Google Cloud has introduced Data Agent Kit, a preview product that embeds AI-powered data investigation directly into development environments like VS Code, allowing developers to query databases, data warehouses, and storage systems without leaving their code editor. The toolkit combines AI agents with the Model Context Protocol, an open standard for connecting software to tools and databases, enabling natural language queries that can span multiple data sources in a single workflow.

What Problem Does Data Agent Kit Actually Solve?

Data practitioners spend considerable time moving between separate consoles, databases, and storage systems to answer business questions. Data Agent Kit reduces that friction by letting developers ask open-ended questions in plain language and having an AI agent execute queries across multiple systems on their behalf. The agent can query BigQuery, inspect Cloud SQL databases, and examine files in Cloud Storage, then return results directly in the development environment.

Google Cloud illustrated the toolkit with a real-world retail scenario. A company noticed that average order value had dropped from about $110 to $103 in January, even though overall revenue remained flat. Using Data Agent Kit, an analyst asked the agent to investigate. The agent queried BigQuery, discovered the decline, then broke down the numbers by order type. It found that a new B2B-Wholesale channel had launched in January with an average order value of about $75, dragging down the blended average. The investigation then moved to Cloud SQL, where the agent examined customer records and discovered all 100 wholesale accounts were newly created business entities added within the previous 30 days. Finally, the agent traced the issue to a marketing campaign in Cloud Storage offering a 25% discount on a promo code used in 92% of B2B orders.

How Does Data Agent Kit Work in Practice?

• Natural Language Queries: Developers type questions in plain English within their IDE, and the agent translates them into queries across multiple data systems without requiring manual SQL writing.
• Multi-System Investigation: The toolkit can query BigQuery data warehouses, Cloud SQL production databases, and raw JSON files in Cloud Storage within a single workflow, eliminating the need to switch between separate consoles.
• Approval-Based Execution: Before running actions such as read-only SQL queries, the development environment pauses to request permission, giving users control over what the agent does.
• Audit Trail Visibility: Users can inspect the execution trail in their IDE, including individual Model Context Protocol tool calls and raw SQL sent to BigQuery, enabling teams to review and audit agent activity.

The toolkit is available as an extension for VS Code forks, including Antigravity IDE and Cursor, and as a plugin for tools including Antigravity 2.0, Antigravity CLI, Claude Code, and Codex.

Can Data Agent Kit Turn One-Off Investigations Into Reusable Projects?

Google Cloud positioned Data Agent Kit not just as a query tool but as a way to transform ad-hoc investigations into repeatable data projects. In one example, after identifying the wholesale channel issue, a user asked the agent to build a dbt project, a data transformation framework, that joined BigQuery staging models with Cloud SQL customer and pet profile attributes. The user also requested a uniqueness test on order IDs and asked the agent to run the build.

The first build failed because customers could own multiple pets, causing duplicate order rows when pet profiles were attached directly to each order. The agent reviewed the terminal output, rewrote the dbt logic, and reran the build until the test passed. This example highlights both the appeal and the limits of the approach. The software may cut the time spent writing repetitive queries across multiple systems, but users still need to review generated code and apply data quality checks to the output.

Google Cloud emphasized visibility into agent activity as a core feature. The company stated that users can inspect generated queries because "It's important to keep an eye on generated code, though reading a query can take much less time than writing one against schemas you're unfamiliar with".

Google Cloud

Why Is This Launch Significant for Cloud Providers?

The introduction of Data Agent Kit reflects a broader industry shift toward embedding AI agents into developer workflows and data operations rather than limiting them to chat-based assistance. For cloud providers like Google, the opportunity lies in linking natural language interfaces to the databases, object stores, transformation frameworks, and operational systems that businesses already use. By making these systems accessible through natural language queries within the development environment, cloud providers can deepen integration with their customers' workflows.

Google Cloud is positioning Data Agent Kit around the practical work of investigation: tracing a business anomaly across warehouse data, production databases, and unstructured files, then turning the result into a reusable model within the same working session. Its preview status suggests the company is still testing that workflow with developers before a broader rollout.