Data Agents Over Mirrored Databases: Natural-Language Querying in Fabric

Microsoft introduced Data Agent capabilities over mirrored databases, supporting natural language querying over mirrored datasets. In this session, we’ll prototype a mirrored database setup, connect the data agent, and explore how NL queries get translated into SQL or unified queries.

We’ll also discuss latency, consistency, caching, and how semantic models factor in, and brainstorm best practices for user adoption, prompt design, and training.