Part one of this research described how to test AI-infused applications (AIIAs) based on what we know about testing automatic software and how data scientists test autonomous software. We then presented the integration and performance of the combined automatic and autonomous software (the AIIAs). In this report, we address testing and behavior of AI embedded or integrated in AIIAs and the technology stack needed for automated and continuous testing in an automated delivery pipeline. Large language models (LLMs) bring an additional level of nondeterminism and complexity to testing as we briefly introduce in this report, knowing that we need to do more research on this topic.