In this technological age, artificial intelligence can do almost anything! As long as you have your prompt well-crafted, then you will get your desired results. However, developers face issues regarding access to enterprise data, and it can take a long time before they get what they want. This article explores the newly developed Model Context Protocol (MCP) to help bridge this gap.
You can connect your AI agents to your database using this simple tool
Connecting AI agents to your enterprise data can be a very daunting experience. Sometimes, it might require complex custom integrations or weeks of development. But this can now be done much more easily. With the release of the Model Context Protocol (MCP), which was launched last month, you can now use the BigQuery MCP server to give your AI agent a direct and secure way to analyze data. This MCP server eases your work and makes you focus on developing AI agents.
The building process of AI agents can be aided using this technology. It does this by giving LLM-powered applications access to your analytic data via a set of tools. Also, integrating the BigQuery MCP server with the Agent Development Kit (ADK) is a simple process that involves the Google OAuth authentication method. Integrating other platforms and frameworks, including LangGraph, Claude Code, or other MCP clients, also doesn’t require much work.
How to use BigQuery MCP server with Agent Development Kit (ADK)
This involves a six-step process:
- Prerequisites: This involves setting up the project, necessary settings, and environment.
- Configuration: Enable MCP and the required APIs.
- Load sample dataset
- Create an OAuth client
- Create a Gemini API Key
- Create and test AI agents
These MCP tools have built-in security and observability
Using the Cloud API Registry and Apigee API Hub, developers can locate trusted MCP tools from Google and their organizations. This ability is, however, paired with strict control. Therefore, administrators can only have access via Google Cloud IAM, utilize audit logging for observability, and employ Google Cloud Model Armor to defend against advanced threats, such as indirect prompt rejection, just like this massive industrial AI cloud by NVIDIA and Deutsche Telekom.
In a statement from David Soria Parra, Co-creator of the technology and Member of Technical Staff at Anthropic, he said:
“Google’s support for MCP across such a diverse range of products, combined with their close collaboration on the specification, will help more developers build agentic AI applications. As adoption grows among leading platforms, it brings us closer to agentic AIs that work seamlessly across the tools and services people already use”.
The true potential of agentic AI is shown when your agent can access your entire application stack from containers to your relational databases. Soon, its support for the following services will be updated:
- Projects, Compute, and Storage: This includes Cloud Resource Manager, Cloud Storage, and Cloud Run.
- Database and Analytics: This includes Dataplex Universal Catalog, Cloud SQL, and Spanner.
- Google services: This includes the Developer Knowledge API and the Android Management API.
- Security: Google Security Operations.
MCPs improve the actions of agentic AIs in the long run and make work easier
Following the capabilities of these newly built technologies, the developer’s experience is being shaped to be seamless by creating a bridge that connects intelligent agents with a vast ecosystem of data. The goal is to ensure that AI is useful and actionable.
If these agents can interact with databases, APIs, and real-world services, they can be transformed from passive language processors into able assistants that can retrieve information, execute tasks, and improve work efficiency. This enables developers to build complex, agentic applications without being delayed by custom integrations for every new tool. In the long run, this improves the pace of innovation and deployment.
Artificial Intelligence is making the world a better place. If AI agents have free access to your data, it makes your work much easier. Using tools like the BigQuery MCP server can be helpful, making your agentic AIs have secure access to your database, similar to this Microsoft commitment to AI development in Europe.
