Google has launched a new set of tools designed to help businesses connect their AI agents directly to data stored in BigQuery. This move comes as more companies seek to develop smarter, more autonomous AI applications that can perform tasks without constant human input. These agentic applications are becoming popular because they allow organizations to do more with fewer resources, making data access and accuracy critical.

What the New Toolset Offers

The new toolset from Google includes several features that let AI agents run queries inside BigQuery and gather metadata about datasets and tables. For example, the list_dataset_ids tool helps find all dataset IDs within a Google Cloud project. The get_dataset_info tool provides detailed info about a specific dataset, while list_table_ids lists all tables within a dataset. The get_table_info tool fetches details about individual tables, and the execute_sql tool lets users run SQL commands directly in BigQuery to get results.

This collection of tools is designed to give AI agents a clearer understanding of the data they are working with. By providing access to metadata and the ability to execute queries, these tools help improve the accuracy and relevance of responses generated by AI applications.

How to Use the Toolset with Google’s Frameworks

However, the toolset alone isn’t enough. Companies need to use it alongside Google’s open-source Agent Development Kit (ADK) and MCP Toolbox for Databases, which was formerly called the Generative AI Toolbox for Databases. This combination is necessary to connect AI agents to BigQuery effectively.

To assign the toolset to an AI agent, organizations must import it from the agents.tools module within a Python environment. They can do this using Google’s command line interface (CLI) or SDK. Google also offers a feature called the tool_filter parameter, which allows users to select specific tools to expose to their AI agents, giving more control over what functions are available.

The MCP Toolbox for Databases supports BigQuery’s pre-built tools natively. To access these, companies need to create a folder called mcp-toolbox in their project directory and install the toolbox. There’s also an option to define custom tools in SQL, giving organizations flexibility in how they tailor their data interactions.

Industry Impact and Future Plans

Experts believe that Google’s integration of these tools will speed up the development of more advanced agentic applications. Charlie Dai, a vice president at Forrester, notes that having pre-built frameworks from Google reduces the amount of custom work needed. This makes it easier for companies to build AI that can access enterprise data accurately and efficiently.

Google isn’t the only player in this space. Recently, other data platform providers like Databricks, Snowflake, and Teradata have introduced similar offerings to help enterprises connect their AI agents to data stored in data lakes and databases. This competition encourages faster innovation and more options for businesses looking to enhance their AI capabilities.

Google has mentioned plans to add more tools to this set, but no specific timeline has been shared yet. As more features roll out, expect this integration to become even more seamless, helping companies unlock the full potential of their data with AI-driven insights.

Inspired by

• https://www.infoworld.com/article/4031238/googles-new-toolset-to-help-connect-ai-agents-to-bigquery.html

Sources

• gmpg.org
• github.io
• databricks.com
• snowflake.com
• podigee.com