An AI chatbot can make a product claim a merchant never authorized โ a real regulatory risk for health, wellness, supplement, and skincare brands. Flayr grounds every answer in what a store has actually published, backed by an explicit "never claim" boundary and a live-preview test before anything ships.
An AI chatbot that answers questions on its own carries a specific risk most merchants don't think about until it's a problem: it can make a claim about your product that you never authorized. For most product categories, an overstated claim is an awkward correction. For health, wellness, supplement, and skincare brands, an overstated claim can be a regulatory issue. The FTC and FDA both hold the brand responsible for what a chatbot says on its storefront, the same way they'd hold you responsible for a claim printed on packaging.
Flayr is built to prevent that from happening in the first place.
Every answer comes from what you actually published
Before Flayr's AI talks to a single shopper, it reads your store. Not a general model of what an ecommerce site probably says โ your actual homepage, your actual shipping and returns pages, your actual product catalog, and any additional pages you point it to, like an ingredients list or a sizing guide.
That becomes the source the AI answers from. If your product page doesn't say something, the AI doesn't say it either.
A dedicated guardrail for product claims
Inside the Product Knowledge section, alongside your categories and best sellers, there's a field called Never Claim. This is where you tell the AI directly what it must never say about your products, regardless of how a shopper phrases the question.
For a supplement brand, that might mean the AI never implies a product treats, cures, or prevents a condition. For a skincare brand, it might mean no claims about medical-grade results without the certification to back it. You set the boundary once, and it applies to every conversation the AI has from that point forward.
This matters because shoppers ask leading questions. "Will this help my anxiety?" "Is this safe during pregnancy?" A generic chatbot might answer helpfully and inaccurately at the same time. Flayr's AI is built to stay inside the boundary you've set, even when the question invites it to step outside one.
Legal and compliance, read directly
The Legal, Compliance & Restrictions section pulls your required disclaimers and the key points from your terms of service directly from your site. When a shopper asks something that touches on a legal or regulatory point, the answer reflects what you've actually published, not a paraphrase built from general knowledge of similar products.
It stays accurate as your policies change
Compliance requirements shift. A claim that was fine last year might not be fine after a label change or a new regulation. Auto Sync checks your site on a regular basis and re-analyzes automatically when something changes, so an outdated claim doesn't linger in the AI's answers after you've already corrected it on your site.
Test it before a shopper ever sees it
Every field in the Critical Store Info section is visible and editable. Nothing here runs as a black box. Before publishing, you can open the live preview and ask the AI the exact questions a shopper is likely to ask, including the ones designed to test a boundary. If an answer isn't right, you go back to the relevant field, correct it, save, and test again.
This step is the difference between hoping an AI chatbot behaves and confirming it does.
When to bring in a human instead
Some questions shouldn't be answered by AI at all โ medical questions in particular. The Customer Service Ops section lets you set specific escalation triggers, so a question that touches on health, safety, or anything outside what you're comfortable automating gets handed to a real person in the same conversation, instead of getting an AI-generated answer it shouldn't have given.
The result
A shopper asking about your product gets an answer grounded in what you've actually published, bounded by claims you've explicitly approved, and handed to a human the moment a question needs one. For a health, wellness, or supplement brand, that's the difference between an AI chat feature that helps you sell and one that puts you at risk every time it talks to a customer.