Engineered AI for Google Ads.
Free, open-sourced skills I use in production across 118 Google Ads accounts. Each one is a guardrail, framework, or workflow that makes AI reliable for PPC work. Drop any of them into your own Claude Code setup.
Step-by-step walkthrough to get Google Ads API access working the first time. The prerequisite for every other PPC AI skill.
Two-step approval system for Google Ads AI mutations. Dry run, human review, then execute.
Classify Google Ads search terms into four intent categories. Judgment-based, context-aware, designed for AI reliability at scale.
Three independent AI classification passes with consensus voting. The production version of search query classification that I actually trust.
Upload approved negative keywords to Google Ads shared lists across hundreds of accounts unattended. The last mile of SQR automation.
Mandatory verification protocol for AI-generated ad copy. Every claim sourced from the website. Empty fields over unverified claims.
Replace LOW-performing RSA assets with website-verified headlines and descriptions. The ad copy cycle that doesn't hallucinate.
Diagnose Google Ads underspend through the three IS metrics — Search, Budget Lost, Rank Lost — in a Smart Bidding world.
5-tier triage framework for multi-account managers. Answers the daily question: which account do I look at first?
Reusable GAQL query templates for the Google Ads API. Skip the syntax trial-and-error and start from queries that actually run.
Hypothesis-driven framework for diagnosing Google Ads performance issues. The process that separates senior strategist analysis from generic AI output.
Audit and fix Performance Max asset automation settings at scale. Opt out of Google's auto-generated headlines, images, and URL expansions.