As of 13 August 2026, AI can only partly choose the best creative for your paid ads.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, power-user skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the relevant ad platform and analytics reports, then choose one campaign or test period and record the primary outcome you actually care about, such as purchases or qualified leads.
   2. Export one row per creative with its ID, format, spend, impressions, clicks, conversions, conversion value, audience, placement, attribution window and dates, then remove personal data and note any tracking changes.
   3. Gather the corresponding image or video files, headlines, primary text, calls to action and landing-page links so the model can compare the creative itself rather than metrics alone.
   4. Paste the context, creative table and creative descriptions into a chatbot with the supplied prompt, then save its data-quality warnings, ranking, confidence and proposed test plan.
   5. Compare every reported figure with the original platform export and analytics records, correct any calculation or definition error, and reject recommendations based on non-comparable audiences, placements or conversion windows.
   6. Apply the recommendation only as a controlled test with a pre-set success metric and stopping rule, leaving the other relevant campaign conditions unchanged before deciding whether to keep or pause a creative.

   Prompt

   You are helping me choose which paid-ad creative to keep testing, not pretending to prove a winner. Analyse the data and creative information below.
   
   Business and campaign context:
   - Business: [business and offer]
   - Campaign objective: [objective]
   - Primary success metric: [for example, qualified leads or purchases]
   - Target audience: [audience]
   - Platforms and placements: [platforms and placements]
   - Attribution window and conversion definition: [details]
   - Important constraints: [brand, compliance, budget or production constraints]
   
   Creative data:
   Paste one row per creative with its ID, format, message or hook, audience, placement, spend, impressions, reach, frequency, clicks, landing-page views, conversions, conversion value and the dates covered. Include the creative file, a link, or a faithful description for each ID.
   
   Tasks:
   1. Check whether the rows are comparable. List missing fields, inconsistent definitions, obvious tracking problems and reasons not to compare particular creatives.
   2. Assess each creative against the primary metric, while also reporting relevant secondary metrics such as click-through rate, cost per click, conversion rate, cost per conversion and conversion value where the supplied data supports them. Do not invent figures or calculate a metric when its denominator is missing.
   3. Separate observed results from possible explanations. Do not claim that a creative caused the result if the data cannot establish that.
   4. Recommend one creative to keep testing, one to pause only if the evidence supports that action, or state that there is no defensible winner yet. Explain the decision in plain English.
   5. Give me a short test plan covering the comparison, the success rule, what must stay constant and when to reassess.
   6. State exactly which checks I must make in the ad platform and analytics system before acting.
   
   Use only the information I provide. Treat missing or thin data as a reason for lower confidence, not as permission to guess.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.