Just tested something wild with our Claude AI project for Kokomo Botanical Resort.
The experiment: Upload our best-performing Facebook ad creative and see if Claude can tell me WHY it's working based on the image, copy, and our ideal customer profile.
What I gave it:
- Screenshot of the ad image (yoga class at the resort)
- Screenshot of the ad copy
- Screenshot of the CTA
- Simple prompt: "Tell me why this is the best performer"
What Claude extracted (without me explaining anything):
✅ Authentic wellness capture - recognized the yoga class imagery resonates with wellness seekers
✅ Strategic CTA effectiveness - identified scarcity psychology in "Reserve VIP Access"
✅ ICP alignment analysis - explained exactly why this resonates with our target audience
✅ Performance improvement suggestions - spotted cognitive overload in the copy (7 value props crammed into one post)
✅ Timeline concerns - caught that the late 2026 occupancy date creates uncertainty
Here's what blew my mind:
Claude didn't just say "nice ad." It broke down the psychological triggers, identified friction points, and provided actionable improvements - all from analyzing the creative.
The bigger lesson: You can upload ANY ad creative to properly trained AI and get a thorough performance breakdown including:
- Why it resonates with your ICP
- What psychological triggers are working
- Where you're losing people
- How to improve performance
This changes ad testing completely.
Instead of waiting weeks for performance data, you can get AI-powered insights on whether your creative will resonate BEFORE you spend a dollar on ads.
Watch the full walkthrough to see exactly how I set this up and the prompts I used.