An app wrote a reason next to your supplement. Who is liable for it?
Personalisation tools increasingly write a line of text next to a product: why this one, for this shopper. If that line drifts into a health claim, the question is who answers for it. The short answer is that it is probably you.
- A brand can be liable for claims made on its behalf, even ones it did not write or review.
- Structure/function claims are allowed. Disease claims are not, and plain wording is no defence.
- Claims need substantiation before they are made, not after someone asks.
- So the question for any vendor is not "is it accurate" but "what is it allowed to say, and who checked".
Not legal advice. This is a summary for someone making a decision without a lawyer in the room. Your own counsel and your regulator have the final word.
The FTC’s guidance for health products is the reference point, and the distinction it draws is not intuitive.
| This shape | Status | Note |
|---|---|---|
| Supports normal cardiovascular function | Structure/function claim | Permitted, with substantiation |
| Helps unclog arteries | Disease claim | Not permitted. Lay wording does not make it one |
| Supports normal sleep | Structure/function claim | Permitted, with substantiation |
| Treats insomnia | Disease claim | Not permitted |
The FTC also expects "competent and reliable scientific evidence" to exist before a claim is made. And a brand can be responsible for claims made on its behalf by someone else — the endorsement rules reach further than most people assume.
The published guidance addresses what a brand says on its own pages and what an endorser says. It does not address generated or personalised text. We looked; the compliance guides that rank for supplement claims do not cover it, and the app round-ups for supplement stores do not mention compliance at all.
That gap is not a free pass. It means nobody has set the norm yet, so the obligation falls back on the brand.
- 1Show me the actual sentences your tool produces, for twenty of my products.
- 2Are they generated per shopper, or drawn from a fixed set?
- 3Who wrote the rules about what they may not say, and can I read them?
- 4Can I edit or veto the wording, per product?
- 5What stops the text naming a condition?
- Keep a record of what the tool said and when
- Review the wording the way you review your own PDP copy
- Ask for the text to stay non-clinical and name no condition
- Assume a vendor’s compliance badge covers your claims
- Let the tool reference a condition, even helpfully
- Ship text you have not read at least a sample of
The reason shown beside each product is written for the shopper, is deliberately non-clinical, and names no condition. It appears on your surface, so you can read it by using your own store.
We say the honest part too: it is a sentence, not a code. That is exactly why our terms prohibit using the output to infer health status, and why the wording is held to a fixed, reviewable standard rather than invented freely per shopper.
Can we edit the reason text?
The wording is held to a non-clinical standard rather than being freely editable per shopper, because editable claim text is how a compliant system becomes a non-compliant one. Tell us where it reads wrong for your category and that is a product conversation.
Does this apply outside the US?
The specifics here are FTC. UK and EU rules on health claims for foods and supplements are separate and in places stricter. The principle — substantiate before you claim, and you answer for claims made for you — holds in both.
Is a recommendation itself a claim?
Ranking a product is not the same as claiming an effect. The risk sits in the sentence next to it, which is why that sentence is the thing to check.