TheBigQuery MCP Toolboxprovides a set of MCP tools designed to simplify the interaction with BigQuerydata. It can be used to query theFCM BigQuery exported data. Thisdocumentation shows how to use the BigQuery MCP Toolbox and an AI agent to queryand analyze this data quickly and effectively.

Set up BigQuery Export

• Make sure your FCM project is configured to export data to BigQuery. Seeunderstanding message deliveryto get started.
• Once enabled,FCM will automatically populate a BigQuery datasetwith message delivery events.

Set up the BigQuery MCP Toolbox

1. Follow the instructions inUse BigQuery with agentsto install and configure the BigQuery MCP toolbox.

Query BigQuery Export Data using an AI agent

Once the setup is complete, your AI agent should be able to list the availabletools. Check your AI agent to make sure the following tools are available:

• execute_sql
• get_dataset_info
• get_table_info
• list_dataset_ids
• list_table_ids

Now, you can use the following example prompts with your AI agent:

• How many FCM notifications were sent in the last 7 days?
• Graph the data by date.
• What are some common errors for missing notifications?

Benefits of using AI agent for data exploration

Some of the benefits that come with using AI agent for data exploration include:

• Accessibility: This lets your users to query data using natural language.
• Easy Visualization: You can use LLMs capabilities to visualize datareturned from BigQuery.

Feedback channels

To provide feedback, contact the following:

• For AI agent response issues, contact the model owner or the AI agent team.
• For BigQuery MCP toolbox issues, contact theGoogle Cloud supportteam.