We were awarded yet another patent on Decision Logic Elicitation. This is awesome!
As much as it is an ego boost to receive these acknowledgements, I would like to explain why we have been investing continuously on elicitation. After all, most of the research happened in the 1980’s and we are part of a very small group that has continued exploring new ways to simplify this task
Decision management system have demonstrated over and over that they could:
- automate operational decisions in large volumes
- reduce errors and inconsistencies
- refocus expert personnel on value-added tasks, and complex decisions
- allow companies to update their decision-making strategies quickly and efficiently
- improve the bottom-line as a result of all these benefits
So why aren’t there more projects using decision management technologies?
My hypothesis is that harvesting knowledge remains a heavy investment. Once the business rules are know, projects can be implemented in a matter of days or weeks. The hard part is to know what to put into these systems.
The current state of the art is to rely on consultants or decision modeling experts to interview subject matter experts, or to review manuals and other artifacts. In the worst case, I have seen teams of developers reverse engineering COBOL or C code to find out what the rules are.
The good news is that we now have standards like DMN (Decision Modeling and Notation) that help structure decision knowledge. Tools supporting the DMN standard, like Pencil, can support these efforts.
I am a strong believer that, while absolutely valuable, modeling tools do not address the elicitation challenge completely. We will not be able to get rid of the challenge of authoring completely by using modeling tools.
Elicitation challenges
The reason that the investment on elicitation research has been low over the years is that it is a hard problem. How do you make it easier for domain experts, that are not decision management gurus, to structure their know-how into business rules that are complete and accurate?
This is certainly a challenge that we have explored in very different ways.
Should we give them building blocks that are safe to use, yet powerful enough to express all of their logic? This is an interesting path, but we, as an industry, struggle with the level of granularity and expressivity that should go into these building blocks. If we knew from day 1 what rules would need to be implemented, we could design the right building blocks. If you have ever built Lego models on your own or with your kids, you know that each Lego box contains a fair number of custom pieces for that Star Wars or Ninjago model. The problem of anticipation is the catch-22 we have been battling with for a while.
Should we not let business experts author business rules? This is, in my mind, the worst solution. By removing the direct interaction of the business user with the system, you are removing the main benefit of the technology. Adding a layer of translation slows down the effort and introduces potential translation errors. It sets us back to the time of the hard-coded rules!
Why elicitation matters
That brings me back to the title of this post. Why does elicitation matter?
The research that we conducted, and that lead us to the two patents we have at Sparkling Logic in decision logic elicitation. They cover the two main aspects of elicitation.
The first and most obvious challenge is to make it easy for experts to express business rules in the context of what they know. What they know is the business, and that translates into knowing what a business transaction is, and how to deal with it. Our initial research focused primarily on how to express simply and naturally rules, overlaid with data. Not only can you see what the rule is, but you can execute your decision logic and check that it produces the right results.
The second patent we worked on, and that was just awarded last December, is about understanding the context. Some of it has to deal with metrics such as ‘which field is playing a role in my decision logic’, or ‘how often are rules referencing this field or that field’. There is a lot to be said about ‘understanding the decision logic’ and yet it has been ignored. It is natural that understanding is not that important if you struggle to write your rules in the first place, but, assuming that we found a way to address that first problem, understanding becomes crucial.
After all, you are not changing business rules for the sake of changing them… You are changing them to improve your decision logic.
Learn more about Decision Management and Sparkling Logic’s SMARTS™ Data-Powered Decision Manager