Free editable template

Meta ads audit prompt pack: 25 prompts for Claude, ChatGPT and Grok

Replace "how are my ads doing?" with 25 narrow questions an AI can actually answer from your data. Each prompt names its export (level, date range, breakdown, columns), carries your target cost per purchase and break-even ROAS, and tells the AI to use only the file, show its sums and stop if a column is missing. Works in Claude, ChatGPT and Grok.

By GaaS editorial · Sources checked · 8 worksheet rows

Open the CSV in Excel, or import it into Google Sheets and choose comma-separated values. Save your working copy in a controlled location. The CSV contains the rows below and space for your notes; it has no macros, formulas or live account connection. The separate Excel workbook contains editable example inputs and formulas that update the forecast.

How to use the template

Fill in Settings once

Your business in a few words, target cost per purchase, break-even ROAS, target ROAS, the export's date range and attribution setting. Every prompt on the Prompts tab rewrites itself from these.

Pull the export the prompt names

Ten of the 25 prompts run on one ad-level export, so start there. The Exports tab gives the steps for each kind: ad level, breakdowns, ad set level, by day, and the three that need screenshots or Shopify data.

Ask, one prompt per message

Copy a grey Full prompt cell into Claude, ChatGPT or Grok with the export attached. Keep the same chat so prompt 25 can summarise the rest. The AI only reads the file; it can't see or change your account.

Check the sum, then log it

Paste the ad-level columns into the Check tab and compare its totals with the AI's answers to prompts 1 and 8. Log each finding with the dollars a month at stake; unchecked rows stay amber.

Preview the worksheet

Every row shown here is included in the download. Scroll the table horizontally on a small screen.

#QuestionExport it needs
1Rank every ad by spend with 0 purchasesAd level, last 30 days
2Which ads are past 4.0 frequency and losing CTR?Ad level, 8 weeks, by week
3Where is the dead money by placement?Placement breakdown, 30 days
4Which ad sets never left learning?Ad set level, 7 days, Delivery column
8What share of spend sits on 3 ads?Ad level, last 30 days
12What is my real blended ROAS?Campaign level by day, plus Shopify sales by day
20Which ads to scale tomorrow?Ad level, last 7 days
25Write my audit summary for my bossYour earlier answers, same chat

Worked review example

Fictional 12-ad account, $20,000 spent in 30 days, target cost per purchase $40, break-even ROAS 2.5: the Check tab finds 3 ads that spent $2,140 with zero purchases (prompt 1) and 62% of spend on the top 3 ads (prompt 8). Account ROAS 2.63, cost per purchase $45.15.

Findings log: zero-purchase ads $2,140 a month (checked, done), Audience Network at 9% of spend with no purchases $1,800 (to check), Sunday cost per purchase $61 against $38 on weekdays (to check whether there are enough Sunday purchases to trust it).

What this template does and does not establish

The AI only sees the file you attach, and AIs make arithmetic mistakes: check the number a decision rests on. Ads Manager results depend on the attribution setting and won't match store orders exactly. The thresholds in the prompts are our starting points, not Meta rules. Small samples mislead. Don't paste customer personal data into an AI. No account connection, macros or automatic changes.

Primary references

Continue the work