Flag irrelevant paid search terms for the negative keyword list
Run the Google Ads search terms report through Jev to flag queries that do not match what you sell, for a person to add as negatives.
The use case
Anyone running Google Ads pays for clicks on queries that have nothing to do with their product, and reading the search terms report each week is slow. Each query plus a short description of the business goes to Jev with one question: is this search from someone who could buy what we sell? Queries with a high probability of no go onto a review list. A person approves which ones become negative keywords, because a wrong exclusion silently cuts off real buyers. The same pipeline from this build also scores landing pages and mines long-running ads, but the search term check is the easiest place to start.
Example workflow
- Export the search terms report from Google Ads each week
- Add a short product and audience description as shared context
- Ask Jev a Noul question per query about purchase relevance
- Sort queries with a low yes probability into a review tab in Google Sheets
- A person approves negatives and adds them in Google Ads
Tools
- Jev (TypeSafe AI)
- Google Ads
- Google Sheets
Risks and guardrails
- Excluding a term by mistake can quietly block buyers
- Brand and competitor queries need their own rules
- Short queries carry little context, so confidence may be low
Sources
Written by Agents Lookup. Credits and official pages for the tools this use case names:
- Winning patterns out of the Meta Ad Library by Ysato on Made with Jev: madewithjev.com/builds/ad-library-winning-patterns
- Winning patterns out of the Meta Ad Library: original post or repo by Ysato: x.com/Ysato_su04/status/2102199755250229334
- Google Ads MCP server (official page): github.com/googleads/google-ads-mcp (checked 2026-11-09)