Product Research · Podcast episode · source Jun 15, 2026
Product Validation Agents Built with Claude Code
Check product ideas against your research criteria in bulk
“now takes less than five minutes”
Per source Self-reported by the source, not independently verified.
Copy-paste prompt
Here are my 23 product criteria with pass/fail thresholds: [list]. For each ASIN in the attached list, pull the data for each criterion, mark pass or fail with the number, and give an overall verdict. Put anything that fails only one criterion in a separate 'review' list instead of rejecting it.
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
A Singapore-based seller with multiple seven figures on Amazon US validates products against 23 data points, including how competitors rank for keywords. His team uses Claude Code to build the automated processes and agents that run this research for them.
How it went
He says validating one product used to take one or two hours and now takes less than five minutes, so they validate in bulk; the agents report weekly and a person still reviews each product before it's chosen.
We only report what the source says. Results are self-reported unless noted.
Worth knowing
He also sells Amazon coaching, and he gives no build details. Keeping a person on the final pick is right: the agents shortlist, you decide.