This year our presenters represent these 12 countries:Presenters and Presentations: (click on the title to see abstracts, photos, and biographies)
Mike Gualtieri: The Future Of Enterprise AI And Digital Decisions
Dr. Alan Fish: Collaborative decisions: coordinating automated and human decision-making
Denis Gagne: Making the Right Decision at the Right Time: Introducing Temporal Reasoning to DMN
Ron Ross: Brainstorming Next-Generation Rule Platforms
Prof. Jan Vanthienen: Quality of Decision Models
Prof. Robert Fourer: Model-Based Optimization for Effective and Reliable Decision-Making
Mario Fusco, Matteo Mortari: How and Why I Turned a Rule Engine into a First-Class Serverless Component
Gary Hallmark: DMN 2.0?
Silvie Spreeuwenberg: The Decision Model for Gate Allocation
Guilhem Molines: Business Rules – Focus on “Business”
Fernando Donati Jorge: Combining Decision Models for better Decision Management
Yoshihito Nakayama: Business innovation of BPO realized by Task Center and AI and Rule Engine
Daniel Schmitz-Hübsch, Ulrich Striffler: FEEL, is it Really Friendly Enough?
Keith Swenson: Close is not Close Enough
Edson Tirelli, Matteo Mortari: Machine Learning + Decision Management: a standards based approach
Brian Stucky: Beyond Decision Models – Using Technical and Business Standards to Transform Financial Services
Kedar Kulkarni, Harriet Parkinson, Larry Goldberg: Decision Management Journey at Hiscox Claims
Roy Robinson, Alan Giles: Case Study – Building an enterprise ready decision management platform for customer engagement and next best action and how to help businesses align for success
Ulrich Striffler, Daniel Schmitz-Hübsch: Modelling Decisions with Embedded Testing
Dario Campagna, Carlos Kavka, Alessandro Turco: The Role of DMN and BPMN in the Design of Composite Materials
Marjolein Deryck: Meeting the Expectations with DMN and Constraint Solving : The Notary Case
Charlotte Bouvy: Spaghetti Spreadsheets Untangled – Benefits of decision modeling when uncovering complex business logic hidden in spreadsheets
Nicholas Decleyre, Marjolein Deryck, Bram Aerts: Combining DMN, First Order Logic and Machine Learning: The creation of Saint-Gobain Seals’ Digital Engineer
Tim Stuyckens: Exploiting payroll knowledge with Viren
Martijn Tromm, Marten Schokking: A hybrid implementation of multi-channel, multi-modal, high volume financial risk monitoring
Dennis Aarts: Decision Management as a Service
Presentation Abstracts and Brief Biographies
Keynote: The Future Of Enterprise AI And Digital Decisions by Mike Gualtieri, Forrester Research
Machines are amazing learners. Humans are amazing learners. AI is best when powered by both. AI solutions that employ both machine learning and knowledge-engineered rules learn continuously from data whilst at the same time are informed by wisdom and commonsense expressed in rules. This keynote will convey the key trends in enterprise AI and rules/decision management; and discuss how leading enterprises can use them in combination to build truly learning AI-infused applications at scale. Keywords: Artificial Intelligence, Machine Learning, Knowledge Engineering, Rules/Decision Management.
Mike Gualtieri is Forrester Research Vice President and Principal Analyst. Mike’s research focuses on software technologies, platforms, and practices that enable technology professionals to deliver digital transformations that lead to prescient digital experiences and breakthrough operational efficiency. His key technology coverage areas are AI, machine learning, deep learning, AI chips and systems, digital decisions, streaming analytics, prescriptive analytics, big data analytical platforms and tools (Hadoop/Spark/Flink; translytical databases), optimization, and emerging technologies that make software faster and smarter. Mike is also a leading expert on the intersection of business strategy, artificial intelligence, and innovation.
Mike provides technology vendors with actionable, fine-tuned advisory sessions on strategy, messaging, competitive analysis, buyer-persona analysis, market trends, and product road maps for the areas he directly covers and adjacent areas that wish to launch into new markets or use new technologies. Mike is a recipient of the Forrester Courage Award for making bold calls that inspire leaders and guide great business and technology decisions.
Collaborative decisions: coordinating automated and human decision-making by Dr. Alan Fish, FICO
Although decision modelling tends to focus on automated decision-making, most commercial applications of AI also involve human decision-making. Typically, human decision makers are “knowledge workers”, that is, they bring knowledge to bear which cannot easily be encapsulated in automated decision services. Achieving effective collaboration between automated and human decision-makers requires a formal definition of the relationship between these actors, which can be provided using BPMN, CMMN and DMN. Design patterns are suggested for some common approaches, which are differentiated by the formal roles allocated to the automated and human components of the system. Keywords: Process modelling, Case modelling, Decision modelling, BPMN, CMMN, DMN, Knowledge worker, Decision service.
Dr. Alan Fish is an authority in Decision Modelling and Decision Management, especially in the support and/or automation of organisational decision-making. With over 30 years experience in this field, he has been responsible for many projects at the forefront of current technology. He invented the “Decision Requirements Diagram” (DRD) which exposes the structure of a domain of decision-making, and developed Decision Requirements Analysis (DRA): a methodology for building and using such decision models. He is the author of “Knowledge Automation: How To Implement Decision Management in Business Processes” (Wiley), and a co-author of the OMG specification Decision Model and Notation (DMN).
Making the Right Decision at the Right Time: Introducing Temporal Reasoning to DMN by Denis Gagne, Trisotech
Time is a crucial factor for many of business everyday decision activities. Within business operations decisions are frequently dependent on the time, or period of time, in which they are made. Yet the relationships between time and decision making in the current version of the Decision Model and Notation (DMN) remains quite basic. Manipulating various common notions related to time in DMN expressions can get quite complex very rapidly.
In this presentation we will explore the problem of representing temporal knowledge and temporal reasoning within DMN. We will demonstrate how to make decisions in the context of sometime imperfect temporal information by introducing a series of new temporal constructs to DMN. Keywords: DMN, FEEL, Time, Temporal Representation, Temporal Reasoning
For over a decade Denis Gagné has been a driving force in the majority of international BPM standards in use today. He is a member of the Workflow Management Coalition (WfMC) Steering Committee, chair of the Business Process Simulation Working Group (BPSWG), and co-Editor of the XPDL 2.2 process definition standard. For the Object Management group (OMG), Denis is the Chair of the BPMN Interchange Working Group (BPMN MIWG), and a member of the Business Process Model and Notation (BPMN), Case Management Model and Notation (CMMN) team and Decision Management (DMN) team.
Brainstorming Next-Generation Rule Platforms by Ronald G. Ross, Business Rule Solutions
It’s time to brainstorm next-generation rule platforms. The opportunity is not just to get smarter, but to reduce – or eliminate – programmer workload (maybe even programmers?). Let’s consider some rule realities. Many rules can be broken. Why two decades into the 21st century that obvious reality is still not supported by rule platforms is a mystery. Rules that people or organizations can break are called behavioral business rules. Think laws, regulations, contracts, agreements, MOUs, certifications, warranties, etc. Such rules are often one-off and often sourced directly in natural language.
A great impedance mismatch exists between DMN / decision rules vs. SBVR / behavioral business rules based on whether evaluation of rules is modeler-invoked or state-based. The burden of validating states of affairs in modeler-invoked evaluation is shouldered by coders.
Modeler-invoked evaluation does not work well for behavioral business rules. Not catching violations ASAP causes snowballing errors downstream and thus extensive rework. No wonder business software remains so complex and brittle.
Flash points are the specific events when a rule needs to be evaluated based only on its semantics – i.e., using no external model or specification (e.g., procedural model or decision model). Flash points for the same rule can occur in multiple processes, procedures, use cases, etc., or at various points in ad hoc (unmodeled) business activity. Invoking flash points automatically requires ‘stateful’ platforms.
I’m sure you can envision other opportunities for next-generation rule platforms. Come to this interactive session and let’s figure out together how to reignite the rule space. Read more
Ron Ross, Principal and Co-Founder of Business Rules Solutions, LLC, is internationally acknowledged as the “father of business rules.” Recognizing early on the importance of independently managed business rules for business operations and architecture, he has pioneered innovative techniques and standards since the mid-1980s. He wrote the industry’s first book on business rules in 1994. With BRS’s client roster of Fortune 500 companies and governments, Ron consults,speaks and teaches worldwide. He has served as the chair of the International Business Rules & Decisions Forum conference since 1997, now part of the Building Business Capability (BBC) conference. Ron is also the author of 10 professional books, as well as the executive editor of the Business Rules Journal. Through these publications, as well as on the online forum BRCommunity and his blog, Ron enjoys sharing his knowledge and experience in consulting and business rules.
Quality of Decision Models by Jan Vanthienen, KU Leuven
A number of frameworks for data quality exist. This presentation introduces a decision model quality framework, along the lines of data quality indicators, and its implications for decision modeling. Quality of decision models is defined across multiple dimensions, e.g.
Internal: Accuracy, Correctness, Adaptability, Complexity
Contextual: Relevance, Completeness, Fitness, Timeliness, Integration
Representational: Ease of understanding, Compactness, Consistency
Access: Traceability, Security
Obtaining decision model quality requires a number of steps at different levels: Governance, Management and Operational. This presentation provides an overview of success factors in improving decision model quality.
Prof.Jan Vanthienen received his PhD degree in Applied Economics from KU Leuven, Belgium. He is a full professor of Information Systems at the Department of Decision Sciences and Information Management, KU Leuven and (co-)authored more than 200 full papers in international journals and conference proceedings. His research interests include information and knowledge management, business rules, decisions and processes, and business analysis and analytics. He received an IBM Faculty Award on smart decisions, and the Belgian Francqui Chair at FUNDP. Currently he is department chair at the Department of Decision Sciences and Information Management of KU Leuven..
Model-Based Optimization for Effective and Reliable Decision-Making by Robert Fourer, AMPL Optimization Inc.
Optimization originated as an advanced mathematical technique, but it has become an accessible and widely used decision-making tool. A key factor in the spread of successful optimization applications has been the adoption of a model-based approach: A domain expert or operations analyst focuses on modeling the problem of interest, while the computation of a solution is left to general-purpose, off-the-shelf solvers; powerful yet intuitive modeling software manages the difficulties of translating between the human modeler’s formulation and the solver software’s needs.
This talk introduces model-based optimization by contrasting it to a method-based approach that relies on customized implementation of rules and algorithms. Model-based implementations are illustrated using the AMPL modeling language and popular solvers. The presentation concludes by surveying the variety of modeling languages and solvers available for model-based optimization today. Keywords: Optimization, Modeling languages, Mixed-integer linear programming.
Prof. Robert Fourer is a scientist working in the area of operational research and management science. He is currently President of AMPL Optimization, Inc and is Professor Emeritus of Industrial Engineering and Management Sciences at Northwestern University. Robert Fourer is recognized as being the designer of the popular modeling language for mathematical programming called AMPL. Together with David M. Gay and Brian Kernighan he was awarded 1993 ORSA/CSTS Prize by the Computer Science Technical Section of the Operations Research Society of America, for writings on the design of mathematical programming systems and the AMPL modeling language. Robert Fourer was also awarded Guggenheim Fellowship for Natural Sciences in 2002.
How and Why I Turned a Rule Engine into a First-Class Serverless Component by Mario Fusco, Red Hat
These days rule engines are often overlooked, possibly because people think that they are only useful inside heavyweight enterprise software products. However, this is not necessarily true. Simply put, a rule engine is just a piece of software that allows you to separate domain and business-specific constraint from the main application flow. I am the project lead of Drools, the rule engine of Red Hat, and my target was to modernize my project and make it ready to be used in serverless environments. In this talk I will explore and make sense of technologies like GraalVM and Quarkus. In particular I will show, with very practical use cases taken from my experience with this migration, what is necessary to change in a code base making extensive use of reflection, dynamic class loading and other Java sorceries in order to make it compatible with those technologies and demonstrate how this is allowing us to make Drools part of the cloud and serverless revolution. Keywords: Java, Drools, GraalVM, Serverless, Function as a service. Read more
Mario Fusco is a principal software engineer at Red Hat working as Drools project lead. He has a huge experience as Java developer having been involved in (and often leading) many enterprise level projects in several industries ranging from media companies to the financial sector. Among his interests there are also functional programming and Domain Specific Languages. By leveraging these 2 passions he created the open source library lambdaj with the purposes of providing an internal Java DSL for manipulating collections and allowing a bit of functional programming in Java. He is also a Java Champion and the co-author of “Modern Java in Action” published by Manning.
Matteo Mortari is a Software Engineer at Red Hat, where he contributes in Drools development and support for the DMN standard. Matteo graduated from Engineering with focus on enterprise systems with a thesis involving rule engines which sparked his interests and influenced his professional career since. He believes there is a whole new range of unexplored applications for Expert Systems (AI) within the Corporate business; additionally, he believes defining the Business Rules on the BRMS system not only enables knowledge inference from raw data but, most importantly, helps to shorten the distance between experts and analysts, between developers and end-users, business stakeholders.
DMN 2.0? by Gary Hallmark, Oracle
In DMN’s 8 year lifespan from RFP to version 1.3, many features, changes and enhancements have been proposed and discussed. Some changes were rejected because they would not be backward compatible — required for all 1.x specifications. Some enhancements were rejected as being out of scope (too big) or out of spirit (too different). Many issues languish for lack of a volunteer to spend the considerable amount of time to write a detailed proposal to change a mature specification. In this presentation, I will briefly discuss some of the top ideas for a future version of DMN:
1. Context-free grammar,
2. better align DMN, BPMN, CMMN,
3. additional business types,
4. better support for iteration,
5. new boxed expressions,
6. type theory, standard model validation,
7. cyclic requirements, recursion, and
8. solve the model in reverse: what inputs result is a given output?
This talk is meant to be an interactive discussion about possible future directions and all opinions expressed are those of the author and not of Oracle nor of OMG.
Keywords: Decision Model, Decision Requirements, Decision Table, DMN, BPMN, Business Rules, FEEL.
Gary Hallmark has been at Oracle since 1989. He invented Oracle’s distributed transaction technology and received Oracle’s first patent. He invented Oracle parallel query technology. Moving from database to middleware, Gary contributed to Java and XML based middleware technologies, including Weblogic, XML database, JAXB, security, fault-tolerance, SOA, web services, application integration, connectors and adapters, message oriented middleware and JMS, business rules, business processes (BPEL and BPMN), workflow management systems, object-relational mapping, domain-specific languages, Swing and JSP UI development. Gary contributed to standards including OMG Corba, J2EE (JMS, EJB), W3C RIF (co-author), and OMG DMN. Currently, Gary is an Architect for Oracle Process Cloud Service (BPMN, business rules, workflow, and more) and is leading the development of a Decision Modeling product based on OMG DMN. Gary invented the DMN FEEL language. He co-chaired the DMN standards committee, and co-authored the standard.
The Decision Model for Gate Allocation by Silvie Spreeuwenberg, Librt
Schiphol Group creates an integrated plan on a tactical, strategic and operational level that is adjusted daily based on the actual operational situation. Internal research demonstrated that planning procedures are not optimally aligned; many plans depend on the aircraft stand allocation plan which changes heavily on the day of operations. These findings motivated a research project into the decisions and criteria used by gate planners.
Keywords: DMN, Airport, Gate allocation. Read more
Silvie Spreeuwenberg is an experienced entrepreneur and consultant. She
combines the ability to be a holistic thinker while, at the same time, she has
detailed knowledge about artificial intelligence, compliance and software
development. Therefore, she is a good strategy advisor for founders, scaleups
and start-ups. Her inspiration is rooted in a strong need to create
sustainable partnerships based on trust. She has an urge to share
knowledge by being extremely transparent. These are the themes in her work..
Business Rules – Focus on “Business” by Guilhem Molines, IBM
In this talk, we will explore some of the key features that make rules – and decisions in general – more consumable by their target audience, the business user. We will cover the modeling and authoring experience, but also steps that are too often ignored, from authoring to testing to simulation, all the way to actual deployment in production. We will pay special attention to the bottom-up and top-down approaches, especially when it comes to defining the data model that the decisions are written against. Finally, we will open the discussion to potential techniques that can help put the focus again on the user we want to empower: the business analyst.
Keywords: authoring, modeling, design thinking, consumability, user experience. Read more
With a background in fundamental Computer Science and Artificial Intelligence, Guilhem Molines has been involved with decision technology for more than two decades, both on the field as a consultant and as an architect of the ILOG, then IBM product team behind ODM (Operational Decision Manager). With a special focus on the modeling and authoring experience, Guilhem is always in close contact with users and practitioners and willing to find innovative ways to make the authoring of decisions an easier task for the industry..
Combining Decision Models for better Decision Management by Fernando Donati Jorge, FICO
DMN Decision Models are a great visual representation of how decisions are made and the authority over the knowledge used to make those decisions. However, they don’t represent why those decisions matter: what performance metrics relevant to the business might be impacted by the different decision outcomes. By combining DMN Decision Models with other types of Decision Models, businesses can effectively create a visual representation of both how decisions are made and why they matter to the business, with wide application across the Decision Management lifecycle.
Keywords: DMN, Decision Models, Decision Management. Read more
Fernando Donati Jorge oversees the product management of FICO
Decision Tools, including Blaze Advisor, Decision Modeler and DMN
Modeler. He brings more than 20 years of experience to the role,
including 13 years at FICO in various capacities…
Business innovation of BPO realized by Task Center and AI and Rule Engine by Yoshihito Nakayama, NTT DATA INTRAMART
In recent years, the demand for BPO (Business Process Outsourcing) work has greatly increased, but there is room for various improvements in the productivity of its business operation. We are working on the improvement of business productivity especially using AI and rule engine, and the effect has been confirmed through many POC projects.
In this presentation, we introduce the following two features:
– Task Center: Various tasks to be tackled are collected automatically from existing systems and are aggregated them on the screen. On the screen, the priority of the tasks are automatically determined and displayed based on the due date etc., and the worker’s decision-making to improve the work efficiency is supported.
– Dash-board: Dash-board has the mechanism for allocating tasks to team members optimally according to their work volume and skills, etc. during busy season. Dash-boards run simulations to optimize team members’ tasks through business visibility. Optimal task allocation can be performed while repeating simulation based on several policies such as performance, operation leveling, cost, etc.
Furthermore, we will introduce methods for visualizing the situation of complex on-site operations. The conventional process mining is based on the use of system logs, but we propose a method to acquire task execution logs even in environments where there are no systems such as distribution centers and factory sites. This makes it possible to construct a digital twin of on-site business processes in any business. Keywords: BPO, Improvement of business productivity, Allocating tasks to team members optimally, Digital twin of on-site business processes. Read more
Yoshihito Nakayama received the B.E., M.E. degrees from Graduate School of Agricultural and Life Sciences, The University of Tokyo, Tokyo, Japan, in 1990, 1992 respectively. After graduating, he joined NTT DATA and then founded NTT DATA INTRAMART CORPORATION in 2000. NTT DATA INTRAMART has acquired No. 1 share in the Japanese workflow market and listed on TSE Mothers in 2007. Currently it has installed over 5000 companies in Japan and Asia mainly. On one hand, he is enrolled in doctoral course at School of Engineering, The University of Tokyo. He received the best paper special award at ISSJ (Information Systems Society of Japan) in 2017. His research interest includes BPM, Decision Management, Process Mining, and workflow. He is a member of the ISSJ, IEICE, and IPSJ.
FEEL, is it Really Friendly Enough? by Daniel Schmitz-Hübsch and Ulrich Striffler, Materna
DMN FEEL is committed to be a markup language that can be used by business users to model the logic of a decision. However, this claim has been discussed for a long time. Do you really need it or aren’t decision tables enough? Do business users need to be capable of programming to use FEEL? The presentation illustrates why FEEL is useful for business users and where its strengths can be found. In addition, practical experience will demonstrate what you have to consider when using the current FEEL.
Furthermore, alternative representations of FEEL expressions in the form of visual structural programs are shown and discussed. These represent an alternative opportunity to model logical processes, as known from programming. It will also be discussed whether these can be used easier for business users rather than to use a verbal markup language.
The presentation will end with the quite controversial discussion of translating FEEL into country- specific languages. A workflow will be shown how the progress the can be supported automatically in order to transform the country-specific language into standard-compliant FEEL. Keywords: DMN, FEEL, Decision Table. Read more
Daniel Schmitz-Hübsch is a Software Developer at Materna GmbH.Daniel holds a Master degree in Business Informatics with focus on mobile oriented analysis of business processes. For five years, he has been involved in the modelling and technical implementation of business process- and decision management systems. As a software developer for an independent IT company, he is responsible for the development of high-availability decision applications using rule engines like IBM Operational Decision Management.
Modelling Decisions with Embedded Testing by Daniel Schmitz-Hübsch and Ulrich Striffler, Materna
When creating complex decision models, quality management is at least as complex as the actual creation of decisions. The modeler lacks the tools to ensure the technical correctness for business patterns that are decidedly complex. In such cases, the functional testing of the models often takes place after the actual modeling, as a part of the technical implementation. The consequences of this approach are time-consuming adjustments, faulty models and an increase in costs.
A less complex procedure would be to check the rules of the decision development process at every stage. Both the modeler and the developer would benefit from a technically correct decision. The modeler, because he receives an immediate technical feedback at an early stage of the modelling and does not have to wait for the technical development. The developer, because such a technically secure modelling result has significantly less volatility than a still insufficiently tested modelling.
With the open source application “DMN-Tester” such a procedure should be supported. The browser-based application allows, based on the RedHat DMN engine, the structured creation, administration and execution of test cases based on DMN models.
In our presentation we will show how test cases can be created and executed by business users during the decision-making process based on our practical project experience. We show how complex test cases can be created methodically within the framework of the process model. The focus of the presentation will be on the methodical creation of complex test cases based on DMN models. The presentation will conclude with a live demonstration of how to work with the DMN-Tester in practice.
Keywords: DMN, Testing, RedHat. Read more
Ulrich Striffler is a Senior Consultant at Materna GmbH. Ulrich has a degree in computer science with a focus on eHealth and artificial intelligence. For 30 years he has been working as a software engineer and architect, case and process modeler and requirement engineer. Within the last six years he has been involved in the modelling of business process- and decision- management systems with BPMN and lately with DMN as well. As a requirements and method engineer for an independent IT company in Germany, he is responsible for the development of high-availability decision applications for the government.
Close is not Close Enough by Keith Swenson, Fujitsu
For an international standard to work and be viable, it is critical that everything be completely specified and that everything works exactly as specified. For any kind of “language” that means millions of potential edge conditions. The confidence of the audience is effected if there is any slight difference in implementations that stop a user from being successful. To avoid that, you need a test suite. The DMN TCK effort provides exactly that test suite.
The original idea was to simply to back up the written spec with some real running code. The specification can’t possibly express the full detail necessary for an actual implementation. Anyone faced with implementing DMN would have to make thousands of small design decisions between, say, behavior X or behavior Y. Most of those are arbitrary in the sense that the a user could use behavior of either of X or Y, but what matters is that the product is consistent. Moving from product to product requires that all implementations make the same choices at those levels. The specification is, by its nature, ambiguous on these many small points.
The DMN TCK has more than 1000 test cases which any implementation of DMN should be able to run. We use the standard DMN XML file format, and the test cases themselves are easy read and use to automate testing. Test runners are available in several languages for use. All of this is freely available as Creative Commons open source. There is a web site listing the vendors that have demonstrated compliance.
It is most important for consumers to understand that if they are purchasing a DMN solution, they really should have the confidence of knowing that their vendor has pass all the tests which demonstrate that they really run DMN.
Keywords: DMN, TCK, Decision Model Notation.
Keith Swenson is Vice President of Research and Development at Fujitsu America Inc. and is the Chief Software Architect for the Interstage family of products. He is known for having been a pioneer in collaboration software and web services, and has helped the development of many workflow and BPM standards. He is currently the Chairman of the Technical Committee of the Workflow Management Coalition. In the past, he led development of collaboration software MS2, Netscape, Ashton Tate and Fujitsu. In 2004 he was awarded the Marvin L. Manheim Award for outstanding contributions in the field of workflow. His blog is at http://kswenson.wordpress.com/.
Machine Learning + Decision Management: a standards based approach by Edson Tirelli and Matteo Mortari, Red Hat
Machine Learning (and AI in general) is not new, but has had a huge resurgence in the last few years, both in new investments and general adoption. The success of Social Networks and the increased availability of data collected by fully connected, always online systems, makes any solution that can make sense of such data very attractive. On the other hand, the increased demand for transparent, explainable decision making, that is accurate, consistent and effective, has never been greater. Legislations like GDPR are just a result of increasing concerns about privacy, safety and transparency in general. While AI/ML solutions are great at making sense of high volumes of data, the reasoning process for most of the generated analytic models is usually quite opaque.
Decision Management on the other hand, is a discipline that typically aims to provide full transparency on the decision process, but requires translating knowledge into decisions/rules, using some form of knowledge engineering (automated or not).
However, there is a middle ground between these two approaches: combining decision models with analytic models is turning into an increasingly effective method to achieve a higher level of transparency, without losing effectiveness. Such approach achieves several of the goals of an Explainable AI (XAI), but still leverages abstract knowledge extracted from data for quality results.
While many technologies and frameworks exist to implement such strategy, a standards-based solution can be employed with several benefits over proprietary approaches. In particular, PMML (Predictive Modelling Markup Language) is a well established standard for the representation of predictive models, automatically generated from datasets using well known AI/ML technologies. DMN (Decision Model and Notation) is a Decision Modeling standard that provides out-of-the-box, transparent integration with predictive models. During this presentation, attendees will have the opportunity to learn how the combination of these two standards brings us an easy to use, high level, vendor neutral and effective solution for Explainable AI.
Keywords: Artificial Intelligence, Machine Learning, Decision Management, Standards, PMML, DMN.
Edson Tirelli is JBoss Drools/Red Hat Senior Principal Software Engineer, Development Manager, Decision Manager Architect, and former Drools Project Lead. He has a deep experience in a variety of Java related technologies and frameworks, including JEE, JSE, WS/REST, JVM architecture and bytecode generation, ANTLR and JIT compilation, among others. He is the main developer of the Red Hat DMN Modeler.
Beyond Decision Models – Using Technical and Business Standards to Transform Financial Services by Brian Stucky, DecisionX and MISMO
MISMO – the mortgage industry’s standards organization – recently announced they officially recommended the use of the Decision Model and Notation (DMN) standard for documentation, implementation, execution and exchange of business rules and decisions across the mortgage industry. Combining DMN with MISMO’s own mortgage data standard has the potential to truly revolutionize and transform financial services across a variety of areas. This presentation will focus on few of these approaches within the financial services industry:
1. Enabling exchange between trading partners, an enterprise and their clients, and even within previously siloed organizations.
2. Creating APIs and microservices to make automated decisioning accessible in new ways and level the playing field for businesses of all size.
3. Promoting decision models as a means for regulatory agencies to present compliance requirements.
4. Enhancing the concept of a truly digital mortgage by sending decision models along with the loan (perhaps as part of a blockchain) so full transparency, documentation and compliance is established.
Each of these areas will be demonstrated with case studies of real world financial services and mortgage industry examples.
They key takeaway will be the concept of moving beyond the traditional use of decision models and DMN to apply this approach in novel ways to handle current issues and pain points for the financial services (and other) industries.
The audience will be general – business and technical, both new and advances. Industry sector will apply most to financial services. Key technologies include DMN, decisions, APIs and microservices. Keywords: DMN, APIs, Microservices, Financial services, Mortgage, Compliance, Decision modeling.
A recognized thought leader in decision management, Brian Stucky brings three decades of experience designing and implementing business rule and process management systems. Domain experience includes the secondary mortgage market, credit card processing, mutual fund portfolio analysis, insurance underwriting, and Federal civilian agencies.Brian works closely with decision management vendors and frequently speaks at professional events. He was a contributing author to three decision management books, currently serves on the editorial board of the Business Process Management Institute and has well over 60 publications. Brian is now in his fourth year as co-chairman of the Mortgage Industry Standards and Maintenance Organization Business Rule Exchange Workgroup. His efforts there have resulted in finalizing the Decision Model and Notation (DMN) standard as the official mortgage industry standard.
Decision Management Journey at Hiscox Claims by Kedar Kulkarni, Harriet Parkinson and Larry Goldberg, Sapiens Decision and Hiscox
Hiscox UK Retail Claims context: Hiscox has consistently ranked outstanding on claims service. Hiscox UK Claims have the ambition ‘To be recognised as the leading claims service. Digital where it count, human where it matters’.
The Challenge: To deliver fair and consistent claims customer journeys, through digitising claims logic and decision making both in their internal PAS system and online. Put the control of the claims logic in the hands of UK Claims, with limited IT burden. Give Hiscox customers and brokers the ability to notify their claim online, to support the growing demand for online claims management.
The solution: Hiscox’s experience during the PoC was
• to be able to extract the logic out of the business process,
• make this visible and transparent by way of representing in a tool in a normalised business readable format and
• then be able to quickly integrate into the technology landscape almost instantly.
The journey forward with this is approach is to now start looking at widening to all claims journeys and integration with internal PAS system and online
Keywords: Decision Management Journey, Business Logic, Reuse, Digital Transformation, Simplify Process, Transparency.
Larry Goldberg is an evangelist for Sapiens DECISION, and as a member of the senior management team is responsible for all products in the Sapiens Decision company. He was Co-founder and Managing Partner of Knowledge Partners International LLC, acquired by Sapiens Decision, and has over forty years of experience in building technology based companies on four continents. Commercial applications in which he played a primary architectural role include such diverse domains as banking, healthcare, supply chain, property & casualty insurance, and enterprise modeling tools. He has been the business lead and/or business sponsor on many major projects in both the public and private sector, and is a trusted adviser to senior executives from major corporations. Larry is a leading international authority on business requirements, and is the co-author of the best-selling book “The Decision Model: A Business Logic Framework Linking Business and Technology” (Auerbach, New York 2009).
Harriet Parkinson leads the design and delivery of Hiscox UK Claims Strategic Change. The strategy includes both digital and human elements – from internal operating models to digitising the claims experience. The ambition is to be ‘Digital where it counts, Human when it matters’.
Her prior role at Hiscox was Head Of Property Claims, where in 2018 the team won the Insurance Times High Net Worth Claims Solution of the Year.
Previously Harriet worked for AXA for nearly 10 years, leading teams across a variety of departments including travel claims, sales, change and personal lines underwriting.
Harriet is a driven and ambitious change leader, who is excited by the development required in the insurance industry to not only meet but exceed our customers’ increasing expectations.
Case Study – Building an enterprise ready decision management platform for customer engagement and next best action and how to help businesses align for success by Roy Robinson and Alan Giles, Boxever
Boxever have been providing 1:1 customer personalisation capabilities to some of the world’s top airlines and financial services businesses for over 7 years. Engage our customer engagement engine was built on Drools and had a coding interface which allowed coding of rules in a user friendly domain specific language created by Boxever to provide real time personalisation. As our customers matured however they demanded the ability to arbitrate between marketing messages, service messages and customers service messages. They also were seeking ways to apply AI and machine learning across all customer touch points. This case study will discuss the challenges we faced both from a technical, business and user experience perspective in creating a decisioning capability based on the DMN standard for first best action and next best action and how these were overcome.
The key takeaways will be:
– Building the right DMN modeller for enterprise decisioning – augmenting the DMN standard to improve flexibility;
– Why technology is only a part of the solution;
– How businesses need to align to make digital decisioning for NBA/FBA a reality
– Choosing the right scalable architecture to support multi-tenant enterprise
– Decisioning in real time with complex JSON data structures
– Testing Decision Models – Component, Model and Variant
– Challenges of building optimisation workflows (A/B/N) into model deployment.
Keywords: Decisoning, Enterprise, Next Best Action, First Best Action, Organisational Alignment.
Roy Robinson Director of Product, Boxever
Roy has spent the past 10 years as a product specialist working with SaaS businesses to build out data and analytics platforms. He has a wealth of experience in building scalable real time decisioning products which leverage the latest data science techniques in both the security and customer experience fields. Roy has a passion for user experience and practical applications of AI. Roy has a Masters in Mathematics and Computer Science from Queens University, Belfast and is a SciFi and Fantasy novel fanatic!
Alan Giles CTO, Boxever
Alan is a Co-Founder and the CTO at Boxever. He is a highly accomplished software technologist, software visionary, and entrepreneur with over 14 years’ experience leading, developing and delivering highly scalable and fault tolerant distributed systems. He is passionate about large-scale systems, cloud security, and building new architectures to support petabyte-scale data processing. In addition to developing world-class technology for the travel industry, Alan is a mentor and advisor in the Dublin startup scene and you can even find him occasionally at an all-night hackathon.
The Role of DMN and BPMN in the Design of Composite Materials by Dario Campagna, Carlos Kavka and Alessandro Turco, ESTECO SpA
We present a new paradigm for composite material design as implemented in the COMPOSELECTOR H2020 project. The project vision is to integrate material modeling methodologies and business process management practices into a comprehensive framework for decision making: a Business Decision Support System (BDSS).
The BDSS integrates materials modeling, business tools and databases into a single workflow. This is achieved by means of an open integration platform which enables interoperability and information management of material models and data, and connects a materials modeling layer with industry standard business process and decision models.
The BDSS goal is to support the complex decision process involved in the selection and design of polymer-matrix composites. In particular, the BDSS can contribute to a better informed decision-making process in the early development stages of products. The BDSS is built on three “pillar” technologies. Namely: material and process modeling; business decision system; data integration. The three pillars reflect on the BDSS architecture main modules: (1) the business layer, (2) the database and workflow manager, (3) the interoperability platform.
In this session we focus on the Business Layer of the BDSS and its implementation. We present the industrial application cases driving the BDSS development and show it in action. Application cases includes: material and process selection of fuselage thermoplastic frame (Airbus), material and process selection of a leaf-spring (Dow Chemicals), material and process selection of passenger tire platform with fixed carcass (Goodyear). For each case we present an overview of the underlying business processes and decisions. A decision for material model selection is used as a guiding example to show how we model business decisions using BPMN and DMN. With the same example we highlight the decision and process management features of the BDSS.
Keywords: DMN, BPMN, Business Decision Support System, Composite material design, Materials modeling.
Dario Campagna received his Master Degree in Computer Science in 2008 from the University of Udine, Italy. In 2012 he received his PhD in Mathematics and Computer Science from the University of Perugia, Italy. The same year he started working at ESTECO SpA, an independent technology provider that delivers software solutions to the engineering industry. Dario spent four years as developer for the Research and Development group. He took part to different research projects and worked in the field of business process management. In 2016 Dario took the role of Agile Coach in ESTECO while continuing to contribute to Research and Development activities as Senior Researcher. Since 2017 he is involved in the COMPOSELECTOR H2020 project, where he contributes to the modeling of processes and decisions underlying the industrial application cases.
Carlos Kavka has a PhD in Computer Science from the University of Paris Sud (France). He is currently the Head of Research and Development of ESTECO SpA (Italy), a company specialized in multi-objective and multi-disciplinary design optimization. In particular, he is in charge of the coordination of scientific, technical and technological aspects in most of the European research projects in which ESTECO SpA participates. Carlos Kavka has been an Instructor, Lecturer and co-Director of international workshops at the International Center for Theoretical Physics ICTP (Italy) since 1993 till 2012, Professor at the Universidad Nacional de San Luis (Argentina) from 1994 till 2005. He has also participated in research activities at the LHC-CMS CERN experiment from 2005 to 2007.
Alessandro Turco has a PhD in Applied Math from the International School of Advanced Studies (SISSA) of Trieste and a Master in Management from the School of Management of Milan Polytechnic (MIP). He has been working for ESTECO SpA for ten years, starting as a researcher for the Numerical Method Group. He is now the project manager of Cardanit, the BPM solution recently launched by ESTECO.
Meeting the Expectations with DMN and Constraint Solving : The Notary Case by Marjolein Deryck, KU Leuven
DMN proved its usefulness for the analysis and automation of business decision logic. Furthermore, the self-evident syntax of the notation, makes it very suitable as communication tool between technicians and businesses. However, typically the effort put in the analysis, is only used to answer the predefined top question. We explain how we combined DMN with a constraint solver to have access to a larger number of inferences, and get a larger ROI on our analytical effort.
We do this by discussing the registration duties business case. When purchasing a house in Belgium, a notary affirms the process. This registration gives rise to the payment of registration duties, which depend on the region. The standard tax rate can be reduced for certain houses, which leads to a range of possible tax rates with their associated conditions. As the legislation at hand is very complex, the notary office wants support during the client interview.
In this presentation we show how we started to address this problem with the use of DMN. First, we gradually built an executable model. The user-friendly DMN acted both as an analytical and communication aid. Afterwards, we translated the decision tables into the first-order logic language IDP. By plugging the program into an existing IDP web-interface, a user friendly and interactive prototype resulted. Even though this prototype provided some basic interactive characteristics, it missed some essential functionalities, like explainability. Therefore we further developed prototypes to include instantaneous impact propagation, highlighting relevant variables and explanations to meet the notary’s needs.
Keywords: DMN, Constraint solving, case study, knowledge representation.
After more than 10 years of experience in the financial sector, Marjolein Deryck obtained an additional master in Management Information Systems (Business Engineering ) in 2016. Afterwards she started her doctoral research on knowledge elicitation and decision automation. In her current project she combines the advantages of DMN with inferences offered by the knowledge-based system IDP to offer solutions in real life cases.
Spaghetti Spreadsheets Untangled – Benefits of decision modeling when uncovering complex business logic hidden in spreadsheets by Charlotte Bouvy, M.C. Bouvy Consultancy
The Sociale Verzekeringsbank (SVB) is the organisation that implements national insurance schemes in the Netherlands. The SVB has been modeling the decisions from laws and regulations they execute in their primary process for quite some years, using standards (BPMN, DMN), tools that support these standards (BiZZdesigner, Trisotech) and a rule engine (Oracle Policy Automation) in addition to their more traditional IT-solutions and legacy applications.
At DecisionCAMP 2019 the BRM team of the SVB will share their experiences with transforming the business logic that was hidden in highly knowledge intensive spreadsheets used by their operational control and audit departments, into explicit declarative decision and calculation rules. Some of the difficulties encountered were the dependency on external expertise, and the typical intertwining of data, formulas and decisions. These spreadsheets are now successfully replaced by an automated decision and calculation service, containing business rules that are understandable and maintainable by the SVB.
The use case is the statistical analysis of the lawfullness of the execution of the payment schemes: did the SVB pay every client the amount he or she was entitled to? The operational control and audit department of the SVB have to account periodically towards the ministries for which they execute millions of payments every month. To ‘calculate’ lawfullness the SVB uses advanced statistical estimation and sampling.
Transforming these statistical calculations into business rules means carefully choosing the right terms for the many different concepts: they need to be accurate, consistent and meaningful to the business people. It also involves decomposing the often deeply embedded formulas into reusable intermediate results. The creation of a logical data model, with several entity levels with containment relationships, supports expressing the many summations and other aggregations in this domain.
Keywords: explicit knowledge representation, declarative modeling of business logic, decision and calculation rules.
Charlotte Bouvy is a business rules specialist with a knowledge engineering and artificial intelligence background. In her twenty-plus year career she has assisted numerous organizations in various industries with all facets of business rule management. From analyzing and modeling business logic, building executable rule models, knowledge elicitation, requirements management, to developing guidelines and patterns that support business people to model their business logic. She also conducts trainings in rule modeling tools and methodologies.
Her special interest lies in business architecture approaches that bridge business and IT, that relate rules to decisions and data in the primary business processes, and that improve business agility.
Combining DMN, First Order Logic and Machine Learning: The creation of Saint-Gobain Seals’ Digital Engineer by Nicholas Decleyre, Marjolein Deryck and Bram Aerts, Saint-Gobain and KU Leuven
Saint-Gobain is an international construction and high performance materials manufacturer present in 67 countries with more than 180k employees. Having numerous, globally spread product engineers among these, gives rise to multiple challenges, including standardizing the worldwide product offering and filling up the void upon retirement of experts.
The Digital Engineer project is a digitization initiative by the Saint-Gobain Seals department to capture, standardize and automate the different decisions when designing high performance seals. The goal is to develop a tool that provides fully-defined solutions for standard cases and gives support in developing more complex designs, while at the same time enabling the integration of new knowledge from research and experience.
In order to meet these demands, a threefold architecture was developed:
-In the rules-based approach, Saint-Gobain uses the DMN standard to develop a set of executable decision tables to select the standard, fitting seal design for well-known applications.
-To assist the engineer in lesser known application fields, the underlying knowledge of the previous rules is centralized in a knowledge base under the form of first order logic constraints. The engine supports a multitude of inference tasks on the this knowledge base, which allows the engineer to perform a variety of queries.
-Apart from extracting the knowledge directly from the global engineering team, a subproject is devoted to the use of ML techniques for rules- and constraints-mining from historic documents and drawings.
Since the engineering task is only part of the larger manufacturing process, the developed system needs to be integrated with other systems so results and conclusions can be communicated and processed directly. Examples of such linked systems are the ERP-system and software to automatically generate the correct technical drawing.
Keywords: Decision Modeling, Knowledge Representation, Machine Learning, Manufacturing Automation.
Nicholas Decleyre graduated as IT engineer from KU Leuven in 2018 with the master thesis called “A rule-based system for seal design”. This thesis was performed as part of the Digital Engineer project running at the seals business unit of Saint-Gobain Performance Plastics. His work won the Willy Asselman Award for excellent theses in the category “Business Impact”. Nicholas joined Saint-Gobain afterwards as a project engineer to follow up on and expand the Digital Engineer project.
Bram Aerts is a doctoral researcher at the KU Leuven University. His research situates in the field of Knowledge Representation and Artificial intelligence, more specifically applying state-of-the-art techniques in industrial applications.
Exploiting payroll knowledge with Viren
by Tim Stuyckens, Teal Partners
Working in Belgium can be complex. We have six governments and are the second highest-taxed country in Europe. In the Global Payroll Complexity Index report, Belgium ranks third – only preceded by Italy and France – and so it can be challenging to figure out which legal statute is most beneficial for an individual.
In this presentation, we will show how we have modeled three different gross-to-net calculations in our Viren platform, and with that created a powerful simulator that allows users to analyze the net year income and social security coverage of an employee, business owner or sole trader.
Viren is a high performant and flexible cloud platform that allows domain experts to configure the latest legislation. It tracks all calculations and legislation changes and allows any software application to consume that knowledge, using a web API. The expressiveness of our modeling language allows us to have a very short time to market, while the built-in testing functionality and transparency features of Viren ensure the correctness of all calculations.
This talk will show the general concepts and how these contributed to the success of our simulator. Although we started this project with a top-down gross-to-net calculation, we saw all sorts of exciting new possibilities for our customer. For example: “how many days should one work at which daily rate to get to a yearly net income of a certain amount?” Together with the AI department of the University of Leuven, we have found ways to maximally leverage our knowledge model in order to quickly answer those questions.
Tim Stuyckens has been building enterprise software products in .NET for ten years now. He started as a junior developer and gradually became more and more interested in the relation between a software product and a business.
Since he joined Teal Partners 3 years ago, he has worked on the Viren team as a project lead and helped growing tools on top of the platform. He strongly admires Conways law – organizations design systems which are copies of their communication structures – and dreams of a world where all software is built by companies with happy people.
A hybrid implementation of multi-channel, multi-modal, high volume financial risk monitoring by Martijn Tromm and Marten Schokking, Oracle
A real life case study of a dutch bank’s implementation of rule based risk detection across multiple channels, that can be used to gather data, take on-demand decisions in process steps and classify millions of customers in batch.
We will show how the same rules are applied in different modes of operation, enabling projections of rule and data changes by reevaluating customers using either given answers or data from internal and external systems that has been updated and thus potentially resulting in new risk classifications.
The whole system works as a feedback loop: customers’ assessments that are reassessed using given interview data and updated system data may show higher risk classifications, thereby triggering new full assessments leading to updated interview data.
By adding new data streams, replacing interview data for internal and external system data, the bank is working towards a learning and fully automated financial risk monitor.
Additionally, the intention is to have a subset of the indicators of the risk model be determined by machine learning techniques. And vice versa, machine learning can be aided by feeding back employee knowledge (interview data).
Furthermore, the same rules also determine what data from the customer is gathered and are extended with interaction rules to aid efficient and consistent data entry.
Keywords: compliance, financial risk, risk monitoring, rules engine, business rules, customer due diligence, machine learning
Martijn Tromm spent 24 years in IT including development, testing, analysis, design, architecture and requirements engineering in knowledge intensive systems of varying types such as research and data analysis marketing databases, a business support system for telecom network disturbances, public eligibility checks for building permits, content management system for legislation and legal publications, rule based compliance monitoring in tax and financial risk. Main expertise in business rules and knowledge engineering.
Marten Schokking involved in analysis, specification and implementation of all systems around customer processes, including Siebel and rule based tax reporting and compliance monitoring since 2013.
Decision Management as a Service by Dennis Aarts, The Business Analysts
The digital processing of services within the Government of Flanders has a very diverse maturity. In concrete terms, this means that the government still request information from applicants (for a grant) that they already have at their disposal. Furthermore, a lot of repetitive and manual work is being performed, that could be perfectly automated. The project “Automatisch advies” delivers a generic platform to provide advice based on automated services and questions. This is done based on data retrieved from (authentic) sources and/or the application of business rules.
The project “Automatisch advies” focuses on two aspects of the case handling process, namely the composition of an application and the proposition for a decision. In concrete terms: enriching a case with data that is available within the government, so that an applicant only has to provide the missing information; and automatically generating a proposal for a decision with regard to the objective criteria, whereby the ultimate responsibility for the decision remains with the subsidizing entity.
The project is being realized with a so-called orchestration layer (Camunda) and a business rules engine (IBM ODM). The orchestration layer receives the request with limited information from an external application. It than accesses the (authentic) sources to complete the request. This completed request is sent to the business rules engine, which provides a proposal for a decision that is sent back to the orchestration layer. The orchestration layer finally responds the external application with the enriched data and the proposal for a decision, including the reasoning behind this decision.
Up to this point, the project seems like a default application of a business rules engine. However, the big difference with other applications is, that a centralized entity within the Government of Flanders (Informatie Vlaanderen) realizes the project and provides the resulting product as a service. This comes with a lot of benefits that would not have been realized by the individual entities:
• Partition of costs: As may be obvious, the costs of the platform can be divided over the entities that make use of it. This allows certain entities to use a business rules engine, whereas this would be financially impossible should they have to buy one alone.
• No in-house expertise required: Besides the product itself, Informatie Vlaanderen provides the set-up of the rule sets as well. This means, that the entities don’t have to recruit the expertise for data enrichment and the definition of business rules themselves.
• Reuse of business rules: A large number of business rules can be reused for different services. By providing the business rules engine centrally, there is no need to redefine these rules multiple times. This also ensures that the rules are applied consistently cross-entity.
Obviously, such a centralized platform also comes with a lot of difficulties that need to be tackled:
• How to actually divide the costs over the different entities?
• How to apply governance? Allow entities to operate mostly independent, without giving them the ability to break both their own rule sets as well as those of other entities.
• Who is responsible for the rules that are being used cross-entity?
During the presentation I will give an overview of the project and dive deeper into both the benefits and the difficulties that come with the offering of decision management as a service. This gives insight in the possibilities to work together with multiple entities, which we currently see too little.
Keywords: Decision management as a service, Reusing business rules cross-entity, Governance, IBM ODM, Camunda.
After a study in the Netherlands, Dennis Aarts came to Belgium, where decision management was far less mainstream. In Belgium, Dennis realized multiple analysis in terms of decision management, yet always followed by a hard coded implementation. Right now, he supports different customers with the purchase of a business rules engine and guides their analysts in the definition of business rules. The Business Analysts, part of the Cronos Groep, is a company focusing on functional and business analysis. Due to an expanded interest in the domain of Enterprise Architecture, the subsidiary Bespoke Services was founded in 2017. For the same reason a second subsidiary, called B. Adapted, focusing on change management, was founded in 2018. Currently The Business Analysts and its subsidiaries have about 90 employees working in a broad range of market segments for both governmental and privately held organizations.