What to expect

Building an AI agent that works in a notebook is easy. Running one reliably in production is a different challenge. Agents need to survive crashes, retry on failure, and maintain state across restarts.

In this session, we’ll build a documentation agent using Dapr Agents’ DurableAgent. The agent analyzes code changes in a pull request, generates or updates documentation accordingly, and submits a docs PR. The entire process runs as a durable workflow, providing automatic retries, state persistence, and workflow recovery.

Along the way, we’ll walk through the development loop for defining the agent’s prompt and tools, using Dapr’s workflow engine for durable execution, adding retry policies, giving the agent memory with Dapr state management, debugging and iterating on the agent’s behavior, and packaging it as a GitHub Action triggered by pull request events.

What you’ll see in the demo:

• Building a documentation agent using Dapr Agents’ DurableAgent
• An agent that analyzes code changes in a pull request
• Automatic documentation generation and updates based on those changes
• Durable workflow execution with retries and workflow recovery
• Agent memory using Dapr state management
• Packaging the agent as a GitHub Action triggered by PR events

Who this webinar is for:

• Developers building AI agents or internal developer tooling
• Platform and infrastructure engineers exploring agentic automation
• Teams experimenting with agents who want to run them reliably in production

Save Your Seat