Automated Intelligent Assistance with Explainable Decision Models in Knowledge-Intensive Processes.

Predictive monitoring techniques increasingly contain explainability aspects to instill trust in the end-users. However, it currently remains difficult to clearly communicate to the end-user how and why a certain outcome is obtained in a knowledge-intensive process. One improvement is the representation of decisions using executable Decision Model and Notation (DMN) models. These allow to automatically construct an intelligent assistant that reasons directly with any DMN model to provide such explanations. This paper examines the added value of a generic intelligent assistant, based on a DMN model of a decision. A preliminary experiment was conducted using two different explanation sources (text and intelligent assistant) to evaluate the explanation facilities of an automated DMN intelligent assistant. The first findings from this ongoing research provide insights into how organizations could easily provide stakeholders with explainable decisions in processes.