Building the interface for AI agents: MCP, the dbt Semantic Layer, and context engineering

🌐 Global friendly sessions on August 19 & 20

AI agents are only as good as the data they can reach. For analytics engineers, that raises a practical question: how do you expose governed, trusted data to an agent without losing the definitions and guardrails you've spent years building?

In this session, Alex Noonan and Benoit Periguard, will walk through the best practices for making your data agent-ready using MCP, the dbt Semantic Layer, and how to think about context engineering. We'll cover the authoring side, how to structure and expose metrics so an agent gets consistent, correct answers and then move into live demos, including a walkthrough of Claude querying data through MCP. We'll close with a Q&A tackling the questions we hear most often from the community.

You'll leave knowing how to turn your dbt project into a reliable foundation for agentic analytics.

Meet the speakers

Alex Noonan

Senior Developer Experience Advocate

dbt Labs

Benoit Perigaud

Staff Developer Experience Advocate

dbt Labs