Intelligent Automation with Blue Prism

4

Reviewing Predictions and Human in the Loop

One of the biggest differences between RPA and IA lies in how previously certain automated work outcomes can become uncertain. We’re no longer 100% confident that the Digital Worker has taken the correct course of action due to uncertainty in the ML prediction. The main way to get a sense of how well the ML algorithm is performing, and to reduce the overall risk levels of IA, is to perform a partial manual review of the ML predictions. This chapter discusses two different ways to design this manual verification through adding human in the loop (HITL) into the automated process.

In this chapter, we’re going to cover the following main topics:

• Why should we review predictions?
• What does HITL mean ...