In this final session of the Predictable AI Placement series, Jay Stewart, Solution Consultant Fellow at Pega, completes the walkthrough of the five AI placement patterns with a deep dive into GenAI Connect.

Using the same compliance and audit scenario introduced throughout the series, Jay shows how GenAI Connect is used to convert free‑form, AI‑generated text into structured case data that can be acted on downstream. In this example, remediation tasks generated during earlier analysis are transformed into a structured list and written back to the case.

You’ll see how GenAI Connect:

• Takes unstructured text produced by a step agent
• Uses an LLM to extract and normalize remediation tasks
• Writes results directly to scalar and list case properties
• Combines LLM output with business rules (such as due‑date calculation)
• Enables remediation tasks to be routed back to the appropriate team to close the loop

The session walks through both design time and runtime:

• Designing GenAI Connect rules in the AI Designer
• Passing step‑agent output as source content
• Defining extraction instructions and target case properties
• Placing GenAI Connect into the case process at the right point in the flow

Jay also explains how GenAI Connect is commonly paired with step agents—where a step agent produces rich analytical output, and GenAI Connect structures that output for operational use.

This session concludes the AI Expert Circle deep‑dive series, which has covered the Application Agent, Document Agent, Step Agent, Assignment Assist Agent, and now GenAI Connect, illustrating how these patterns work together to deliver end‑to‑end, agentic workflows in Pega.