Team Compliance is the account-wide legal brief generators read before they write copy. Every brand and campaign on the team inherits these rules. A brand or campaign can add more rules. They cannot remove or weaken what you set here.
Open Settings → Compliance. Saving requires permission to manage team settings.
The page subtitle is Account-wide legal and compliance rules for generated content. A second line explains the cascade: brands and campaigns can add rules on top, but they can never remove or weaken the rules defined here.
These rules steer generation. They do not replace a legal review of the finished ad.
If the page shows Compliance rules could not be loaded, do not assume the team has no rules. Choose Reload and try again. The editor stays hidden on a failed read so a save cannot overwrite the real policy with a blank form.
Required disclaimers are lines you instruct the generator to include word-for-word. The model is asked to use your exact wording. That is guidance, not a guarantee — review the finished ad.
Type a line, then press Enter. Each line becomes a chip. Choose × to remove one.
Limits you can save: each line up to 500 characters, up to 100 lines. Blank entries are dropped on save. Keep the list short. The generator only receives a limited legal block (about 4,000 characters for account, brand, and campaign rules combined, including labels). Account lines are sent first. If the block fills, later brand and campaign lines are dropped; a very long account list can lose its last lines too. Saving does not warn you when that happens.
Use this for AI disclosure, paid-partnership language, “results not typical,” age ratings, and other lines counsel wants on every ad.
Example set (book / publisher):
This content was created with the assistance of AI.Dramatization. Book interior shown is from the actual published edition.#ad or Paid partnership with [Brand]Results not typical. Individual results may vary.Reading age 14+. Contains mature themes.Other verticals:
These statements have not been evaluated by the FDA. This product is not intended to diagnose, treat, cure, or prevent any disease.Not financial advice. Investing involves risk, including loss of principal.Write the exact string you want in the ad. Do not rely on the model to invent the legal phrasing.
Prohibited claims are things you instruct the generator never to say or imply. Same caveat: review the finished ad. The model is not blocked from publishing if it misses a rule.
Same chip editor: type a line, press Enter, remove with ×. Same save limits (500 / 100) and the same shared 4,000-character legal block.
Write the banned idea in plain language. One claim per line.
Example set (book / publisher):
That the reader or customer is speaking from personal experience"#1 bestseller" or any ranking claim unless the specific list and date are in the briefThat the book is endorsed by any celebrity, author, or publicationHealth, medical, or therapeutic benefits of any kindGuaranteed outcomes — "guaranteed", "risk-free", "never fails"Price or discount claims not present in the campaign briefOther verticals:
Curing, treating, or preventing any medical conditionSpecific return percentages or "guaranteed" incomeCompliance policy is the markdown box for standing rules that do not fit the two lists — fictional spokespersons, product imagery, competitor mentions, how to treat superlatives.
Placeholder: Add legal or compliance policy that applies to all generated content… Up to 20,000 characters.
Keep it as durable policy, not a one-off campaign brief. Campaigns can add extra legal lines later.
Example posture (book / publisher, adapt to your counsel):
All AI-generated spokespersons must be clearly fictional — never depict or imitate a real, identifiable person. UGC-style ads must not fabricate ownership or usage experiences; use “imagine…” or observational framing instead of first-person testimony. Product imagery must show the actual product — never invented packaging, invented interior pages, or altered covers. Do not reference competitors by name. All claims about the product must be verifiable from the product description provided in the brief; when in doubt, omit the claim. Comparative superlatives (“best”, “greatest”) only as clearly subjective opinion, never as fact claims.
Saved rules apply to later generations on every brand in this team. They do not rewrite ads that already exist.
| Level | Where | What it can do |
|---|---|---|
| Account | Settings → Compliance | Sets the floor for the whole team |
| Brand | Brand Settings → Brand Guidelines, then expand Advanced guidance → Legal & Compliance | Adds brand-only disclaimers, claims, and policy. Inherited account rules are read-only |
| Campaign | Campaign strategy section Legal | Adds disclaimers, claims, and policy for that campaign. Cannot remove account or brand rules |
The same disclaimer or prohibited-claim line is sent to the generator once (account first, then brand, then campaign). Compliance policy text is not merged that way — identical policy at two levels is sent twice, labeled by origin.
See Brand guidelines and website data.