and unstructured documents. It provides pre-trained solutions for other key geographies such as the UK, Continental Europe, and APAC is limited

faster time to deployment as well as self-service controls to allow the flexibility to build custom  The majority of its clients use the platform for F&A, BFSI industry-specific, and web-based models processes. Therefore, its capability to serve enterprises in other process areas / business  It provides pre-built out-of-the-box solutions for various use cases / process areas such as bank & functions is somewhat untested

wealth management statements, claims processing, forms extraction, handwritten forms, insurance  The platform can integrate with third-party complementary technology solutions via APIs and documents, invoices, mortgage documents, and receipts, among others comes with out-of-the-box integrations with third-party RPA tools; however, it does not have any  Ingested documents can be classified into different document types at a document or page level. A pre-built connectors for other enterprise applications (SAP, Oracle, Microsoft, etc.) or BPM tools

patented technique using NLP and CNN is used to classify documents by understanding contextual  While API integration with third-party BI platforms is available for generating insights, users cannot relationships between documents/pages generate custom reports using the native dashboards within the platform  It also has a managed services offering, wherein the human-in-the-loop validation is also managed by it

 Business users can create custom documents, add new document types to the platform, and train models without any ML expertise in a low-code environment

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 71 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Infrrd (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 The platform allows business users to define and monitor queues to correct and validate extracted  The platform does not have the ability to auto-redact information nor does it allow the user to data based on configurable field-level confidence thresholds. They can also define validation rules mask/blur confidential data, which limits its use when dealing with sensitive information/documents

for the extracted data  It currently supports 16 languages for extraction; however, it cannot process multiple languages  An analytics dashboard is available to monitor performance at a model- and corrections- level. The within the same document and the user interface is available only in English, which may limit its model dashboard allows the user to track key metrics such as STP rate, model confidence score, ability to serve enterprises that need broader language support

and processing time, whereas the corrections dashboard allows managers to track user  It provides virtual classroom training to users. However, lack of training via partners as well as performance, corrections made, and correction time online training and certification courses limits the accessibility of its training program  It offers an outcome-based commercial model tied to field-level accuracy  Clients have indicated scope to provide better clarity on its product vision and roadmap, and have  Clients have expressed an overall high level of satisfaction with the product, appreciating its also pointed out a need for better communication on the infrastructure requirements of the extraction capability, scalability, and cognitive features such as NLP and computer vision platform

 Clients also appreciate the quality of the people across different teams such as IT, sales, and management, their ability to understand the client's business, and the responsiveness of the customer support teams

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 72 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

JIFFY.ai (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 JIFFY.ai aims to provide intelligent self-service capability to process structured, semi-structured,  While JIFFY.ai serves clients in North America, Continental Europe, MEA, and APAC, it lacks and unstructured documents through its JIFFY.ai Automate platform. Its vision is to provide a experience in serving clients in other geographies such as the UK and LATAM

configurable IDP solution that allows configuration of the key components needed for straight  It has deployed the solution in F&A and BFSI industry-specific use cases and has limited through processing experience in other key processes including contact centers, mailroom, and healthcare industry-  The platform can classify and extract data from various input file formats and data types including specific processes

handwritten text, bar codes, logos, and stamps  It has a relatively low number of implementation and training partners, which could limit its ability  It provides in-built models for various use cases including invoice processing, financial statements, to serve a broader client portfolio

mortgage documents, and service ticket automation. Custom models for extraction and  The company currently offers a usage-based commercial model, rendering it unsuitable for clients classification of documents can also be generated based on client requirement looking for outcome-based progressive pricing constructs  The ML models can be trained using historical data as well as by business user using the point-and- click interface. Any correction done by the user on the user validation interface is also recaptured for reinforced learning

 JIFFY.ai has enabled its partners and customers to build browser-based HyperApps to solve industry-specific problems such as invoice processing, trade finance, and mortgage. The HyperApps and automations developed are available on a centrally managed library

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 73 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

JIFFY.ai (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 The system is bundled with commonly available OCRs including ABBYY, Tesseract, V7, Ephesoft,  The platform lacks the ability to auto-redact or allow users to mask/blur confidential information – and Google Vision. However, it is an OCR-agnostic platform and can work with any OCR available which can be a roadblock for clients with high data security requirements. However, this is in its with the customer roadmap

 Its task designer is a web-based, drag-and-drop interface that allows developers to configure  NLP capabilities are offered for documents only in English. Moreover, the out-of-the-box user complex workflows across multiple systems and interfaces in a low-code environment interface is also available only in English. This can be a deterrent for clients looking for IDP

 It provides standard rule sets as well as allows users to configure business rules for data solutions to process unstructured documents in regional languages transformation and validation  Clients expect a more robust product roadmap, faster implementation timeline, and better

 JIFFY.ai provides its own analytics and visualization layer to help track and monitor SLAs, extraction performance and stability of the solution accuracy, and manual worker performance  Clients also cited focus on user experience and interfaces as other areas of improvement

 The solution can be hosted on private/public cloud as well as on-premise. It is also available as a SaaS offering, providing flexibility of deployment to its clients

 Clients appreciate the company’s flexibility and responsiveness in providing customer support and services

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 74 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Nividous (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Nividous aims to expedite enterprises’ digital transformation journey through its integrated  Nividous is relatively new to the IDP market and launched its IDP offering in 2018. Its ability to automation offering that encompasses RPA, IDP, process mining, BPM, and insights from analytics. maintain the robustness of customer support and meet product requirements as its clientele It started offering its IDP solution in 2018 expands is somewhat untested

 Nividous leverages its proprietary OCR and computer vision techniques for extracting data from  Nividous’ IDP client base primarily belongs to BFSI, healthcare, and manufacturing verticals in documents. It has a layer of ML model for classification and extraction of data from unstructured North America and APAC regions. It has relatively low experience of serving enterprises in other documents, which can be trained through past data or continuous feedback geographies and verticals

 It also natively offers NLP capabilities based on deep learning algorithms for text classification,  While its IDP platform has been deployed for some key processes such as F&A, contact center, entity recognition, and sentiment analysis. It also allows users to choose from a list of ML models and vertical-specific use cases for insurance and healthcare-payer industries, the robustness of available OOTB for processing documents the platform is somewhat untested for procurement, HR, mailroom, and banking industry-specific

 The platform offers pre-trained models for processing standard templates such as identity cards for use cases KYC verification. It combines other technology components such as RPA and BPM with IDP to offer  It presently supports only English language for document extraction. This could be a deterrent in pre-packaged solutions for standard processes such as accounts payable, medical coding, service adoption for global enterprises looking to deploy IDP solution for regional languages such as desk automation, and insurance processing Arabic, Chinese, Japanese, and Cyrillic languages

 On-device processing of documents on mobile devices is supported through a dedicated application  Nividous does not offer the platform as a SaaS offering. This could deter adoption by smaller to upload and process documents enterprises looking for lower upfront infrastructure investments and deployment time

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 75 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Nividous (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 The platform allows users to add, configure, or modify fields to be extracted. Enterprise users can  While the platform provides out-of-the-box dashboard for analytics and reporting, it presently does also configure and manage business validation rules for the extracted fields. It also highlights fields not offer pre-built connectors for third-party BI platform providers (Tableau, Power BI, etc.). This based on low confidence scores or violation of business rules could be a constraint for enterprises in gathering insights from the platform

 It provides key insights around the extracted documents such as confidence level, STP rates, field-  Training to enterprise users is provided through classroom mode and only in India and the US. level accuracy, and workforce performance metrics. It also displays confidence levels of extracted Online training and user community portal with self-paced training modules is part of the roadmap fields and allows users to configure the threshold for extraction for 2021

 Enterprise users can develop custom reports for monitoring and analytics. Further, user actions are  It has very few technology vendors and service providers in its partner ecosystem. This could limit also logged in audit trails with time-stamps the reach of the product and make it less attractive for enterprises looking to integrate with best-of-

 Enterprises can create tailored roles for accessing the platform based on their requirements. It also breed vendors of complementary technologies such as RPA, IVA, process mining, and BPM provides capability to mask/blur sensitive information in documents to ensure confidentiality  Clients have indicated the need for a smoother implementation and deployment process by

 Referenced clients have highlighted a flexible licensing model, customer support, and an integrated providing more out-of-the-box models and better managing the infrastructure requirements for platform out-of-the-box as key areas of strength for Nividous deployment

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 76 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Parascript (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Parascript’s FormXtra.AI uses cognitive technologies to classify and extract data from documents in  A large part of Parascript’s clientele is based out of North America, the UK, and Continental a low-code environment. Its Smart Learning module is a self-learning module that automatically Europe, and its experience of serving enterprises in other geographies such as APAC is limited

chooses the ML models needed for optimized results  It provides solutions for the BFSI and government sector; however, it lacks experience of serving  FormXtra.AI uses reinforcement learning for its human-in-the-loop training, which works on firms in other sectors such as healthcare and pharma, high-tech & telecom, media and background data verification to curate the learning data set, compares the results of learning with entertainment, and professional services

past performance, and automatically sets document, page, and field-level thresholds. It also allows  Its current client portfolio is slightly skewed toward large enterprises with revenues greater than bulk import of historic data to train the system in an unsupervised mode US$5 billion. Therefore, its ability to successfully cater to small enterprises and SMBs is  The platform provides a software learning system that can automate various aspects of document somewhat untested

processing such as image pre-processing, document classification, splitting the document, and data  Currently, the platform is not offered in a SaaS model, which could deter adoption in enterprises extraction across various document and data types looking for a lower total cost of ownership, especially enterprises in small and SMB segments  It can process various data types within structured and semi-structured documents such as free-  The platform currently has limited capabilities in processing of unstructured documents with long- flowing text, handwritten text, bar codes, and logos, and also has the ability to detect and verify form text such as contracts, lease agreements, letters, and news articles signatures against a specimen signature

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 77 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Parascript (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 It comes with pre-built models that can be used out-of-the-box for various use cases such as claims,  While it has the capability to extract English and other Latin languages, the platform does not invoices, checks, signature location and verification, and mortgage document classification support processing of documents in Asian or Middle Eastern languages. The UI is currently

 It offers services such as data collection, quality assurance, and model training based on customer- available only in English language specific needs. It can also use synthetic data when there are issues related to data accessibility  Clients have indicated integration with other complementary technologies as one of the limitations.

 Enterprise users can define business validation rules at a field-level through a GUI as well as FormXtra.AI can currently integrate with third-party platforms via APIs and comes with out-of-the- validate extracted fields using external data box integrations for some RPA tools, however, it does not have pre-built connectors for other enterprise applications or BPM tools  The solution is flexible in terms of hosting options as it can be hosted both on-premise as well as on public/private cloud  Clients have highlighted that availability of analytics capabilities would help generate better insights from the platform. FormXtra.AI currently does not have a native analytics dashboard and  Clients have appreciated the platform’s ability to process handwritten documents and its self- also does not offer pre-built integrations with third-party BI platforms learning capabilities  Clients have also expressed a need for higher flexibility in the available commercial models.  Clients have also indicated pricing of the product and technical support provided by the team as Parascript currently offers a usage-based model with scope to offer more progressive models that major strengths of Parascript are tied to outcomes

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 78 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Rossum (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Rossum’s vision is to automate document-related enterprise communications with IDP as a key  While Rossum’s portfolio presently includes clients from travel & logistics, high-tech and telecom, component, leveraging deep learning to enable template-independent document extraction professional services, and manufacturing sectors, its ability to serve firms from other sectors such

 The platform uses a proprietary OCR, deep learning-based neural networks, and computer vision to as BFSI and healthcare is somewhat untested perform image pre-processing, classification, and extraction of semi-structured documents  While it serves clients in North America, the UK, and Continental Europe, its experience in APAC,

 The platform comes with pre-trained models for documents such as invoices and purchase orders, LATAM, and MEA is relatively low. Also, it has relatively lower experience serving large enterprise which can be further extended to other document types such as delivery notes, packing lists, and clients (revenue > US$5 billion) bills of lading  The platform does not offer flexibility in deployment. It is currently offered only as a cloud-based

 It offers a "Dedicated AI Engine" as an add-on service, which enables business users to fine-tune SaaS platform and cannot be deployed on-premise, which could inhibit the adoption for the engine for their specific use case and deploy a customized data extraction model. It also allows enterprises looking for an on-premise solution users to add and train custom fields for extraction and continuously trains the AI models on  It presently lacks NLP capabilities such as named entity recognition, sentiment analysis, and text production data via a human-in-the-loop summarization, which may deter clients looking for processing of unstructured documents and long-form text

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 79 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Rossum (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 The platform comes with a native analytics dashboard, which tracks metrics such as worker  While the platform allows integrations with other systems through APIs and comes with pre-built performance, field level accuracy, and process-level SLAs connectors for SAP/Oracle applications and third-party RPA tools, clients feel that the quality of

 Over 20 languages are supported by the platform for extraction including most Latin languages, the existing connectors and their approach toward integrations can be improved Arabic, Korean, Greek, and Hindi. The UI of the platform can also be configured in four languages –  The platform allows roles-based access to the system; however, it currently does not have the English, German, Czech, and Slovakian ability to auto-redact or allow masking of data, which limits its use when dealing with sensitive

 Clients have appreciated the intuitiveness of the user interface and have called out its proactiveness information in communicating its product vision and roadmap, while continuously developing new features and  Configuring business validation rules within the platform requires some technical know-how, functionality based on client requirements necessitating editing of schema or code which may not be suitable for enterprise business users

 Clients have also indicated the quick implementation time of the solution and the customer support  The platform does not detect and verify stamps and signatures, and ingestion of documents in provided as strengths CSV, TXT, and XLS formats is currently not supported

 Clients have indicated that there is scope to increase the accuracy of its pre-trained models, which would help them realize a faster return on their investments

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 80 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Singularity Systems (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 Singularity Systems applies deep neural networks to extract data from unstructured documents and  Singularity Systems serves clients in APAC and North America; it is yet to establish its presence in images. It aims to provide an intuitive software that allows regular business users with no data other geographies. Hence, its ability to serve clients in LATAM, Continental Europe, MEA, and the science background to build custom solutions in a no/low-code environment UK is somewhat untested

 The SinguAI platform comprises four engines – SinguTXT (NLP solution), SinguIMG (computer  Its clientele primarily includes buyers from BFSI and manufacturing industries; thus, it has limited vision solution), SinguOCR (customer OCR solution), and SinguPREDICT (predictive analytics experience in catering to some high-growth industries such as healthcare, high-tech and telecom, solution) – that work in tandem to extract data from documents and images in multiple formats and travel & logistics

 It offers packaged out-of-the-box solutions for classification and extraction of business contract  Its client portfolio is heavily skewed toward large enterprises having annual revenue greater than information, bank statements, medical reports, customs declaration, identity documents, and US$5 billion, with limited experience of serving small and medium-sized enterprises

financial statements  Currently, the solution supports extraction of documents only in English and Chinese – this can be  The software can gather training data from databases, PDF/text archives, and any existing data a roadblock for enterprises looking for multiple language processing

repositories in addition to human-in-the-loop training that enables real-time AI. The OCR is trained  While end-users can manage and configure validation rules from the platform, validation using through the "Automatic Feeding Machine" that crawls text online and feeds it into the training external data sources can be performed only at the back-end – limiting the flexibility for business module with no supervision users

 Presently, it lacks a robust partner ecosystem and primarily relies on direct sales channels, which may limit its capability to cater to a broader set of enterprises

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 81 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Singularity Systems (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 The solution comes with a front-end labeling tool and administrator user interface to enable data  It lacks pre-built integration with third-party BI platforms, RPA tools, and BPM tools, which may labeling and manage projects in production. The user interface also highlights the fields with errors deter clients looking for an IDP solution that readily integrates with their existing systems

and exceptions for review by enterprise user  The solution tracks sensitive data within documents; however, it lacks the ability to auto-redact the  It performs various pre-processing techniques (such as greyscale binarization, blurring, skew information. The ability to mask/blur confidential data by enterprise user is available only on correction, and dilation) to adjust the quality of captured images and documents demand

 It allows business users to configure the confidence threshold for classification and extraction,  The analytics dashboard currently provides metrics around accuracy rates and process-level providing them the ability to control the exception queue SLAs. Clients looking for comprehensive dashboards with drill-down options for tracking operator

 The product is built on microservices architecture, supporting multi-tenant deployment. It can be performance besides other metrics may find this solution unsuitable hosted on cloud and on-premise and is also offered as a SaaS offering  Singularity Systems offers training only via classroom programs. Lack of training via partners,

 Clients highlighted the accuracy and turnaround time of the machine learning models, the ability to online training portal, and online community forum limits its ability to train and increase adoption train the model in real-time, and ease of use for non-technical business users as key strengths among a broader set of enterprises  Clients stated that there is scope to improve the solution to cater to highly complex business scenarios. Few clients also stated that model configuration is not very easy and there is scope for improvement

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 82 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

UST SmartOps (page 1 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 UST SmartOps envisions to help enterprises in transforming disparate data from unstructured  UST SmartOps’ current client portfolio for IDP is skewed toward the healthcare and pharma documents into valuable information using a combination of computer vision, AI, and ML techniques sector; therefore, its ability to serve enterprises in other industries such as banking & capital through its SmartVision™ platform markets, public sector, and professional services is somewhat untested

 The platform comes with pre-packaged solutions, which can be used out-of-the-box for various use  A large part of its clientele is based out of North America and APAC. Its experience of serving cases such as KYC, invoices, contract management, check extraction, and annual stakeholder enterprises in other geographies such as the UK, Continental Europe, LATAM, and MEA is limited

reports  The platform can integrate with third-party tools via REST APIs; however, it does not have pre-  The platform uses NLP capabilities to perform sentiment analysis, entity extraction, and intent built connectors for various enterprise applications, legacy information systems, or BPM tools

detection from emails and plain text  The platform can currently process handwritten documents via partner integrations; however, it  A native analytics dashboard is available, which tracks various performance metrics such as field- does not have the ability to extract logos, stamps, and signatures

level accuracy and manual time spent for document review. Integration with Grafana enables users  While the platform operates via a browser interface and is compatible on mobile devices, a native to build custom dashboards. It also keeps records of each batch of documents processed for the mobile application is not available purpose of auditing  The user interface only supports two languages – English and Spanish – which may limit its  The platform comes with a skills workflow designer where enterprise users can edit existing application across different geographies workflows or create new ones using a drag-and-drop interface

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 83 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

UST SmartOps (page 2 of 2) Everest Group assessment – Major Contender

Measure of capability: High Low

Strengths Limitations

 The platform allows role-based access control, which can be configured based on the business  While the platform supports various languages for extraction, it currently cannot process multiple requirements. The roles can be segregated into development, test, and production environments languages within the same document without manual intervention. It also supports only English

 It also provides out-of-the-box integration with UiPath’s RPA tool, which allows enterprises to language for handwritten documents integrate with their existing UiPath platform and trigger automation  It has limited partnerships with service providers and system integrators, which might be a

 The platform is offered through a PaaS model, which reduces time to implementation. It has the deterrent to serving a broader set of enterprises flexibility to be deployed via any public cloud. It is also offered on-premise, albeit as a lighter version  While the vendor provides training to users based on requirement, lack of training via partners and

 It offers high flexibility in terms of commercial models and pricing. Apart from usage-based pricing, it online training modules limits the accessibility of its training programs also offers outcome-based pricing and a gainshare model where the customer has no upfront cost,  It currently does not have NLG capabilities that can be used to summarize long-form text from enabling a faster ROI documents

 Aspirants – GuardX – i3systems 05 – qBotica – SortSpoke – TAIGER

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 85 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

GuardX (page 1 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 GuardX envisions to provide an enterprise-grade platform to help individuals and organizations  Foreseer AI is a relatively new product in the IDP space and has limited experience of serving extract data from unstructured documents, through its IDP platform – Foreseer AI enterprise clients. It primarily focuses on the BFSI segment within North America and has low

 It provides pre-packaged solutions to process documents focused on financial institutions. Pre-built presence beyond these domains out-of-the-box use cases provided by the platform include data extraction from US public finances,  The platform can presently extract data from structured, semi-structured, and unstructured loan agreements, and real estate deals documents. However, it does not have the capabilities to detect and extract data types such as

 The platform leverages OCR engines of ABBYY and Tesseract along with its in-house OCR. The stamps, logos, and signatures extraction model can be trained based on inputs received from users as part of the human-in-the-  It also has limited capabilities in processing handwritten documents, with the primary focus being loop process vertical-specific printed financial documents

 Some of the key IDP capabilities of the platform include extraction of data from nested tables,  Presently, it does not provide an option for users to add/modify or configure a new field for classification of pages, and GUI with drag-and-drop features to define/edit the process. Additionally, extraction. Configuring business validation rules based on external databases and lookups is also it also provides NLP capabilities for analyzing free-flowing text part of the roadmap

 It offers an out-of-the-box analytics and reporting tool to measure the software’s performance on  While the platform has NLP capabilities, it lacks sentiment analysis of text and comparison of broadly two categories – quality of data extraction and processing time documents based on intent and meaning. NLG functionalities to summarize long-form text is part of the roadmap and is expected to be made available in upcoming releases

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 86 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

GuardX (page 2 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 It provides insights such as accuracy levels of fields extracted, percentage of false positives, and  Although the platform provides an out-of-the-box analytics and reporting capability, it does not modified and missed data points. Further, users can also drill down for more granular analysis for allow users to generate custom reports by defining KPIs based on enterprise requirements

the time spent by users on data validation based on a geography or document type  It lacks features to auto-redact sensitive information or mask/blur confidential data in the  Developed on a microservices architecture with support for containerization, Foreseer AI can be documents. This could inhibit the adoption of the solution among enterprises processing deployed on-premise, on private cloud, and through managed services-based models. Public cloud documents with highly sensitive information

deployment is presently part of the roadmap  Foreseer AI has very few service provider and technology partners. Enterprises, especially with  Foreseer AI offers a flexible commercial model where customers have the options to choose from centers in remote locations, could face challenges in implementation and support as the vendor an enterprise-wide licensing and usage-based pricing. It also offers variable price based on the expands its global footprint

volume of pages processed  GuardX does not offer an online training and certification portal and courses are offered only  Referenced clients have overall expressed a high level of satisfaction with the vendor, appreciating through classroom mode in North America and India for IDP. This could restrict the accessibility of the flexibility and ease of deployment of the solution, customer support, and the understanding of the training program for its clients

financial domain and enterprise requirements of the vendor  Clients have indicated a need for improvement in NLP capabilities and a more holistic and broader approach for intelligent automation to better serve enterprises

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 87 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 i3systems (page 1 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 i3systems has built an integrated product that augments IDP with cognitive AI decision models to  i3systems’ portfolio only includes clients from the insurance sector and its solution has been automate business processes along with data capture deployed only for insurance-specific use cases. Hence, Its ability to serve clients in other sectors

 It can process both semi-structured and unstructured documents with free-flowing text for medical such as banking & capital markets, healthcare & pharmaceuticals, and manufacturing as well as claim-related documents such as discharge summaries other use cases such as F&A, procurement, and HR is untested   The platform comes with pre-trained models for use cases such as financial/medical adjudication, A majority of its clients are based out of the APAC region. It does not have experience serving medical underwriting, invoices, and structured documents such as IDs and forms clients in other geographies such as North America and Europe   The solution provides flexibility in terms of deployment options. It can be hosted by the client on- Currently, it only has experience serving small enterprises and SMBs, and its ability to serve mid- premise, on desktop/laptop, or over public/private cloud. It is also offered in a SaaS model, which size and large enterprises with revenues in excess of US$1 billion is yet to be proven reduces the total cost of ownership for small enterprises and SMBs  The platform currently does not have the capability to process data types such as bar codes,

 It can classify documents into different document types using language models. It leverages NLP logos, stamps, or signatures. Handwriting extraction is supported via a partner capabilities to extract phrases and numerical data from free-flowing text, as well as semi-structured  It does not allow enterprise users to add or modify fields to be extracted. This limits the use of the documents platform only to fields already available in the platform

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 88 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 i3systems (page 2 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 The platform provides roles-based access to the system and the number of roles can be configured  The level of interoperability with other systems is low. While it comes with APIs to create based on business needs. It can also time-stamp each user action to maintain audit trails integrations, it does not currently have pre-built connectors for other enterprise applications,

 It provides different types of roles for users to access the platform based on business process legacy systems, RPA, or BPM tools requirements. It also gives users the option to review and mask/blur confidential data in documents  It supports only two languages – English and Arabic – for both extraction and from a UI

 It offers flexible commercial models such as per process-based, usage-based (per page or perspective. Lack of support for other languages limits its use for global operations document), and fixed capacity-based pricing. It can also provide SLAs that are tied to field-level  While users can download operational data from the platform and create MIS reports in CSV accuracy format, the platform currently does not provide visualizations within the GUI for tracking metrics

 Clients have indicated that i3systems has a stable platform and is flexible to accommodate such as accuracy levels and time spent manually for document review customer requirements in the product roadmap  Users do not have the ability to configure confidence levels or accuracy threshold for extraction or classification

 Clients have indicated that there is scope to reduce the implementation and turnaround time for the solution. They also mentioned that a more streamlined QC process would be beneficial for enterprise users

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 89 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 qBotica (page 1 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 qBotica aims to create digital BPOs through an ecosystem of solutions that automate enterprise  qBotica’s portfolio includes clients from the banking & capital markets, manufacturing, and functions traditionally performed by humans. Its IDP product, Doqument, can classify and extract professional services sectors. Hence, its ability to serve clients in other sectors such as healthcare data from semi-structured and unstructured documents for a variety of use cases such as invoice & pharmaceuticals, insurance, CPG & retail, and public sector is untested

processing, resume scoring, and tax forms processing  Clients have deployed Doqument for process areas such as legal, F&A, procurement, and HR.  The platform comes with pre-trained solutions for various use cases such as invoices, purchase The platform’s capabilities in other business areas such as contact center and mailroom as well as orders, W2 forms, and structured handwritten forms other industry-specific process areas such as banking, pharmaceuticals, and utilities are yet to be

 NLP is used for document search capabilities. It can extract, read, and understand document demonstrated clauses and compare them against a set of clauses to provide a match  Currently, all its clients are based out of North America; therefore, its ability to serve clients in

 It can auto-redact sensitive information as specified by the client and allows users to mask/blur other regions such as APAC, Europe, LATAM, and MEA is somewhat untested confidential data, which makes it unavailable for any processing downstream. This allows the  While the platform uses APIs for integrations and comes with out-of-the-box integrations for RPA platform to be deployed in process areas that deal with sensitive information and BPM tools, it currently does not have pre-built connectors for other enterprise applications or

 It supports extraction and processing of documents in over 200 languages for printed text, and four legacy information systems languages for handwritten text  It does not allow enterprise users to add or modify fields to be extracted from a document, nor does it allow users to add a new document type for extraction

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 90 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 qBotica (page 2 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 Errors and exceptions get flagged by the system based on low confidence scores, missing  It currently does not have the ability to perform sentiment analysis from the extracted text, or NLG information, or violation of business rules, which are then sent for human verification. Users can capabilities to summarize text in documents

also configure the confidence score for extraction at a field level  The UI is currently available only in English, which may limit its use in certain regions  The platform can be hosted on-premise or can be availed as a SaaS offering hosted on a virtual  Lack of progressive commercial models such as outcome-based pricing can be a deterrent for private cloud. It can also be hosted on laptop/desktop for small scale implementations certain organizations. Its current commercial model consists of fixed platform fee plus volume-  A native dashboard is available, which tracks various metrics and can be customized as per client based pricing linked to the number of pages processed

needs  Clients have indicated that there is scope to improve the user training provided by the vendor.  Clients have appreciated the platform for its technological capabilities, design, and ease of Currently, self-paced online training and certification courses are not available. It can also improve implementation its communication around complex AI terminologies for better understanding by users

 Clients also praised the qBotica team for its flexibility in terms of tailoring the solution to suit  Clients have also highlighted a need for teams and support in other geographies to ease scaling customer’s needs, consistent delivery, and customer service the platform

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 91 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

SortSpoke (page 1 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 SortSpoke leverages proprietary and patented neural network-based machine learning technology  The geographic presence of SortSpoke is currently limited to only North America. Hence, its ability for classification and extraction of semi-structured and unstructured documents. It aims to provide a to serve clients in other regions remains untested

self-serve IDP solution that is easy to implement, maintain, and integrate – for use cases in mid-  It primarily caters to clients in BFSI and professional services verticals. Its experience of serving market and enterprise operations clients in other industries such as healthcare and pharma, CPG & retail, manufacturing, and high-  The solution has flexibility to use Omnipage OCR, Azure OCR, Google Vision, and AWS Textract tech and telecom is low

for OCR capabilities, as per customer request  The solution allows document ingestion only in PDF format, limiting its flexibility for ingestion in  The software can learn to extract data from documents in any Latin-based language by uploading other commonly available formats such as JPG, PNG, TXT, XLS, and CSV

the documents and letting business users label the data using point-and-click interface  While it can extract data from tables, free-flowing text, handwritten text, and checkboxes, it lacks  Supervised machine learning allows it to automatically learn from human-in-the-loop corrections. It the ability to process bar codes, logos, stamps (non-text), and signatures

continues to learn in near real-time as every document is processed, allowing it to deal with new  The solution currently lacks certain NLP capabilities such as sentiment recognition and text variations and new fields without IT involvement summarization  SortSpoke can automatically classify and split a multiple-page PDF that contains multiple document  It lacks pre-built connectors with leading enterprise systems such as SAP, Oracle, and Microsoft classes Dynamics, which could limit its ability to provide faster time-to-value and shorter implementation time. Integration with third-party BI platform providers is currently in its roadmap

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 92 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

SortSpoke (page 2 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 SortSpoke utilizes deskew, contrast enhancement, noise removal, as well as line removal for pre-  Reporting data is currently available in a spreadsheet format. A self-service dashboard for processing documents and enhancing image quality to achieve higher OCR accuracy performance metrics is in the roadmap

 It allows business users to define workflows for review, add new users, and assign tasks for  The solution lacks capability to perform business rules validation through external database lookup

document review, while facilitating role-based access control  Auto-redaction and manual blurring of sensitive information is currently unavailable and is in the  It can perform entity recognition for extraction of data from unstructured documents roadmap

 Business users can create custom rules or select any pre-defined rule-sets (such as date and  The solution can be hosted only on cloud. Thus, it will be an unsuitable option for clients looking currency formats) for validation of extracted data specifically for an on-premise solution

 The platform has pre-built connectors with leading RPA and BPM vendors for easy integration with  As per client feedback, SortSpoke can invest more in proactively conveying its product vision and existing systems and workflows. It is also looking to further expand its network of complementary roadmap to clients

technology partners  Clients feel that there is scope to improve its ability to extract handwritten text, its reporting and  Clients appreciate the intuitiveness of the interface, responsiveness to client needs, understanding analytics capabilities, as well as support for additional languages (such as Asian languages). of use cases, and effectiveness of the ML engine Clients also suggest that providing quality scores (on extraction accuracy) through the reporting interface would be helpful

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 93 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

TAIGER (page 1 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 TAIGER is a Singapore-based provider of AI solutions. It leverages semantic and NLP capabilities  TAIGER is a relatively new vendor in the IDP space and currently has few IDP clients. Therefore, for extracting information from unstructured documents. Its portfolio of offerings includes Omnitive it is yet to establish itself as a vendor of choice in the IDP market

Extract, its IDP solution, in addition to a semantic search engine and a virtual assistant  It has predominantly served small and SMB segment enterprises in Continental Europe and APAC  Omnitive studio provides pre-built models out-of-the-box for extracting data from documents such as regions. Hence, the vendor’s abilities to cater to requirements of a global enterprise is not proven passports, identity cards, payslips, and bank statements. It also allows users to train new models or yet

improve existing ones based on new data and can classify uploaded documents into different  Its client portfolio consists of enterprises from the banking sector and the solution has been document types deployed for banking industry-specific use cases only. The platform’s capabilities to serve  Users can ingest documents in various formats such as TIFF, JPG, JPEG, PNG, BMP, DOC, enterprises in other industries and use cases is yet to be proven

DOCX, and TXT through the UI itself for processing. The extracted data is highlighted in different  While it can detect signatures, verifying them against a specimen sample is part of the roadmap. It colors to indicate the confidence level of extraction or a validation rule also does not support extraction of objects such as logos and bar codes and has limited  In addition to extracting data from structured and semi-structured documents, Omnitive Studio can capabilities for processing handwritten documents

also process and extract information from unstructured documents with free-flowing text by  ML models for processing of documents and data extraction are pre-configured at the starting of leveraging NLP capabilities projects. Auto-selecting the best model for a given document type and use case based on benchmarking of results is part of the roadmap

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 94 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

TAIGER (page 2 of 2) Everest Group assessment – Aspirant

Measure of capability: High Low

Strengths Limitations

 It augments the extraction and classification capabilities of IDP with NLP to search and extract data  It does not provide an analytics dashboard and presently only generates performance reports from documents and also perform sentiment analysis indicating process-level SLAs and field-level accuracy. Offering analytics dashboard out-of-the-box

 The platform supports English, Spanish, and Russian languages for processing documents, and is part of the roadmap for 2021 adding Japanese is part of the roadmap. The UI of the platform is available in English and Spanish  TAIGER has very few service provider partners for providing implementation and training support. languages It also does not partner with technology vendors for complementary technologies and does not

 It allows users to add or configure a new field to be extracted from the document. Users can also have pre-built connectors for downstream systems configure the threshold confidence level for extraction of fields and add new document types  Owing to its small client base, it provides customer support based on client requirements and in a

 The platform is developed on microservices architecture and supports multi-tenant, containerized tiered manner. It does not provide 24/7 support to all clients, and does not have an online user deployments. It can be deployed on-premise and on private cloud community forum either   It offers a subscription-based commercial model where customers are charged based on usage of Referenced clients have stated that providing greater clarity in technical requirements for product pages or documents. Typically, the cost involves a base price and enterprises are charged for integration could lead to smoother implementation and integration of the product including add-ons such as the virtual assistant. TAIGER also guarantees accuracy level as part of the SLA depending on the document types

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 96 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

IDP products capabilities, features, and functionalities (page 1 of 4)

Improve low-quality images and handwritten documents with features such as auto-crop, noise removal, background editor, and the ability to Documents pre-processing configure custom enhancement criteria

Refers to automatic classification and sorting of incoming documents and the ability to route them to desired destinations. It also identifies Document classification different sections within the document (e.g., faxes, letters, or invoices) and classifies them according to categories before extraction

Pre-trained out-of-the box solutions with reasonable accuracy (~70-80%). These are generally trained by ingesting a variety of documents for a Pre-trained models particular domain or use cases into ML models, which users can further train to match their own needs

Accelerated learning by ingesting a labeled set of sample documents or Excel files with sample data in order to efficiently classify documents Supervised learning and extract the relevant fields

 Learn through corrections performed by business users during manual review by automatically generating training batches in the background User-based learning  Ability for business users to train the software by defining the criteria through anchors, associations, expressions, and lookups by uploading a set of keywords

Leveraging the ML models developed by various enterprises for use cases to accelerate the training process. Cross-training is generally Cross-training performed on specific use cases depending on the willingness of enterprises to share their trained models

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 97 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

IDP products capabilities, features, and functionalities (page 2 of 4)

Complementary Tighter integration with complementary solutions including RPA and BPM solutions, either through in-house capabilities or via partners, to technologies enable end-to-end process automation for enterprises

Process documents in multiple languages apart from English such as European, Latin American, Japanese, Chinese, and Arabic languages. Multi-lingual documents Some leading vendors have built the capability to process documents in 100+ languages

Processing Classify and extract unstructured data fields from emails, documents with free-flowing text, images, handwritten documents, etc. unstructured data

Signature extraction Refers to extraction and matching of signature within a document, along with displaying the confidence-level of the match. Particularly useful for and matching KYC-related use cases

Developing mobile applications and websites to process data from images captured through devices such as mobiles and tabs. Users can Mobile capture directly upload the captured image of the document into the app for processing

 Availability of APIs for faster integration with enterprise systems and applications Integration with  Pre-built connectors with leading enterprise applications enterprise IT architecture  Support for integration with upstream and downstream applications of enterprises such as ERP, CRM, and HCM

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 98 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

IDP products capabilities, features, and functionalities (page 3 of 4)

 Modify and customize the confidence level or accuracy thresholds for classification and extraction of desired fields from documents Configuration and set-up GUI  Add new processes or use cases from a set of available domains for enterprise administrators  Ability to define, add, and modify the fields to be extracted  Upload the documents by batches and manage the processing  Define and manage user access controls  Visual tool to adjust the accuracy threshold for documents or extracted fields by evaluating the criticality of processes and the desired automation rates  Add or modify business validation rules such as performing mathematical calculations on extracted fields and looking up data from external databases for verification

 Categorize documents into different types and display the confidence level for classification or highlight the categorization below the Review or correction GUI defined accuracy thresholds for business users  Display the confidence level of extracted fields or highlight the extracted fields that failed to meet the defined confidence or accuracy thresholds  Highlight the fields that violate the business rules or fields with incorrect/missing data  Ability for business users to manage the work queue of processed documents  Display error messages with user-friendly explanations for the extracted fields that require manual review

 Flexibility to experiment on the best-fit ML algorithms for use cases Workbench for  Workbench for enterprise IT users to manage the workflow of processes, access to machine learning libraries, and integrated enterprise IT users RPA capabilities

Ability to review and approve the learning data sets that are automatically generated from user-based training Learning queue administration

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 99 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

IDP products capabilities, features, and functionalities (page 4 of 4)

Analyze the running text in documents, understand the context, consolidate the extracted data, and map the extracted fields to a defined Natural language taxonomy. Further, to also recognize the sentiments from the text (e.g., from emails and other unstructured data) and classify into processing different categories

Searching a keyword across a repository of scanned images and PDF documents for shortlisting. Useful in use cases where organizations need Searching through to search for details in documents to analyze a trend or group them together unstructured documents

 Analytics dashboard provides a unified view of multiple document processing projects and allows tracking of various parameters such as STP Analytics dashboard rate, process-level SLAs, batch-level & field-level processing accuracy, manual worker performance, number of errors fixed, and time taken to fix the errors  Ability to drill down the metrics by specific classes of documents and reviewers  Consolidated view of processed documents and fields that contain exceptions to help in identifying the root causes of lower processing accuracy rates

Security features of IDP solutions include the ability to encrypt, hide, or redact confidential data fields using various technologies before review Security and and adherence to enterprise IT security standards & regulatory compliance requirements data protection

IDP solutions can be deployed on cloud, on-premise servers, and on desktops. Cloud is the most widely adopted deployment mode, whereas Hosting on-premise and desktop deployment could be considered by industries such as BFSI and healthcare in use cases with stringent data security and compliance requirements

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 100 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021 What do the latest developments in IDP products mean for enterprises?

Increased configurability of platforms to provide greater control for enterprise users

SaaS offering of the solution to lower TCO for Enhanced integration with complementary technology solutions enterprises and increase accessibility including RPA, BPM, and process mining

Vertical-/horizontal-specific pre-trained solutions out-of-the- Dedicated mobile applications to facilitate document processing box; app store-like channels for easy access through handheld devices

Increasing maturity for processing Advanced image recognition and processing capabilities using a unstructured documents combination of computer vision and deep learning algorithms

Extraction capability for a large set of languages Availability of benchmarking analytics for particular including Asian and Middle Eastern languages processes such as invoice processing

 Glossary 08  Research calendar

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 102 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Glossary of key terms used in this report (page 1 of 3)

Artificial Intelligence (AI) Ability of machines to use cognitive computing to mimic human intelligence, such as visual perception, speech recognition, decision-making, and language translation Business Intelligence (BI) Technologies, applications, and practices for collection, integration, analysis, and presentation of business information Business Process BPM solutions help to coordinate tasks and orchestrate the flow of information across disparately designed applications, databases, digital workers, and the human Management (BPM) workforce. It includes capabilities of process design, execution (through workflows and orchestration of different BPS technology systems), and monitoring (through analytics) Buyer The company/entity that purchases outsourcing services from a provider of such services Classic process mining Classic process mining refers to the ability to leverage specialized algorithms to analyze process-related information that is captured in event logs generated by enterprise systems such as ERP, CRM, and SCM, to discover as-is processes, generate process maps, perform conformance check with pre-defined input reference process models, and generate insights for process improvement Desktop Process Mining The ability to capture user’s keyboard, mouse, and potentially other system-level activities performed simultaneously on various desktops to virtually reconstruct the (DPM) / Task Mining processes and generate a process map capturing the different process variants Cognitive/smart The ability of a system to learn how to interpret unstructured content, such as natural language, and use analytical capability to derive and present inferences in a automation pre-defined/structured fashion; for example, a system classifying the mood of a person into one of the pre-defined groups based on his/her tone and language Computer vision A technology that uses AI to enable automatic extraction, analysis, and understanding of useful information from digital images Deep learning A subfield of machine learning concerned with algorithms and inspired by the structure and function of the brain called artificial neural networks FTE A way to measure a worker's productivity and/or involvement in a project. An FTE of 1.0 is equivalent to a full-time worker General AI A machine that can perform multiple intellectual tasks across a variety of domains; essentially, it mimics all activities performed by a human Horizontal business Those processes that are common across the various departments in an organization and are often not directly related to the key revenue-earning business, such as processes procurement, finance & accounting, and human resource management IDP Intelligent Document Processing is a software product or solution that captures data from documents (e.g., email, text, PDF, and scanned documents), categorizes, and extracts relevant data for further processing using AI technologies such as computer vision, OCR, Natural Language Processing (NLP), and machine/deep learning

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 103 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Glossary of key terms used in this report (page 2 of 3)

KPI Key performance indicators for processes, services, products, or solutions Machine Learning (ML) A type of artificial intelligence that provides computers with learning capabilities without explicit programming Narrow AI A machine that performs one narrow task; an expert system Natural Language A machine’s ability to interpret human languages Processing (NLP) Natural Language The software process to write text in human languages based on structured data Generation (NLG) Optical Character A technology within computer vision that involves the recognition of printed characters using computer software Recognition (OCR) POC A realization of a certain method or idea in order to demonstrate its feasibility, or a demonstration in principle with the aim of verifying that some concept or theory has practical potential ROI Returns attained from an investment RPA RPA refers to a type of rules-based automation technology that helps automate repetitive tasks by mimicking a user’s activities. It is non-invasive and typically interacts with a computer-centric task/process through the User Interface (UI) of the underlying software applications Semi-structured data Semi-structured content is one that does not conform to the pre-defined structure of content, but nonetheless, contains tags / other markers to separate semantic elements and enforce hierarchies. In short, it has a self-describing structure. The placeholders of the content can be in varied sequences Software-as-a-Service SaaS is a software licensing and delivery model wherein the software is hosted centrally by a third-party provider and is made available to customers over the internet. (SaaS) It is also referred to as on-demand software Structured data Structured content is one that conforms to the pre-defined structure in terms of tags to separate semantic elements and enforce hierarchies of records and fields. Moreover, the placeholders for the content have a pre-defined sequence Transaction-based pricing Output-based pricing structure; priced per unit transaction with significant price differences between onshore and offshore

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 104 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Glossary of key terms used in this report (page 3 of 3)

Usage-based pricing Value-based pricing structure; pricing based on per-page or per-document processed Unstructured data Unstructured content refers to information that either does not have a pre-defined data model or is not organized in a pre-defined manner. Unstructured information is typically text-heavy, but may contain data such as dates, numbers, and facts as well Vertical-specific Vertical-specific business processes refer to processes that are specific to a department within an organization and are often directly related to the key revenue- business processes earning business. Examples include lending process in case of the banking industry and claims processing in case of the insurance industry Virtual agent It is a computer-generated virtual character that can have a conversation with human customers and take decisions. Alternative term for chatbots or virtual assistants

Proprietary & Confidential. © 2021, Everest Global, Inc. | This document has been licensed for exclusive use and distribution by IBM 105 Intelligent Document Processing (IDP) – Technology Vendor Landscape with Products PEAK Matrix® Assessment 2021

Everest Group is a consulting and research firm focused on strategic IT, business services, engineering services, and sourcing. Our clients include leading global enterprises, service providers, and investors. Through our research-informed insights and deep experience, we guide clients in their journeys to achieve heightened operational and financial performance, accelerated value delivery, and high-impact business outcomes. Details and in-depth content are available at everestgrp.com.

Stay connected

Website Dallas (Headquarters) London everestgrp.com [email protected] [email protected] +1-214-451-3000 +44-207-129-1318 Social Media @EverestGroup Bangalore New York [email protected] [email protected] @Everest Group +91-80-61463500 +1-646-805-4000 @Everest Group @Everest Group Delhi Toronto [email protected] [email protected] Blog +91-124-496-1000 +1-647-557-3475 everestgrp.com/blog

Podcast DigitalRealITy

This document is for informational purposes only, and it is being provided “as is” and “as available” without any warranty of any kind, including any warranties of completeness, adequacy, or fitness for a particular purpose. Everest Group is not a legal or investment adviser; the contents of this document should not be construed as legal, tax, or investment advice. This document should not be used as a substitute for consultation with professional advisors, and Everest Group disclaims liability for any actions or decisions not to act that are taken as a result of any material in this publication.