What is a Long-Running Decisionning❓
📝To recall our previous discussion, continuous decisioning is one in which the model's outcome will not be determined with a single iteration (or cycle) of decision execution on a case. In each execution cycle, they transition the case's state from one to another over time until the outcome is finalized.
✨A long-running decision is a special type of continuous decision. There are similarities between continuous decisions and long-running decisions:
🔹They both are multistep decisions, meaning a series of connected decisions are executed on a case
🔹They both implement a state machine, allowing decisions to determine the next available states for a case (they are stateful)
🔹They both require multiple iterations (or cycles) of execution to finalize the decision outcome
🔹They both may or may not determine the outcome successfully. The outcome of a decision can be inconclusive.
🔦However, in the long-running decision, the execution of the decision model can be suspended (or ⏸️paused) temporarily based on some criteria, external events, or other conditions.
⛳Because business decisions are situation-aware, depending on the situation (of the case or its environment), there will be different sets of outcomes for a case. But also, in any given situation, there will be a need for different sets of information, and if they are not available, the decision outcome will become inconclusive.
A long-running decision can help 🫵you in the above scenario. When a decision unit as part of the decision model requires extra information and if they are not yet available, rather than making the decision outcome inconclusive, the execution will be paused, and the decision model will wait until the required information is received, and then, the execution proceeds from that point.
🚩A long-running decision can be implemented by integrating the decision model into other systems, such as workflow management, process management, durable functions, or custom code implementation.
However, any of those approaches introduce a disconnected decisioning experience. In short, no cohesive view to allow modeling everything about the business decision: the decision model, states, transitions, situations and techniques of the information access.
💥A disconnected decisioning experience means all the related behavior to the decision and the case is not encapsulated in one unified form and they are spread across systems and processes. Therefore, to understand cause-and-effect, measure success, define KPIs, and follow the impact of the decisions, you have to jump into multiple systems and databases, look at reports, export information to Excel and compare results, write custom ad-hoc queries, build custom integration points and so on.
💣I firmly believe #DecisionIntelligence Platform or DIP must support long-running decisions out-of-the-box!
Posted here.
Published March 27th, 2024 at 07:30 am
