What was the result?
self-reportedAgency co-founder Alexandre Almeida says work that took 4 to 8 hours every week now takes 30 minutes after building five Google Ads skills for Claude Code. He also says they saved clients thousands in wasted spend, but gives no exact amount or measurement period, so this play leads only with the stated weekly time comparison.
Alexandre Almeida of Beyond Six describes five reusable Claude Code skills: a search-term analyzer, product-page audit, responsive-search-ad copywriter, product-feed audit and custom skill builder. His thread says the analyzer classifies intent and outputs a negative-keyword list, while the feed audit uses the real Google taxonomy file. He offers a free repository through an engagement request, but no repository was obtained or installed here. The useful pattern is fixed checklists and reference data for recurring analysis, with a human controlling actual ad-account changes.
Who did this?
How do I try it?
- Editorial adaptation of his five-skill pattern, not a tested copy of his unpublished-to-us repository: start with a read-only search-term export, campaign goals and approved keyword policies.
- Write a reusable analysis checklist that labels commercial, informational and branded intent and flags possible waste with the supporting spend/conversion rows.
- Keep product-page checks, ad-copy drafting and product-feed checks as separate tasks, with reference files such as your brand rules and the current official category taxonomy.
- Ask for a review table and proposed changes, not live edits. Verify brand exclusions, conversion lag and commercial intent with the account manager.
- Time the weekly analysis and review end to end. Apply only explicitly approved changes through your normal platform process.
Editorial adaptation, not Alexandre Almeida's exact skill or prompt. Using only the attached read-only search-term export and campaign context, classify query intent and propose negative keywords with supporting cost/conversion rows and uncertainty. Do not block brand, competitor or broad terms without an explicit rationale. Separate recommendations from applied changes. Provide a manager-review table; do not access or modify the live ad account, raise budgets or publish ads. Ask for any missing attribution window or brand rules.
Make it run every week
A play works once. Turn it into a routine in three small steps:
- Run it once and check the result yourself.
- Give your AI a schedule. Tell it when to run this play (every week, or every month) and to message you with what it found before you act.
- Keep a log. One line per run: what was checked, anything flagged, what you did.
What should I watch out for?
Self-reported agency operator workflow, not a vendor case study or independently timed test. Weekly 4 to 8 hours to 30 minutes covers his combined workflow, not one audit or an assured saving for every user. Exact client-spend savings are unverified and omitted from the headline. Thread has a free-resource engagement CTA; no need to reply/repost to use the general pattern. Skills are supported today; the exact repository and account connector were not inspected or tested. Build it with Page: a client approval page for proposed campaign changes.
Did it work for you?
PageMotor keeps this site free.