Can you really let an AI agent work on a Power BI project, or will it confidently break something?

PBIP makes that question worth asking. Reports and semantic models live in files, which means tools like GitHub Copilot can read and edit them. Add the right project context, skills, and MCP connections, and Copilot can do more than answer questions: it can help with real development work.

In this session, we’ll show what that looks like in practice. We’ll use Copilot’s agent mode with PBIP to draft DAX in TMDL, document a model, make bulk edits, review a pull request, and connect to Power BI Desktop and Fabric through MCP.

But “the agent can do it” is not the same as “the agent should do it.” We’ll look at where this workflow saves time, what context makes a difference, and the kinds of confident mistakes you still need to catch.

You will leave with a realistic view of agent-assisted Power BI development and a workflow you can try in your own projects.

Prerequisites: basic Power BI development, PBIP/Git familiarity is helpful.