Guides
Connecting AI to Google Ads
Everything written down, in the order it tends to be needed: what the thing is, how to wire it up, and what to check first.
What it is
The written guides
What an AI agent for Google Ads actually does, where the limits of autonomy are, and what a hosted server exposes.
- AI agent for Google Ads
- Definition of an AI agent for Google Ads, how it differs from Smart Bidding and automated recommendations, the read/write boundary, and what to check before granting one access.
- Autonomous agents
- Levels of autonomy in Google Ads management, which optimisation work is safe unattended versus which needs approval, and why an action that cannot be walked back should not be automated.
- Google Ads automation
- The layers of Google Ads automation from Smart Bidding to scripts to AI agents, what each layer decides, what none of them will do, and what to automate first.
- AI Google Ads audit
- How to audit a Google Ads account with an AI assistant: the checks that matter, the exact questions to ask, wasted-spend patterns, and which of Google's own recommendations to distrust.
- Negative keywords
- How to automate negative keywords: the three negative match types, account, campaign, ad group and shared list placement, the intent patterns worth blocking, and how an AI connector applies them.
- Performance Max AI
- What is visible and controllable in a Performance Max campaign, what an AI agent can genuinely do including creating PMax campaigns and asset groups, and where it cannot help.
- Daily audit routine
- How to set up a daily Google Ads audit routine in your own AI client's scheduler using the built-in morning-audit playbook, and what the six-check report covers. The routine reads and reports, and AdCopilot proposes each change first.
- Google Ads MCP server
- Whether Google's Google Ads MCP server is read-only (it is), what AdCopilot's server adds — writes, audits and server-side limits — how the tool list is counted, the URL, and connecting Claude or ChatGPT. AdCopilot proposes each change first. For a hard stop, set your AI client to ask before AdCopilot's write tools: Needs approval in Claude; in Cursor, set the run mode to Allowlist and keep them off the allowlist (Cursor's cloud agents never ask).
- GA4 MCP server
- What an AI assistant can do in Google Analytics 4 through AdCopilot (reports, key events, data streams, custom definitions, attribution settings and the Google Ads link), what it cannot do, prompts to try, and how to connect.
- Tag Manager MCP
- What an AI assistant can read and stage in Google Tag Manager through AdCopilot (tags, triggers, variables and folders in a workspace it manages), why it can never publish or delete, prompts to try, and how to connect.
- Search Console MCP
- What an AI assistant can do in Google Search Console through AdCopilot (search performance, URL inspection, sitemaps), what requesting indexing does and does not do, prompts to try, and how to connect.
Getting it working
One address, every client
Step-by-step for the assistant you already use, and what a developer token is for.
- Connect
- One connector URL, added under your own Google sign-in. Claude, ChatGPT, Microsoft Copilot, the command line or any MCP client: connect, add it, ask.
- API access and tokens
- What the Google Ads developer token was and why Google retired it, the four access levels a Google Cloud project now holds, what self-hosting needs, and how AdCopilot connects through its own project.
- Connect Claude
- Add the AdCopilot plugin to Claude, press Connect and approve, then ask about your own Google Ads account in your chat. A custom connector works too.
- Connect ChatGPT
- Turn on developer mode in ChatGPT, create an app with the AdCopilot URL, sign in and approve, then type @AdCopilot to ask about your own Google Ads account.
- Connect Microsoft Copilot
- Add the AdCopilot connector in Microsoft Copilot, sign in with Google, and ask about your own Google Ads account in the chat your team already has open.
- Any MCP client
- One MCP endpoint, added by URL or by one command. Claude Code, Codex, Gemini CLI, Cursor, Windsurf or anything else that speaks the Model Context Protocol.
Before you connect
What to check first
What it can reach, what it cannot do at all, how a team shares it, what it costs, and what we do not claim.
- Security and data
- What AdCopilot can reach, what it cannot do at all, where your Google credential lives, what is written down, and how to withdraw access.
- For agencies and teams
- Teams v1 is live: owner, member and viewer roles, seats and invitations, each member on their own Google sign-in. Agency governance is on the roadmap.
- Every tool, counted
- The connector's tool list, grouped by product and counted with the condition that makes each number true. No delete tools, and a read-only tier on request.
- Changelog
- What changed in AdCopilot, release by release, in plain words: new features, limits and fixes, with the day each reached production.
- Pricing
- Free or from $19 a month, by accounts, operations a day and seats. 5 days on Pro for your first workspace. In US dollars, euros in the EU or rupees in India.
- Book a call
- Up to 30 minutes: an audit on one of your accounts. Book it with more than 9 people, more than 50 Google Ads accounts or $50K or more a month in ad spend.
The library
Written as it came up
The blog, the glossary, the comparisons, the tutorials, the questions and the use cases — every entry a page of its own.
- The AdCopilot blog
- Working notes on running Google Ads with an AI agent — what to automate, what to approve, and what the API actually permits.
- Google Ads + AI glossary
- Every term you meet when an AI starts managing your ads — defined in plain language, with what it means for automation.
- Google Ads AI tools, compared
- AdCopilot next to the alternatives — free self-hosted servers, PPC suites, doing it by hand — with the trade-offs stated plainly.
- Google Ads AI tutorials
- Step-by-step setup for every MCP client, and the working habits that make an ads agent productive in week one.
- Google Ads AI questions, answered
- Direct answers to the questions people actually ask before connecting an AI to a live ad account.
- Google Ads AI use cases
- What an ads agent does for an agency, an in-house team, a founder — concretely, account by account.
For machines
If you are an assistant reading this
Two plain-text files index the site and state its facts once, in a form you can quote.
/llms.txt indexes every page on this site with a one-line summary, and /llms-full.txt is the canonical fact sheet — exact tool counts with their conditions, the access model, the prices, and a plainly stated list of what this product does not claim.