06/2019 cView

Artificial Intelligence in Case Management F2 Version 6.2

In this issue Artificial intelligence in the public sector New F2 version 6.2 with emphasis on automatisation The basic principles of artificial intelligence AI case management

F2 version 6.2 emphasises

Artificial intelligence in case management

Though the public sector is in- to its missing emphasis on the rule-based management can be supported when vesting heavily in artificial intel- type, which is exactly what is needed for relevant. The model also facilitates the im- ligence, many projects are still the automating of case management. plementation of artificial intelligence, so it in their research phase. Many Instead, recognition-based techniques happens as part of a controlled and agile solutions are characterised by are tested and implemented, and while transformation. supporting simple, stand-alone they may be appropriate, they are mostly work routines. The enthusiasm is used for point solutions or in relation to F2 as a platform allows for great impro- present, but so far results have estimation and counselling. vements, as the model ensures that the been modest. administrative tasks are the starting point. Integrated artificial intelligence F2 also facilitates an ongoing optimisa- Artificial intelligence offers a set of IT Based on a model of digital bureaucracy, tion of the authority while experience tools that provides great opportunities, F2 offers a new and effective approach to is accumulated and new techniques are especially in relation to the public sector. artificial intelligence. The model ensures implemented. Artificial intelligence, then, Already today there are great improve- an effective integration and supports both becomes not only an independent science ments to be made, because a large part types of artificial intelligence as well as project, but a natural part of the authori- of administrative case management is agile implementation. ty’s digital transformation and evolution rule-based. Coincidentally, the rule-based in both the short and the long run. type of artificial intelligence is the most The model helps describe an authority’s advanced. work flow in detail, down to case types, In this issue of cView, we present F2 versi- tasks and roles. With a basis in this de- on 6.2 with a special emphasis on artificial Artificial intelligence means “mimicking scription case processes and tasks can be intelligence. We will tell you about the human cognition”. There are several automated by implementing artificial in- new functions in F2 version 6.2, including kinds of human thinking, however. telligence directly in the flow and on each a number of fully integrated robots, and Today we are working on mimicking two individual step. F2 ensures an effective we will tell you about the theory, model basic ways of thinking: integration and supports and utilises both and techniques which make F2 a unique types of artificial intelligence as a fully platform for artificial intelligence in relati- • Rule-based thinking, .g. used for com- integrated part of the standard system on to case management. puter chess and in both the authority’s processes and • Recognition-based thinking, e.g. speech case management. or image recognition The many case areas which are exclusi- The public sector’s limited success with vely rule-based can then be fully auto- artificial intelligence so far may be due mated, and estimations during the case

2 | cView – AI case management F2 version 6.2

F2 version 6.2 emphasises Automatisation

In F2 version 6.2 cBrain has added several • Robot case creation • Robot for hearings involving external new functions and options to the standard • Automated tasks for case steps participants product, not least when it comes to • Robot for producing documents • Mass processing of cases artificial intelligence and automatisation. governed by a rule- and context-based • Robot for case monitoring Support for both rule-based and recog- phrase archive nition-based artificial intelligence is fully • Support for rulings In addition, new functions and improve- integrated in F2 version 6.2, making it an • Robot for case guide and variance ments have been implemented across the especially effective platform for automa- processing product, several of which are inspired by ting work tasks and case management. • Robot for rule-based self-service user suggestions and feedback including - The ability to delete a case, e.g. to meet data protection regulations.

Robot for case Automated tasks creation for case steps

Central to every case process is Based on the “digital bureaucracy” As an example F2 may automatically the case creation. But creating a model, F2 enables the creation of case offer information on parts of the case or case can be laborious and quality guides which determine the steps and verify reported case information through assurance difficult. F2 solves this tasks that a given case type must go the use of rules and lookup functions. using automated case creation, a through. When a case is created, it is context-sensitive robot that supports possible to attach a case guide from the Automating tasks saves time and helps fast and secure case creation. case guide library. increase quality. Additionally, and crucially, it ensures compliance with The robot is based on a library Each step comprises a number of tasks. administrative requirements. Basing of case templates. When the user Automation of these tasks is possible the automatisation on a case guide creates a case, a template can be since rules can be attached to a step, ensures that even an entirely automated attached based on which the robot and each step may call on external case process is fully documented as if performs a number of tasks in rela- functions. An entire task or parts the case were handled manually. The tion to e.g. attachment of security thereof may be completed by calling on documentation covers both the overall groups, participants, keywords or a external services. A recognition routine process as well as each step and its given case guide. which offers counselling on estimations tasks, decisions and rulings. based on available data may also be With case templates, case creation activated. This enables F2 to perform is faster and many errors are elimi- many of the work tasks which the case nated. The robot may be activated manager before performed manually. by a user or by F2. For example, in relation to self-service the robot may be activated on a given case step Case guide model after case creation, fully automating the process. A case guide consists of a number Case guide of steps, each with a number of tasks to be performed. Many tasks can be automated Casesteps through rules which are applied to a given step. The tasks may be performed by the authority or Tasks by a citizen or another external participant.

cView – AI case management | 3 F2 version 6.2 – continued

Robot for Robot for case guide producing docu- and variance processing ments governed The F2 standard product allows for the In this relation, the robot also supports by a rule- and definition and storage of case guides. a so-called “navigation function”. If context-based In this context variance is central. navigation is present on a given step of It is possible to define variance in a case or record, the robot determines phrase archive relation to the case guide itself, e.g. what is presented to the user. For the which steps to follow, and in relation user, this means that clicking on a to automating tasks performed on the record in a given list in F2 opens those A central element of case mana- individual steps of the case process. tasks that must be performed in order gement is the production of notes, to progress to the next step. It is a emails, letters and other documents. This means that rules may be attached means of guiding the user’s work using F2 offers a robot which allows for which govern a concrete case guide, lists, or bookcases, made possible by large parts of document production and the case guide itself may be redefi- the F2 robot that governs the overall to be automated through an ad- ned with additional case steps if chan- execution of the case. vanced rule-based phrase archive. ges are made to the case’s information. When the robot is activated, it Variance may also be governed reviews the relevant case informa- A case that progresses through the steps by checklists. If the case is missing tion and relevant templates and of a case guide will often be represen- information, for example, the F2 phrases. The phrase archive is ted by lists that are presented to the robot can add an additional step to comprised of rules with full access user in F2’s main window. The user will ensure that the case is fully examined to case information and is able experience the robot for case guide and before proceeding. Through rules and to utilise templates for e.g. Word variance through the case’s progress variance, the robot can also determine documents or emails. from list to list, which can be thought of which further case steps are relevant in as a relocation from one bookcase to relation to a concrete ruling. This means that F2 can generate another. even complex documents automati- cally. It also allows for automatisati- on of even the complex steps of the case process.

Example: When a new case is received from a self-service site, the case creation robot ensures that the Robot for rule-based case is created correctly and atta- self-service ches a case guide. As a first step of the case guide, the document pro- ducing robot generates a document Case guides in F2 can support the The rule-based self-service is gover- or an email based on the informati- entire case process from self-service ned by a robot with full access to the on which the citizen has entered th- to case management, ruling, finalising self-service solution, the case informa- rough the self-service site. The robot and registration. tion, internal and external services as retrieves a template which merges well as other data from F2. The robot relevant phrases and information. Many self-service sites are often just facilitates an interactive and intelligent Then the robot generates and sends advanced forms. In contrast, F2 offers a self-service, guiding the user and custo- an individual confirmation email, robot which enables a fully integrated mising the interface to fit a given case. including relevant attachments, to and automated self-service solution the citizen. based on rules and case context.

The robot can also generate a This eliminates many of the case pro- pertinent document with a ruling cessing tasks, because the case itself is based on the case information, created by the citizen, and a number of such as an individual complaint checks are made already at case crea- guide that matches the given ruling tion. In addition, the quality is markedly and context. increased since relevant information is verified and made consistent already from the beginning.

4 | cView – Artificial intelligence in case managemnet Utilising artificial intelligence, F2 becomes an efficient platform for automating work tasks and case handling

Robot for Support for case monitoring rulings

F2 offers a robot for advanced case Utilising information from case steps, For case steps that demand an monitoring, which includes the setting metadata, including progress codes, estimation, the case manager up of alarms. and variance for each case type, it can be supported by both rules is also possible to attach alarms to and recognition, which includes F2 offers an effective list control to and initiate messages for each case utilising external services. This is monitor the case progress, which also process. possible since each case step in makes automated, online case moni- F2 is open and able to call on toring possible. By utilising information Thus, overview and intelligent moni- external services through web from each case type, including each toring are fully integrated parts of F2. services. On a given step, F2 can individual case step, metadata and call on an external AI service with variance, the standard system is able select information after which an to display progress and status in the answer is presented to assist the form of lists, dashboards, etc. case manager.

Mass proces- Robot for hearings involving external sing of cases participants

F2 allows for the configuration of F2 offers a robot which supports auto- requesting clarification of the reported rules for automatically processing mated hearings of external participants information. When the citizen submits multiple cases. Such a rule set is on a given case step. an answer, the case is automatically typically configured as an autono- updated, and the case guide registers mous case guide in order to docu- This happens through the so-called that a response has been received. ment the mass processing itself. clarification request sites generated by the robot and sent to the citizen. It is possible to create a case The citizen receives a link to a website guide which selects a number of cases and processes them according to a given rule set and then document the results in the form of a new case. Examples of this could be sending letters to a group of citizens, collecting fees or managing grants. Case deletion

With F2 version 6.2 comes the ability to delete cases, for example to support the “right to be forgotten”and in relevant documentation is produced to comply with data protection regulations.

cView – Artificial intelligence in case handling | 5 AI – Artificial Intelligence

The basic principles of artificial intelligence

Artificial intelligence (AI) is not something new. What is new is the processing power.

In 1950, the mathematician Alan Turing to better utilise the mathematics and im- ding to rules. In principle, these parts developed an ability test for artificial intel- plement an even more advanced artificial of the case management could be fully ligence, and in 1957 Nobel prize winner intelligence. automated. Naturally, a premise for fully Herbert Simon predicted that some day automated case management is that legi- computers would solve greater tasks than The two basic types of artificial intelli- slation allows it for each case type. humans were able to. But it is within the gence last 5-10 years that serious progress has Artificial intelligence revolves around The subsymbolic methods will often be been made. mimicking human cognition and is more effective in relation to counselling divided into two fundamental paradigms: and supporting the case manager’s esti- The reason is not new logic nor algorith- symbolic and subsymbolic. In connection mations, e.g. pointing out irregularities or ms. The mathematics and methods that with F2, we designate the symbolic as fraudulent behaviour that calls for a closer are used in relation to artificial intelligence rule-based and the subsymbolic approach examination of a given case. today have not changed much. What did as recognition-based. Both approaches change, however, is the processing power may be of use in case processing, but in The greater part of the administrative and data capacity that make it possible different areas. case management is based on rules. It is also the rule-based artificial intelligence The symbolic paradigm has its roots in based on the symbolic paradigm that has the 1950s. Its basic idea is to mimic the been developed the most. This means reasoning of the conscious human mind, there is a great potential for artificial e.g. planning of tasks, searching for data intelligence in case management already Authority Service and defining rules for decisions. This today. paradigm is characterised by its robust and explicable methods. In principle, its Based on the model for Business Proces results are “faultless”, since they are 100% digital bureaucracy predictable, exactly based on the rules We consider F2 a unique platform for

Casetype used by the given method. The symbolic both types of artificial intelligence and paradigm is used e.g. in computer chess; automatisation because it is fully integra- the computer calculates its next move ted with cBrain’s model for administrative Modeling services based on algorithms. work, the “digital bureaucracy” model.

F2 is based on a model for the In contrast, the subsymbolic paradigm is The integrated, model-based approach digital bureaucracy, which is based based on simulating basic physical pro- enables the utilisation of artificial intel- directly on the administration’s cesses of the human brain, e.g. through ligence in the context of the authority’s work. simulated neural networks. work, from its process and data down to The authority has legitimate respon- the tasks of the individual case manager. sibility for managing and providing Case management is mainly rule-based This allows for the significant advantages a range of services (the authority’s Case management work is often divided of well-defined solutions and for ongoing resort area). The authority delega- into rules and estimations. It is also chara- impact measurement and learning. tes this responsibility to a number of specialist offices, each responsible cterised by zero tolerance of errors. for given processes, where each The agile approach means that artificial process consists of one or more This makes the symbolic, i.e. rule-based, intelligence can be implemented through case types. methods particularly effective in the parts easy-to-control, smaller projects in which of case management performed accor- the basic (bureaucratic) principles of the

6 | cView – Artificial intelligence in case managemnet Context Examples of rule-based AI Examples of recognition-based AI

Services Rule-based AI makes it possible to Recognition-based AI makes it possible to Proces & Data define rules for monitoring on a general recognise patterns and developments related level of the production from a process to production and the authority’s data or data perspective

Case type Rule-based AI makes it possible to Recognition-based AI makes it possible to Case guide and case define rules for automating an entire reveal patterns related to an entire case Artificial intelligence (AI) is not something new. steps case process or parts thereof (governed process or parts thereof and to inform the case What is new is the processing power. by case type) manager

Function Rule-based AI makes automation of Recognition-based AI makes it possible to Complex and simple concrete system functions possible, thus reveal patterns related to an entire case pro- minimising the user’s work tasks cess or parts thereof and to counsel the case manager

administration, such as requirements of The fully integrated, agile approach Could artificial intelligence replace the compliance and transparency, are kept. means that artificial intelligence becomes case manager? In addition, the impact on production is not just a peculiar science project, but It is natural to ask the question, “to what minimal, as artificial intelligence can be a natural part of the authority’s digital extent will artificial intelligence replace implemented using a gradual transition in transformation and evolution in both the jobs?” which an increasing part of the authority’s short and the long run. work tasks and processes are automated. There is no doubt that we today have The table above offers context and the technology to replace large parts of Using F2 as a platform, then, great examples for the applications of rule-ba- rule-based case management. However, improvements can be made to quality, cost sed and recognition-based artificial this process would require several legal reduction, citizen service and transparen- intelligence in F2. clarifications and decisions on both a na- cy, as it is based directly on the work of tional and international level. Conversely, the administration. Evidently, artificial intelligence can be it is evident that machines possess neither applied to several levels, and support verbal nor social intelligence. Based on the model it is possible to define thereof may be implemented in the form case guides in F2 directly and store them of well-defined projects. This ensures In principle, artificial intelligence could in libraries. When a new case is created, a that artificial intelligence becomes neither replace the rule-based parts of the case case guide can be attached to define case disruptive or too complex, but instead a management, while parts that demand steps, rules, ruling, responsibility, docu- natural, controlled and agile tool included estimations and intelligent interaction mentation, etc. as a fundamental part of the authority’s with citizens are likely to be carried out by digital strategy and projects. cBrain’s humans for the foreseeable future. F2 offers an agile user experience in which ongoing product development as well digitisation and automatisation of the as the development of new services are The potential of supporting the rule-ba- authority’s work are facilitated through grounded in the three levels shown in the sed case management with artificial intelli- the support and automatisation of each AI context table: services, case type and gence is great and the support and de- case type. Each case type can be further function. velopment of case management through digitised and automated through ongoing pattern recognition will only increase. revision of the case type library and through the utilisation of new types of artificial intelligence.

With F2, then, the authority is able to keep optimising and improving while gaining experience and implementing new techniques.

cView – Artificial intelligence in case management | 7 Customer index Our Danish Customers

Danish Ministries Unions and Active Patient Support - Across Four Ministry of Employment Regions Organisations Ministry of Industry, Business and The Evangelical Lutheran Church - The Danish Union of Teachers Financial Affairs Deaneries The Ministry of Employment The Prime Minister’s Office Danish Unemployment Funds MinistryMinistries of Finance Ministry of Social Affairs and the Interior Ministry of Transport and Housing Danish Psychological Association Ministry of Ecclesiastical Affairs TheState Ministry of Greenland of Ecclesiastical Affairs Ministry of Higher Education and Science Danish Union of Social Counsellors Ministry of Climate, Energy and Utilities MinistryIT University of Industry, of Business and Financial The Ministry of Foreign Affairs Affairs The MinistryDjøf - Unionof Immigration for Graduates and Integration in Social Ministry of Environment and Food of - Internatio- The Ministry of Climate, Energy and Utilities MinistrySciences, of Children, Business Education and andLaw. Gender nal House, Copenhagen Ministry of Finance EqualityDanish Association of Occupational Ministry of Taxation Environmental Protection Agency The Ministry of Environment and Food State Therapistsof Greenland Ministry of Social Affairs and the MinistryAgency of forTaxation Modernisation Danish Association of Pharmacologists Interior Agency for Digitisation DanishThe Labour Danish Court National Union of Upper The Prime Minister’s Office The Ombudsman for Greenland Authorities, Rail Denmark Vestforsyning,Secondary Water School and Heating Teachers Ministry of Transport and Housing The Ombudsman for The Faroe Islands and Danish Meteorological Institute The OmbudsmanDanish Diet Greenland and Nutrition Association Ministry of Higher Education and University Danish Energy Agency AgencyDanish for Development Medical Association and Simplification Science DanishRudersdal Business Municipality Agency Agency of Employee and Competency other The Social Democratic Party of Den- Ministry of Foreign Affairs The Evangelical Lutheran Church - The The National Board of Social Services mark institutions Deaneries Odsherred Municipality Ministry of Immigration and Integration Agency for Governmental IT Services Active Patient Support - Across Four Regions Copenhagen Municipality - International DanishAgency Economy for Governmental Agency Administra- House, Copenhagen Environmentaltion Protection Agency Agencies, Municipalities TheDanish Ombudsman Agency Faroe for International Islands Recruit- VordingborgPrivate Municipality Sector Thement National and Integration Board of Social Services LiverpoolCompanies City Council and Other Authorities AgencyDanish for Maritime Governmental Authority Administration Agency for Governmental IT Services IT University of Copenhagen Agency for Development and Simplifi- Danish Labour Court Agency of Family Law Roskilde University cation The Danish Bar and Law Society Danish Agency for International Recruitment University of Bristol Rail Denmark Beierholm - State Authorised Auditing andVestforsyning Integration Water and Heating Agency for Digitisation DanishVordingborg Maritime Municipality Authority Nilfisk - Manufacturing Company Danish Meteorological Institute Rambøll Management Consulting Danish Energy Agency Danish Business Authority Danish Union of Teachers Danish Association of Pharmaconomists Unions, Danish Unemployment Funds The Danish National Union of Upper Agency of Family Law unemployment Danish Psychological Association Secondary School Teachers Danish Union of Social Counsellors Danish Diet and Nutrition Association funds and DJØF - Union for Graduates in Social Scien- Danish Medical Association ces, Business and Law The Social Democratic Party of Denmark organisations Danish Association of Occupational Therapist

cView cView published by cBrain A/S Responsible publisher: Ejvind Jørgensen, CSO & CFO Editor: Inger Vinther Veber, [email protected]

www.cbrain.com · cBrain A/S · Dampfærgevej 30 · Copenhagen · Denmark