Monday, 8 AM. Josian walks in and opens Slack. Its client prep is already done. Spend trends and anomalies flagged. Luke had taken care of it overnight.
Request a free auditJosian has a weekly performance review at 9 AM with a major event ticketing platform. The Golden State Warriors ad group is outperforming everything else. He wonders why.
Luke already knows. "Warriors 126 – Clippers 121. NBA Play-In game. That explains everything."
Luke is our new PPC account manager. He works 24/7. He never misses a Monday prep.
Luke is also not human.
He is our PPC Marketing AI Copilot: OpenClaw to run it, Claude to think, Slack to talk to it, and Google Ads MCP to feed it live account data.
We were sold on the idea of trying out OpenClaw, after listening to Claire Vo's intervention in Lenny's podcast.
Luke's stack consists of 4 pieces. It took us about one afternoon to set up, but the result was well worth the hassle.
1. OpenClaw: the operating system. OpenClaw manages Luke's identity, memory, and channel integrations
Install it, run the onboarding wizard, and you have a self-hosted AI gateway running as a daemon on your machine. That's the foundation everything else plugs into.→ Getting started · Onboarding CLI
2. Claude AI: the brain. All reasoning and analysis run through Anthropic's Claude.
Add your Anthropic API key during onboarding. OpenClaw handles the rest — model selection, token routing, and session management.→ Anthropic provider setup
3. Slack: the user interface. Luke lives there. The team talks to him like a colleague.
Create a Slack app, grab the bot token and signing secret, and paste them into OpenClaw's config. Luke appears in your workspace and responds in any DM or channel you invite him to.→ Slack channel setup
4. Google Ads MCP: the data layer. A MCP server connected to our MCC. Luke queries campaigns, keywords, and performance metrics directly. No copy-paste, no exports
Google's official MCP server connects directly to the Google Ads API. Point it at your MCC credentials, register it in OpenClaw, and the agent can query any account; campaigns, keywords, performance data; in plain language.→ Google Ads MCP setup guide · GitHub repo
Done.
With the correct prompt, Luke will audit a full account in under a minute. He will catch anomalies before anyone notices. And last but not least, client prep is done before the team even walks in.
Less time pulling data. More time making decisions.
Reading data is just the start. We're connecting Luke to Equeco's generative AI capabilities — keyword research, ad copy, campaign recommendations. Spot the problem, generate the fix, apply it.
And we're considering making it available to our clients in Slack, Teams, or whatever chat tool they already use. Would you like to speak to Luke? Reach out to us!
