To make the supplement compliance thing concrete, here's roughly what it looks l
By Ted Gibson · July 24, 2026 · Curated by George's Blog
To make the supplement compliance thing concrete, here's roughly what it looks like in practice now.
A brand we work with has a catalogue built up over several years, written by various people at various times, all optimised to convert. With the new pressure from Amazon on claims matching the supplement facts panel, we needed to know how exposed they were. A couple of years ago that would have been a super tedious week of someone reading every listing line by line against every label.
We did it in an afternoon. AI pulled every claim from every detail page, lined each one up against the product's actual documentation, and flagged the mismatches. The scan itself was almost trivial.
Then the real work started, because it turned up around thirty claims that didn't cleanly match, and every one of those is a decision rather than a delete. Some were clumsy wording that was easy to tidy. Some were claims worth keeping and properly substantiating, because they genuinely set the product apart. A few had to go, even though they were almost certainly helping conversion, because the risk wasn't worth it.
That afternoon is the whole argument for how I think this works now. The machine collapsed a week of grunt work into nothing. What it couldn't do was make thirty judgment calls about risk, evidence, and brand. That part still needs someone who's done it before, and that's exactly where the value has moved.