🚫NLY #machinelearning #optimization #rules capability alone on its own is not a Open Decision Intelligence Platform (DIP).
🤔Why? The whole promise of DIP is to allow business decisions to be explicitly defined and executed. Composite AI is an essential part of the #decisionintelligence platform because business decisions are complex in nature.
✨Business decisions are composed of many smaller business decisions, and those smaller business decisions themselves are also based on some other business decisions, and so on. Therefore, this hierarchy of business decisions requires multiple #AI techniques to work together in orchestration and create the final outcomes, such as judgment, suggestion, recommendation, the next best action, etc.
As we 👀see here, a holistic decision model representing a complex business decision has a multi-step behavior. It means outcomes of one decision node, will be the input to the other ones. Not only that, each of them can (and should) use different AI techniques such as #rules, #machinelearning #optimization #data, etc., based on their own individual requirements and their nature.
💥For instance, if one of the decision units in the hierarchy is probabilistic in nature, it is fine to use #ml #machinelearning, but if it is deterministic, you should NOT use #machinelearning for that specific decision unit.
💣Based on the nature of individual decision units in the holistic view of a decision model, we should choose the right #AI techniques that satisfy the decision unit's requirement and are suitable for the decision unit based on its nature. This is called the Composite AI technique, which is the core capability of a Open Decision Intelligence Platform or #DIP in short.
Posted here.
Published September 14th, 2024 at 07:30 am
