Summary

This chapter set out to build on the previous 11 – extending some of the ideas that we met earlier in the book and giving you signposts for where to go next.

In this chapter, we swapped out the previous machine learning algorithm for decision trees and saw how this allowed us to explain what is going on within our decision model. Lack of explainability became an issue when we introduced neural networks as more powerful classifiers, but we solved the problem using cutting-edge TrustyAI tools.

We introduced three ways to deploy combined machine learning and rules-based AI models at a high level via Business Central, and in more depth using KIE and Kogito samples, and again using Python and Power Automate workflows. We introduced another method using Red Piranha as a template enterprise solution to build on, including its ability to read entire Excel spreadsheets and apply rules to them.

We looked at the more advanced DRL rule format and saw that it gave us additional...