dbt and the Future of Analytics Engineering | A Conversation with Jason Ganz

Jason Ganz, Senior Manager of Developer Experience and AI, at dbt Labs woke up early with an idea, opened Cursor, and built the first prototype of the dbt MCP server. That proof of concept led to a production server now used across the dbt community.

In this conversation, Adam Azzam, VP of Product at Prefect, and Jason cover:

Designing for the model as a user, not just wrapping your API
The three core use cases: Develop, Discover, Consume
Shipping infrastructure when the landscape changes every 3 months
What "agent experience" actually means
Why organizational data is the hardest context problem

Also covered: the future of analytics engineering in an AI-native world, why semantic layers still matter (but differently than expected), and what makes MCP fundamentally about orchestration, not just tool calling.
---

About the guests:
Adam Azzam - VP of Product at Prefect
Jason Ganz - Senior Manager of Developer Experience & AI at dbt Labs

Links:
dbt MCP Server: https://github.com/dbt-labs/dbt-mcp
dbt Labs: https://getdbt.com
FastMCP: https://gofastmcp.com/
Prefect: https://prefect.io