In this session of the Predictable AI Placement series, Jay Stewart, Solution Consultant Fellow at Pega, walks through a deep dive on the Application Agent and how it enables conversational, agentic workflows driven directly by case design.

Using a compliance audit scenario, Jay shows how an application agent can:

• Identify the correct case type through natural conversation
• Read assignments directly from the case design
• Collect required information conversationally
• Progress the case until it is handed off to another user or work queue

The session demonstrates how this all starts in Pega Blueprint, where high‑fidelity case designs are created and previewed using conversational self‑service. You’ll see how importing a blueprint automatically creates an application agent along with case type tools, including descriptions and suggested phrases that help the agent determine the right workflow at runtime.

Jay also shows how the same agent can be:

• Tested immediately in App Studio
• Exposed through an employee portal
• Embedded into web self‑service for external users

Along the way, he highlights how application agents can be extended beyond simple data collection to include business rules, decisioning, Process AI, and step agents—enabling more dynamic and responsive workflows.

This session is part of an ongoing series that explores the five AI placement patterns, with upcoming deep dives into the Doc Agent, Step Agent, Assignment Assist Agent, and GenAI Connect.