What it does: Reads your Google Ads search terms report, finds the searches that cost money and never turned into a lead, and drafts a negative keyword list for you to add.
Use it when: Your Google Ads spend goes up but the leads don't, and you want to know where the money went.
You get: A 3 line summary, a negative keyword table with match types, and up to 10 keywords to think about adding. Nothing changes in Google Ads.
You are my search ads lead, a senior Google Ads manager with 15 years of cutting wasted spend for small businesses. Goal: read my Google Ads search terms report, find the searches that cost me money and never turned into a lead, and draft a negative keyword list, so I have a short list I can add myself and know where my money went. Context: I run a small business and sell [what you sell] in [where you sell]. Every click costs me. My real buyers want to find me fast, and I need my budget there for them, because clicks from people looking for jobs, free stuff, or something I don't sell eat it first. First, check that you can read the search terms report I gave you (a CSV downloaded from Google Ads for the last 30 days, with search term, campaign, clicks, cost and conversions). Tell me how many search terms and the total cost you found. If it's missing or a column is gone, stop and ask me to download it again or paste it instead. Ask me these, then run. Skip any I already answered. 1. What do I sell and what don't I sell? (Default: read it from the search terms with the most conversions, and show me your guess first.) 2. How much spend with no conversion makes a term worth cutting? (Default: 2 times my average cost per conversion in this report.) Do this: 1. Add up the cost, clicks and conversions. Check your total cost matches the report. 2. Sort the search terms by cost, highest first. 3. Mark each term with no conversions and spend over my line as a cut maybe. 4. Read the words in each cut maybe. Put it in a group: job seekers, free or cheap, a service I don't sell, wrong place, a competitor's name, or other. If the negative word you would use is also in a term that converted, put that term under Check by hand, with a narrower word if there is one. 5. For each group, write the negative keyword in the shortest words that block it, and say if it should be phrase or exact match. Check it won't block a term that converted. 6. Find the terms that converted at or under my average cost per conversion and aren't in my keywords yet. List them as keywords to think about adding. 7. Show me the lists. Don't add, pause or change anything in Google Ads. I will make each change myself after I say yes. Example row, match this: "plumber jobs near me | 14 clicks | $38 | 0 conversions | job seekers | negative: jobs, phrase match". Give me: a 3 line summary (total spend, spend on terms with no conversions, how much of it the new negatives would cover), the negative keyword table (term, clicks, cost, conversions, group, negative, match type), and up to 10 keywords to think about adding. Before you show me, check: - Your totals match the report to the cent. - No negative keyword would block a search term that converted. - Every number in the table comes from the report. Rules: - Drafts only. Never change a bid, budget, keyword or ad in Google Ads without my yes. You give me the list, I make the change. - Never promise more leads, lower cost or any result. Say what the report shows, nothing more. - Never make up a number, search term or keyword. Use only the report and what I tell you. - If something is missing, write [check: what is missing] in that spot and keep going. Last, save these steps as a skill called 'Search Terms Cleanup' in Claude or ChatGPT so you do it this way each time I check my Google Ads search terms. In Google Ads, save the search terms report so it's one click next time. To run it again, just say: "Run my Search Terms Cleanup skill." Run it every 2 weeks.
You need![]()
Your files
Google Ads
![]()
![]()
Any AITested on sample data
Connect first: your files. The steps for each AI are on one page. Then the prompt reads it for you. It ends by saving itself as a skill.
Built on Fanatical Prospecting's rule to know your numbers.
Copilot version
There is no separate Copilot version of this one yet. Paste the prompt above into Microsoft 365 Copilot Chat.
See a real example
Tested on sample data Sample data run (simulated tool).
Watch: no video for this one yet.
AuthorCorey Leon