Give your AI agent direct access to your Google Ads campaigns
n8n exposed as an MCP server with two read-only operations on Google Ads — the reusable pattern for opening any internal system up to an AI agent.
What it solves
When an AI agent needs to query external data, the cleanest way isn’t to hand it a report — it’s to give it direct access to the source via MCP (Model Context Protocol): a standard that lets an agent use external tools as if they were its own functions. This workflow turns n8n into an MCP server that exposes two read-only operations on Google Ads.
How it’s built
The flow goes from credential to agent call, with no intermediate steps:
- MCP server — n8n publishes an MCP server URL that connects to the Google Ads account through an OAuth credential configured once.
- MCP client — that URL is pasted into the configuration of any MCP-compatible AI agent — one built with Claude, or n8n’s own agent.
- On-demand call — when the agent needs campaign data — to answer a question, generate a report or make a decision — it calls the matching tool and n8n returns the real Google Ads data in the moment.
The two exposed tools are list all campaigns in an account and get the detail of a specific campaign by its ID. Both are read-only: the agent can’t modify anything, only read.
Why it’s interesting
The beauty of this approach is that the agent decides on its own, based on the conversation, whether it needs to list all campaigns or pull the detail of a specific one — no one has to pre-program it. It’s a small but reusable piece: the same pattern works for exposing any other system (CRM, ERP, in-house database) to an AI agent without writing custom integration code. Once the MCP server is in place, opening more data sources is just a matter of adding more tools.
The workflow, in the open
The full workflow is published on GitHub: you can download it, import it into your own n8n and see exactly how it’s built, node by node → google-ads-mcp-server.json