PPC & Ads · Blog post · source The case study page shows no date.
Agency Wasted-Spend Checks and Unit Forecasts with Claude and Intentwise
Find wasted spend across client accounts and build forecasts without gathering reports
“a bottoms-up unit forecast for all of the agency's Amazon brands in 10 minutes”
Per source From the tool's vendor, not independently tested.
Copy-paste prompt
For each client account, list every campaign where average CPC in the last 7 days is more than double its 30-day average. For each one, check: placement bid adjustments (top of search, product pages), and any non-brand search terms showing up in campaigns named as branded. Give me a table of campaign, CPC change, likely cause and the fix.
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
Austen Cayne, Head of Amazon at the agency Motif Digital, first uploaded Amazon reports into Claude by hand, but found the answers imprecise and spent 15-30+ minutes per client gathering reports for each question. With Intentwise's AI Gateway MCP connecting the data, he asks Claude diagnostic questions across client accounts, has it build dashboards and forecasts, and runs workflows that build custom client decks.
How it went
On Prime Day, some clients saw CPCs as high as $30, from a faulty placement modifier or a non-brand term pulled into a branded campaign, and the case study says Claude located those sources and the reasons. Claude built a bottoms-up unit forecast for all of the agency's Amazon brands in 10 minutes, and the agency saves 3-5 hours per week on building reports.
We only report what the source says. Results are self-reported unless noted.
Worth knowing
A vendor case study from Intentwise about its own product, with no date on the page. The $30 CPC causes (placement modifiers, non-brand terms in brand campaigns) are worth checking in any account.