This page lists three turns, the messages you send to the AI one after another in the same chat, and then a check of the answer. Keep it open next to your chat. Each turn shows the next thing to paste, with the values you entered already filled in. Nothing is sent from this page, and your conversation stays in the AI tool you are using.
Headline, subhead, body, image text, testimonials, and every disclosure, including the fine print.
Determines whether a disclosure can physically sit near the claim it qualifies.
Everything offered as substantiation, with study design, sample size, and dates where you know them.
Which regime governs, plus who else could come after this copy.
Sets whose takeaway defines the implied claim, and which overlay rules apply.
Watch for these throughout
The model does not know current enforcement priorities or what any agency has recently done. Treat every regulatory example it offers as a lead to check, not as authority.
Naming or clearly implying a competitor adds a private plaintiff with a real incentive to sue, and a self-regulatory challenge that moves far faster than an agency would.
An "up to" number without typical results disclosed near it is a recurring enforcement theme. Check both the number and where the qualifier physically sits on the page.
Regulated categories (health, safety, children, earnings, environmental, and capability claims about automated systems) carry overlays this review does not cover. Escalate rather than approving them yourself.
Approving copy without seeing the underlying study is how legal ends up owning a claim it never verified. Ask for the document, not the summary of the document.