Data Cloud connectors cover BigQuery, Spanner, AlloyDB, Cloud SQL and Looker. Gemini uses Knowledge Catalog to find relevant data, map everyday terms such as product categories onto database structures, generate SQL, and run it through the connector. Data is queried where it sits through federated query mode rather than copied into a Gemini data store.
Administrators configure default read-only connections in the Gemini admin console, and end users enable the connector to build reports. The BigQuery connector also reaches open table formats such as Apache Iceberg, Delta Lake and Parquet, along with datasets in AWS or Azure and AWS S3 through the AWS Glue federated catalog.
Alongside the connectors, Conversational Analytics agents are scoped agents built and certified by data teams for governed, multi-step analysis. Gemini also surfaces published agents next to the connectors, so users can choose between ad hoc exploration and a predefined agent.
Why it matters if you build agents
Developers building agents on Google Cloud data can use first-party connectors with managed authentication, which removes the need to deploy MCP middleware or configure their own OAuth client for those services. Query costs can be controlled with a default billing project set by administrators. Because the connectors are in preview, teams should check the setup documentation for each engine before relying on them.
Get stories like this every weekday
One short email with the day's agent tooling news.