<<

THE ATTUNE

A New ADS for Apps that Program Themselves with AI

Abstract HFP Technologies LLC (HFP) has successfully harnessed AI technology to develop and execute software applications (apps), supplanting ~95% of traditional instructions. The Attune System is a first-of-its-kind Application Development System (ADS) incorporating AI technology as its foundation to make easier, faster, and less expensive. The Attune System operates as a next generation, low code/no code, high performance application platform (hpaP) for end-user AI- powered enterprise apps that we call Smartware. The Attune System demonstrates that ~95% of source code, as ubiquitously employed today in software engineering, is no longer required to be written by humans (or generated and compiled) in complex end-user apps. Instead, the HFP AI- enhanced ADS represents Smartware processing instructions as reusable, loosely-coupled, data-driven functional specifications (specs) stored in a database. Specs are linked at run-time to an AI engine, running as a service, that operates to find, match, and interpret data-driven specs as machine language virtual programs for processing. A Smartware app essentially programs itself at the business logic (process) level as it is used. Complex businesses embracing cloud-based digital transformation and seeking very low- cost, apps for internal and customer-focused operations can be efficiently managed with configurable enterprise Smartware modules created with the Attune System ADS. With enhancements, this ADS can operate as a next generation AI-enhanced, low code/no code hpaPaaS.

The Attune System is a technology asset of HFP Technologies, LLC 555 North Point Center E, Suite 400 Alpharetta, GA 30022 www.hfptechnologies.com [email protected] 404.271.2240

HFP Technologies LLC Atlanta, GA

1

THE ATTUNE SYSTEM A New ADS for Software Apps that Program Themselves with AI

Overview Definitions We define an Application Development System (ADS) (sometimes Attune System: referred to as a platform, framework, environment, or software system) as a comprehensive, integrated set of tools and utilities with a common interface for developing, testing, modifying, and executing end-user software applications.

The Attune System represents the first-of-its-kind ADS, enhanced Smartware: with artificial intelligence (AI) technology, that creates and executes end-user AI-powered enterprise apps (Smartware). This application development invention is based on a services-oriented architecture (SOA) that supports end-user requirements in complex, networked/cloud-focused businesses. The Attune Functional Specification: System has eliminated the need for representing business logic as written or generated source code through creation of smart processing. Smart processing involves run-time linking of two basic ingredients: 1) a reusable, loosely-coupled, data-driven functional specification containing algorithmic processing logic (data patterns that an AI engine understands), and 2) an AI engine Algorithmic Processing (“Harrison”) running as a service that interprets it. When Logic: Smartware processing is required, “Harrison” finds and interprets the appropriate server-side functional specification stored in a knowledge base and generates a run-time machine language virtual program that enables processing. Functional specifications are an estimated order of magnitude more efficient and flexible than source code in representing or modifying business logic processing instructions. Smart Processing: Smartware supported by smart processing represents a new generation of nimble, Smartware-as-a-service (SMaaS). Smartware can enhance the transformation of any networked/cloud-based digital enterprise. Smartware modules can be configured to blend internal business applications with e-commerce capabilities to seamlessly manage the complete integrated enterprise of any complex business from any device in a single, real-time program. Smartware is quickly assembled, low cost, network optimized, and can efficiently adapt to changing customer needs, new business opportunities, evolving technologies, and new business processes.

2018 © HFP Technologies, LLC / All Rights Reserved.

2

Smartware looks and performs like traditional software but differs in two major respects: 1) prepackaged source code (written or generated) is not needed to represent business logic, and 2) “Harrison,” an AI engine operating as a service, creates run-time machine language virtual programs to facilitate processing. The Attune System’s ability to represent business requirements as structured, reusable, meta- language functional specifications and use AI technology to streamline information processing has important development business advantages including: 1) faster delivery and dramatic cost reduction of software development and change management, 2) increased reusability or easy modification of data-driven business requirements that support a continuous delivery model, 3) creation and storage of domain expertise in functional specifications for future use in information processing, and 4) ability to employ efficient and adaptable Smartware and smart processing in complex enterprises with high volume internal business and customer- focused requirements (including e-commerce). The Attune System invention achieves four previously unattained milestones in enterprise apps: 1) creation of a formalized generic model of knowledge representation of industry entities and relationships and structuring processing instructions in a reusable data pattern that can be interpreted by an AI engine (FUNCTIONAL SPECIFICATION); 2) creation of a practical process (satisfying a DevOps approach) that transforms any business into Smartware (METHODOLOGY); 3) invention of a calculus-based AI engine that operates as a service and transforms algorithmic processing logic in functional specifications into run-time machine language virtual program processing instructions (AI TECHNOLOGY); and 4) assembly of all traditional and AI-based utilities for Smartware development, configuration, execution, and change management (including an AI engine running as a service) into a general-purpose, integrated ADS (ATTUNE SYSTEM). The Rationale for Smart Processing A complex domain or industry can have thousands of data elements, conditions, and events in its business process cycle. These conditions create potentially millions of processing and workflow scenarios to support in end-user applications. A complex monolithic application could have millions of lines of source code to build/generate and support. Despite the resources invested in writing source code, it is seldom possible to efficiently reuse source code in a new , thus adding more time and cost to new . Project failure is not an uncommon outcome when using traditional approaches to build complex applications. Traditional approaches relying on source code involve many skilled workers, high costs, and long development times, despite industry efforts to implement methods (e.g., Agile, DevOps, micro-services, etc.) that claim to streamline this approach.

2018 © HFP Technologies, LLC / All Rights Reserved.

3

Improved methods and systems for digital data processing have long been sought with the desired outcomes of high quality, reusability, rapid completion, web device accessibility, flexibility, adaptability, and low total cost of ownership of end-user applications. Conceptually, a smart processing approach to apps, using run-time AI technology to supplant repetitive business logic traditionally performed by humans, has high potential to meet these goals. In practice, however, this approach has faced the absence of a coordinated, all-in-one ADS for capturing and using business specs (functions, processes, formulas) that can be interpreted and executed by an AI engine in a way that facilitates information processing in large-scale end-user enterprise apps. The Attune System offers a practical solution to this challenge with its general purpose, coordinated, smart processing approach that supports the entire software development life cycle. Reusable Expertise + AI Technology = Smart Processing Software at its basic level allows a computer to automate work performed by a human expert in a specific domain. End-user software is fundamentally a set of written instructions that capture aspects of domain expertise stored in the brain of human experts that can be transformed (compiled/interpreted) so that a microprocessor can perform work based on that expertise. Capturing domain-specific human expertise that can be understood by a computer is at the heart of the software process. Traditionally, capturing this expertise is done with source code using various high-level programming languages and extensive business process analysis. Subject matter experts have a significant amount of built-in knowledge about how a specific domain or industry is organized and how work is performed. Examples of topics that qualify as expertise that might be known by a human expert include the definition of an industry-specific event or instance of activity, industry-pertinent formulas and algorithms, and identification of the sequence of events in an industry. An event is a domain-specific work product (e.g., a payment, an order, a report) that can be automated in application software. An industry or domain can be complex, compounding the challenge of capturing events quickly and inexpensively in a reusable fashion, especially when relying on source code. A critical element of a smart processing approach is the requirement that domain expertise about events be represented and stored as a type of reusable non-source-code processing instruction such that an intelligent expert (AI engine/algorithm) can interpret it directly to complete processing. Fulfilling this requirement calls for solving the industry challenge of finding a reproducible, reliable way to represent and capture these AI-compatible instructions.

2018 © HFP Technologies, LLC / All Rights Reserved.

4

Information processing instructions about events can be expressed as indexed models of sequenced workflows (FUNCTIONAL SPECIFICATION). The Attune System supports an ontological approach to domain expertise that can be applied to any industry. This approach establishes a structured system of categories (a model) that represent entities, ideas, and events together with their properties and relationships. The Attune System METHODOLOGY systematically captures the expertise associated with an event as data-driven functional specifications and stores them in a reusable format as data. Functional specifications are composed of workflows and knowledge segments. Knowledge segments are represented as meta-language that includes English-like knowledge statements (representations of the subject matter) plus SQL statements. Knowledge statements are the algorithmic processing logic whose data patterns can be used by an AI engine. Smartware functional specifications can be modified quickly to enhance existing Smartware or reused with simple changes to create new Smartware. This capability to systematically create, reuse, and modify functional specifications in information processing that can be interpreted by AI TECHNOLOGY is an Attune System breakthrough. The Attune System AI engine contains a unique algorithm capable of recognizing all arithmetic and logical operations in any Smartware functional specification and interprets them as machine language for information processing. Functional specifications together with the AI engine constitute smart processing. They are operationally linked at run-time (see below). Smartware can operate as a simple end-user module or support a complex business or enterprise as a cloud service. This approach significantly reduces the time and cost of development and change management in the systems development life cycle and enables enterprises to capture their domain knowledge and archive it in a knowledge base for future use in information processing. Attune System Key Characteristics 1. The Attune System is a general-purpose, multi-component ADS that can create, modify, and execute Smartware applications. The Attune System operates as an ADS and includes a set or stack of separate tools and utilities. 2. The Attune System employs a SOA where functional specifications can be menu-configured into Smartware modules. Multiple Smartware modules can support a complex enterprise within the ADS. 3. The Attune System creates Smartware. • The Attune System uses a highly structured, software-enabled methodology for transforming business requirements/processes (traditionally found in written form) into fully operational Smartware. The methodology addresses all fundamental components of creating software usually addressed by business analysts and : a user

2018 © HFP Technologies, LLC / All Rights Reserved.

5

interface (UI), database, security, connectivity, and processing instructions (Figure 1, p. 8).

• Processing instructions are a key differentiator of the Attune System methodology. Processing instructions are contained in data-driven functional specifications. A functional specification is composed of multiple expert processes containing workflows and knowledge segments (containing algorithmic processing instructions and SQL statements). These instructions are written in data, not source code, and stored in a database. Algorithmic processing instructions are interpreted at run-time by the AI engine (see below) to generate a machine language virtual program.

• The Attune System Knowledge Manager is a work bench and library to build or modify functional specification expert processes. These processes are based on a model that converts business requirements into functional specifications that can be used directly in information processing and be reused or modified. In traditional programming, the written specification is a detailed document and is usually the starting point for writing source code. An Attune System functional specification has non-traditional business logic in it. Practically speaking, in the Attune System, the “spec is the code.” This is a far cry from traditional practices where specs and code are very different entities. 4. The Attune System executes Smartware with AI technology: At the heart of Smartware processing is “Harrison,” a calculus-based intelligent agent based on narrow AI principles. When information processing is required within a Smartware user interface, an API message containing data and a functional specification reference is generated within the user interface and sent to the AI engine running as a service. When received, “Harrison” searches the knowledgebase for the appropriate functional specification, makes inferences, completes pattern matches, interprets the data patterns of the algorithmic processing instructions, creates a run-time machine language virtual program to complete the processing requirements, communicates with and updates the database, and returns a result audit to the user interface. “Harrison” is capable of supporting machine learning with utilities such as neural networks (Figure 2, p. 8).

2018 © HFP Technologies, LLC / All Rights Reserved.

6

5. As Smartware is developed by following the methodology, its expert processes are derived from libraries and work benches of reusable resources. • The Database Manager is based on a flexible and has a data dictionary containing thousands of reusable data elements. • The Central Directory organizes all persons, places, and things and the security levels for Smartware users. • The Algorithm Library is a database containing >1000 text-based, reusable and modifiable algorithmic processing logic files that perform many different functions and supplant the need for source code to represent business logic in Smartware. Because they are expressed as data, they can be copied and modified for reuse in new Smartware. • The User Interface Manager creates and manages Smartware user interfaces. • The Connectivity Manager supports connectivity with outside systems and databases (APIs) and makes the Attune System IoT compatible. It can accommodate standard interface engines including HL7, SQL access and messaging, XML socket access messaging, XML input/output files (ebXML), and ASCII and ANSI transactions. 6. The Attune System is equipped with configurable, ready-to-use Smartware modules. • Multiple configurable Smartware modules sharing the common Attune System are available in a library to support complex enterprises. Loosely coupled Smartware modules allow them to be quickly configured in any combination to support complex cloud applications and scale to any user need.

Value and Benefits The Attune System, a systems development invention, is equipped with a methodology to capture any business requirement as a reusable, loosely-coupled, data-driven functional specification containing algorithmic processing instructions. This system represents the first integrated (all-in-one), practical approach to smart processing that combines functional specifications with AI technology to supplant traditional source code for processing instructions. The smart processing approach supplants as much as 95% of traditional source code instructions previously required in complex software applications. This approach streamlines the creation, testing, execution, and maintenance of enterprise-level SMaaS for any complex business. The Attune System’s principal value and benefits include:

2018 © HFP Technologies, LLC / All Rights Reserved.

7

Savings in time and money: As a low code/no code approach with many common expert processes included in the Attune System library, creating new Smartware is both efficient and fast compared to traditional systems development approaches based on source code. Existing Smartware modules can also serve as a template for creating new ones through modification of functional specifications, not writing, rewriting, or generating source code. Rapid change management: In most complex businesses, frequent updates to software applications are required to keep up with changing business opportunities, customer demands, and technology advances. Often, change management is significantly delayed and costly because of the need to modify source code. Attune System Smartware updates require only modest changes to data in a functional specification. This advancement permits Smartware updates in near-real-time with rapid testing and near-instantaneous redeployment to support a continuous delivery model. Summary There is great interest and ample opportunity to leverage AI technology in software engineering in 2018. Much emphasis is currently placed on embedding AI services and functions into software applications to make work activities easier for users. In contrast, HFP has successfully employed AI technology to build and operate software applications themselves and make the task of software engineering itself easier, faster, and less expensive. With the Attune System ADS’ integrated utilities (including its AI engine running as a service), we demonstrate that ~95% of source code, as ubiquitously employed today in software engineering, is no longer required to be written by humans (or generated and compiled) in end-user software applications. Knowledge about domain events and processes can be captured and stored as reusable smart data in a functional specification. The Attune System’s generic model for representing domain expertise for information processing without source code can be applied to any industry. Smartware exploits an for representing business requirements as reusable functional specifications containing algorithmic processing instructions that are used by an AI engine for information processing. “Harrison,” HFP’s AI engine, simulates a , operates as a service, and generates run-time machine language virtual program processing instructions. An AI-powered Smartware app essentially programs itself at the business logic (process) level as it is used. This new ADS, (incorporating breakthroughs in knowledge representation, methodology, and AI technology) yields complex, continuously delivered, end-user software applications that are efficient, agile, and dramatically less expensive to develop and manage than are available today. Attune System Smartware apps are positioned to accelerate enterprise transformation in the digital economy while significantly lowering the total cost of ownership.

2018 © HFP Technologies, LLC / All Rights Reserved.

8

Figure 1: Attune System Smartware Methodology

• Knowledge Manager • Knowledge Templates Smartware Knowledge • Functional Specifications Methodology Library • Algorithm Library (processing logic)

• Data Dictionary • Connectivity Manager Smartware Database • Central Directory Library • Database Tables • Data Loading

• UI Manager • Display, Menus, & Presentation Branding • Dashboards • Security Management

Figure 2: Attune System AI engine Running as a Service (AI Service)

• Data Inputs ✓ Files • Updates Results to ✓ Messages • Sends API Message Database ✓ UI • Sends Audit Trail to User

• AI Service • AI Service Socket Out AI Engine Socket In

• Identifies Unique • Receives & Parses Functional Specification API Message • Interprets Processing • Calls AI Engine Logic • Creates Virtual Program • AI Engine in Memory

2018 © HFP Technologies, LLC / All Rights Reserved.