The brand new toolset comprises instruments that can enable AI brokers to execute queries inside BigQuery in addition to fetch metadata from the cloud-based information warehouse.
These instruments embrace the list_dataset_ids device, which retrieves all dataset IDs inside a Google Cloud mission; get_dataset_info device, which offers detailed metadata a couple of particular dataset; list_table_ids device, which lists all desk IDs inside a dataset; get_table_info device, which fetches metadata for particular person tables; and execute_sql device, which permits customers to run SQL queries immediately in BigQuery and retrieve outcomes.
The toolset, nonetheless, can’t be applied in itself, and enterprises want to make use of the toolset together with Google’s open supply choices — Agent Improvement Equipment (ADK) and its MCP Toolbox for Databases, previously referred to as Generative AI Toolbox for Databases, to have the ability to join their brokers to BigQuery.
If an enterprise desires to make use of the ADK, it must assign the toolset to an agent created inside the framework itself.
