I managed to get out of bed and down to the conference in time for James Taylor’s 7am presentation on advanced decisioning. If you’ve been reading here for a while, you know that I’m a big proponent of using decisioning in the context of processes, and James sums up the reasons why: it makes your processes simpler, smarter and more agile.
Simpler: If you build all of your rules and decisioning logic within your processes – essentially turning your process map into a decision tree – then your processes will very quickly become completely unreadable. Separating decisions from the process map, allowing them to become the driver for the process or available at specific points within the process, makes the process itself simpler
More agile: If you don’t put your decisioning in your processes, then you may have written it in code, either in legacy systems or in new code that you create just to support these decisions. In other words, you tried to write your own decisioning system in some format, but probably created something that’s much harder to change than if you’re using a rules management system to build your decisions. Furthermore, decisions typically change more frequently than processes; consider a process like insurance underwriting, where the basic flow rarely changes, but the rules that are applied and the decisions made at each step may change frequently due to company policy or regulatory changes. Using decision management not only allows for easier modification of the rules and decisions, it also allows these to be changed without changing the processes. This is key, since many BPMS don’t easily allow for processes that are already in progress to be easily changed: that nice graphical process modeler that they show you will make changes to the process model for process instances created after that point, but don’t impact in-flight instances. If a decision management system is called at specific points in a process, it will use the correct version of the rules and decisions at that point in time, not the point at which the process was instantiated.
Smarter: This is where analytics comes into play, with knowledge about processes fed into the decisioning in order to make better decisions in an automated fashion. Having more information about your processes increases the likelihood that you can implement straight-through processes with no human intervention. This is not just about automating decisions based on some initial data: it’s using the analytics that you continue to gather about the processes to feed into those decisions in order to constantly improve them. In other words, apply analytics to make decisions smarter and make more automated decisions.
To wrap up James’ five core principles of decisioning:
- Identify, separate and manage decisions
- Use business rules to define decisions
- Analytics to make decisions smarter
- No answer is static
- Decision-making is a process
He then walked through the steps to apply advanced decisioning, starting with identifying and automating the current manual decisions in the process, then applying analytics to constantly optimize those decisions.
He closed with an action plan for moving to decisioning:
- Identify your decisions
- Adopt decisioning technology
- Think about decisions and processes, and how those can be managed as separate entities.
Good presentation as always – well worth getting up early.