The claims you cannot make
Supplement advertising is regulated as a claims problem first and a creative problem second. The claims that cause enforcement action are consistent across markets.
Disease claims. Saying or implying that a product treats, prevents, cures or mitigates a disease moves it out of supplement advertising and into a category it is not authorised for. This includes the softer versions: “clinically proven to reverse”, “helps your body fight infection”, “for people with diagnosed…”. A supplement video that makes one of these claims is not a risky ad, it is an unapproved drug claim in video form.
Guaranteed results. “Lose 10 pounds in two weeks”, “works for everyone”, “results guaranteed” — any promise of an outcome the product cannot substantiate. Even where a study exists, the ad cannot generalise beyond what the study supports.
Implied medical authority. A presenter in a white coat, a stethoscope on the counter, a caption that reads like a diagnosis. The visual carries the claim even when the script is careful, which is why reviewing only the words is not enough.
Fabricated proof. Invented testimonials, invented before-and-after images, invented customer counts. In this category a fabricated review is both an advertising violation and, once AI is involved, a disclosure problem.
The common thread: none of these are creative choices that went too far. They are claims that would fail the same test in a static listing image.
Structure/function language and why the merchant approves the script
The legitimate space for supplement advertising is structure and function: how a nutrient or an ingredient works in the body, and how a product fits into a routine. “Supports normal energy metabolism” is a structure/function statement. “Cures fatigue” is a disease claim. The distance between those two sentences is the entire discipline of the category.
Three habits keep a script on the right side of that line.
Describe the routine, not the outcome. A morning routine with the product in it is a scene. A transformation is a claim. Scenes are cheap to produce and safe to run.
Talk about ingredients, not conditions. What is in the product and why it is in there is factual territory. Who should take it for what condition is not.
Let the merchant own the final wording. This is the part that matters most, and it is why the merchant approves the script in writing before production begins. A production partner can build a careful script; only the brand knows what its own substantiation file supports, what the label already says, and what the local regulator in each market will accept.
That approval step is not bureaucracy. It is the moment the claim becomes the brand’s claim — and the moment someone accountable has read it.
AI disclosure and platform labels
The second rulebook is transparency. If a video contains AI-generated or substantially modified content, viewers should be able to tell. In practice that means three things for a supplement brand:
- A disclosure line where the viewer is looking, not buried in a description.
- The platform’s own AI label set on every variant of the ad, since Meta and TikTok label assets individually.
- A consistent rule for what the brand says when a customer, a retailer or a platform asks how the video was made.
There is a category-specific wrinkle worth naming. Supplement audiences are unusually alert to authenticity — the whole creator economy in this niche runs on “real person, real routine”. A synthetic presenter performing a personal transformation story would be both a disclosure failure and a trust failure, and the second one is harder to recover from. A synthetic presenter explaining an ingredient panel is a different proposition entirely.
The practical rule we apply: AI can present the product, and AI cannot claim to be a customer.
Provenance documentation
Provenance is the paper trail: what was generated, from what inputs, and with what consent.
For a supplement brand this answers a question that arrives from an unexpected direction — a marketplace compliance team, a retailer’s vendor review, a platform’s ad review, or a distributor’s legal team in an EU market. “How was this video produced, and were the assets licensed” is a question that either has a document behind it or does not.
HausTrue’s default position: AI-assisted production is flagged, a provenance note is available on request, merchant assets are not passed to third-party AI services without consent, and an AI-free workflow is available at the quote stage. An AI-free workflow matters more in this category than in most, because some brands would rather not have the conversation at all.
If the product is sold in Europe, the disclosure and the provenance note are the same conversation as the market entry itself. Building both into the first production is cheaper than adding them to a campaign that has already run.
How a compliant project runs at HausTrue
The workflow has one deliberate gate, and it sits before any production time is spent.
Brief. You send product assets, the label, the routine you want shown, and any claim wording your team requires. If you have a script, it goes in the brief.
Concept. Within 48 hours you receive a script pass and a visual direction. This is where claim language gets fixed: the presenter’s lines, the on-screen text, the caption copy, and the disclosure line. Nothing is produced until you approve this in writing.
Production. The video is built to the approved script. One revision round on Pilot, two on Starter. A revision that adds a new claim reopens the approval step — which is the point.
Delivery. Masters in the agreed formats, watermark-free, with worldwide perpetual commercial rights, plus the provenance note on request.
The Pilot exists so this can be tested on one video: $149, fifteen seconds, your script, five business days. If the presenter reads credibly for your category and the claims survive your own review, the monthly tiers are a volume decision rather than a leap of faith.
What we will not do is produce a script we believe makes a disease claim and let the platform catch it. That is not caution for its own sake — a pulled ad costs more than the video did, and an account strike costs more than that.