From BI team to data platform team: What changes when everyone can ask data questions with AI.
For most data teams, self-service has always had a ceiling: analysts who knew the model could explore freely, everyone else waited on a dashboard or filed a ticket. AI removes that ceiling, but when anyone can query the warehouse in plain language, the data team's job changes from delivering BI to running a platform.
Ed, Director of Data at Synthesia walks through what that looks like in practice with Omni:
Curating the semantic layer around how the business actually works
Keeping topics small enough to make the wrong question hard to ask
Feedback loops that raise a PR when an answer looks off, so fixes land in the model instead of a Slack thread
Benchmarking correctness against Snowflake to prove the answers hold up
If you're rolling out AI broadly and worried about correctness, governance, and support load, this is the operating model behind becoming a platform team without the chaos.