Run a lead-gen Google Ads campaign with a daily AI review
The two co-founders of consultancy Biggest Goal used Claude with read-only Google Ads access, plus Claude Code for tracking scripts, to build and monitor a lead-gen campaign. They say it worked better than anything they had run before.
What they did
They gave Claude their offers, messaging, design style and past work to shape the campaign; Claude exported CSVs that they previewed and pushed live in the Google Ads desktop app themselves. Claude Code wrote the JavaScript event tracking for Google Analytics, and Claude helped score leads logged to their CRM. A daily routine pulls Google Ads, Analytics, Microsoft Clarity heat maps and lead data and posts a recommendation to a private Slack channel.
What happened
The hosts say the campaign worked better than anything they had run before. Heat maps showed visitors copying answers from an advertorial page and leaving, and after about 300 leads they found students and job seekers scored as qualified, so they simplified scoring and split campaigns.
Keep in mind
Claude proposes and humans decide; it had no write access to the ad account. No cost, conversion or lead-quality figures are given, their own qualification assumptions were wrong at first, and the episode promotes their workshops.
Try it with your agent
Each morning, using read-only access to <Google Ads account>, <analytics property> and <lead sheet>, compare yesterday's spend, clicks, leads and lead quality by campaign. Point out pages where visitors leave without acting and leads that look unqualified for <your ideal customer>. Post three recommended changes with your reasoning to <Slack channel>; do not change anything in the account.
Replace the <angle brackets> with your own details.