PPC & Ads · Blog post · source Aug 28, 2026
Three Repeatable PPC Jobs with an AI Agent and Scale Insights MCP
Point a listing at winnable demand, promote proven terms and sweep for wasted spend
“organic revenue per day rose by 130% and 84%”
Per source From the tool's vendor, not independently tested.
Copy-paste prompt
Using my search term report and my targeting report for the last 60 days: 1. List search terms where I convert better than the campaign average but they don't have their own exact-match target yet. 2. List every negative keyword that blocks a search term which has orders in the last 12 months, and flag single-word negatives. 3. List search terms with 15+ clicks and zero orders. 4. For each product, show price, estimated margin after fees and ad cost of sales, and flag any where ads can't be profitable at current margin. Propose changes only; I'll approve them row by row.
The source didn't publish its exact prompt. We wrote this starter so you can try the same workflow. Replace anything in [brackets].
What they did
Kevan Soh, founder of Everloop and an 8-figure multi-brand Amazon seller, presented at a Scale Insights event how he runs three repeatable agent jobs from the same two reports pulled live through the Scale Insights MCP: one points his listing at demand he can win, one turns terms he already beats the market on into candidates for their own campaigns, and one sweeps weekly for spend with nothing to show for it. The agent reads and proposes; he approves every change row by row, and writes back to Amazon through his own setup.
How it went
Across three of his products, organic revenue per day rose by 130% and 84% on the two that already had organic sales, and from nothing to a third of sales on the third, while ad cost of sales fell on every one. On one product, conversion went from 12% to 17.7%, ad cost from 20.5% to 7.3%, and daily ad spend from $27 to $13. The agent also found 230 of his own negative keywords blocking good search terms (205 of them single words), and a product selling at $5.99 on 31 cents of margin.
We only report what the source says. Results are self-reported unless noted.
Worth knowing
Written by Scale Insights about its own MCP, and it says plainly that results like these are never down to a single tool. The post doesn't name the AI client outright (it says he 'hasn't handed the keys to Claude'). The negative-keyword finding is worth checking in any account.