Log in / Create free account🌐 ES☀️

Advanced training · Module 19 — ChatGPT Ads — the platform, the market and our diary

Brand, policies and reviews: how to stay out of trouble

⏱️ 6 min read · 🤖 ChatGPT Ads ✍️ Ads · updated on 2026-09-18

The boring part of ChatGPT Ads is the one that can cost you the most days. Not because the rules are strange, but because there are three separate reviews and most people assume there is only one. This lesson is the map, so you are never stuck without knowing why.

Three reviews, not one

Review What it looks at What happens meanwhile
Account brand Your identity as an advertiser: name, icon and website (official documentation) Until it is approved you can have everything built and serve nothing
Account integrity The account as a whole; separate from the one above The documentation warns that an active account with approved branding can still have another review outstanding
Ad Each creative: in review, approved or rejected A rejected ad breaks a policy: you edit it for it to be re-reviewed

📌 We measured this ourselves: in our account the 8 ads were approved, with no low-bid warning, and we still had days of zero impressions. The reason was the brand review, approved on 16/09/2026 at around 11:30. The campaign served the next day. If you see approved ads and no impressions, check the account reviews before touching the bid.

What reopens a review

The official documentation is explicit on one point: brand changes can trigger a review, and the brand review looks at name, icon and website. In other words, changing your trading name, your logo or the domain you send traffic to is not a cosmetic tweak — it is rejoining the queue.

On creatives, the only thing the documentation states clearly is that a rejected ad is edited so it gets re-reviewed. We found no official source detailing which edits to an already approved ad reopen its review and which do not, so the practical rule is the cautious one: assume that touching the title, text, image or destination means another review, and do not leave a creative change until the eve of a seasonal campaign.

Restricted sectors and content

Ads do not appear near personal health, mental health or politics, and political advertising is banned outright. Healthcare and financial services carry strict eligibility criteria and, at best, are assessed case by case — mostly in the US (Headcore manual, August 2026). The basic rules of the advertising policy: nothing deceptive, professional language, no discrimination, no imitating the ChatGPT interface, and a landing page that matches the advertiser and the offer.

Before building anything, the question is not "how do I word this so it passes?" but "is my category allowed in my country today?".

What the card promises must be behind the click

This one caught us out. In our own campaign we advertised two resources as a "free ebook" when they were in fact for subscribers. There was no bad faith: they had been open at another time. It made no difference — the card promised something the page did not deliver, so we pulled those ads.

⚠️ Trap: "free" with a sign-up in the middle is still free if you say so. What does not hold up is a click that leads to a different wall from the one you advertised. The check takes ten seconds: open your ad's link in a private window and ask whether a stranger gets exactly what was promised.

Other people's trademarks: in the description, never in the name

Two rules pointing the same way:

The way out is not to stop naming them: it is to use them as description. Our own product is not called "Ninja GPT Ads"; it is "Ninja Scripts Suite · for ChatGPT Ads" (and its twin, "· for Google Ads"). It says exactly what it does, and the other brand stays where it is allowed to be: explaining which platform it serves, not forming part of the name.

Checklist before launching

  1. Is my category allowed in my country? Does it brush against health, mental health or politics?
  2. Are the advertiser's name and icon final? (If you plan to change them, do it first: it triggers a review.)
  3. Does the landing page match the advertiser, respond, and let OpenAI's crawlers through?
  4. Is what the card promises behind the click, with no conditions I am not announcing?
  5. Am I using someone else's trademark in my product's name?
  6. Is measurement in place before the first impression?
  7. Have I left days of slack for the reviews before the date I want to be live?

What to take away

📎 Sources and further reading

⚠️ Free training with no support. Ninja Scripts support channels (email and Telegram) are only for the use of the scripts, not for Google Ads questions or questions about this training.

Pick up here

← BeforeWhy there are no official performance benchmarks, and how to read the ones going aroundChatGPT Ads — the platform, the market and our diaryAfter →Diary of our campaign, day by dayChatGPT Ads — the platform, the market and our diaryRelatedWhere does your ad send people? The link that breaks most campaignsAdvertising in ChatGPT: first stepsRelatedWhat ChatGPT Ads does NOT let you do (and what you do instead)Advertising in ChatGPT: first stepsRelatedAds and landing pages in ChatGPT Ads: the card, the image, review and variants you can measureChatGPT Ads in depth — ads, bidding and optimisationRelatedPicking a winner between two ads with statistics, not hunchesChatGPT Ads in depth — ads, bidding and optimisation

See the full syllabus

🎓
You're reading the Ninja Academy in the open

Create your free account (no card) to save your progress, take the graded quizzes, earn the Basic level certificate and unlock part of the Intermediate and Advanced training.

Create my free account →
🥷

Subscriber feature

This option is part of the Ninja Scripts Suite subscription.

See the subscription →