Use case · negative keywords
Negative-keyword automation that doesn't throw away converters
The boring weekly job everyone skips. NotFair lets your AI agent run it in minutes — group by intent, propose negatives at the right scope, and approve writes inline with safety checks against converters.
Why automated negatives usually go wrong
Naive automation negativizes anything below a CPA threshold — and accidentally kills a future converter. NotFair's workflow filters out terms that have converted, groups by linguistic similarity, and chooses the right scope based on which campaigns the term appears in.
- Filter out converters before proposing
- Group by intent — “free”, “jobs”, “tutorial”, “reviews”
- Choose scope intelligently — shared list for cross-campaign noise, ad-group scope for one-offs
- Match type: phrase or exact, never broad
The prompt that works
“Run a negative-keyword sweep. Find every search term last 30 days with >$5 spend and zero conversions. Skip terms where the root has converted in the last 90 days. Group by intent. Propose negatives at the right scope.” That single prompt collapses what used to be a 90-minute weekly job.
How the approval flow works
The agent prepares a structured proposal: term, group, proposed scope, proposed match type, supporting cost/conversion data. You approve the batch (or per-row); NotFair applies the writes through the Google Ads API and logs the operation.
Connect Google Ads in two minutes
Authorize once at notfair.co. Then open your AI client and ask real questions about your account.
FAQ
Short answers to the most common questions.
Will NotFair negativize a high-volume keyword by mistake?
No — the workflow warns before negativizing any term that has converted in the lookback window. Approval is per-batch and per-row.
Can I undo a negative keyword I approved?
Yes. The operation log captures every write with undo metadata. You can reverse a batch or a single row.
Is NotFair safe to use on a real Google Ads account?
Yes. NotFair separates read tools from write tools. Diagnostic queries run freely, but every write — bid change, negative keyword, ad copy, campaign state — is approval-gated and logged with full provenance so you can audit exactly what the agent did.