Creative review is OpenAI's evaluation of your ChatGPT ad against its published ad policies, covering the title, body copy, image or media, price text, and destination landing page before the ad becomes eligible to serve. Advertiser review checks whether your business qualifies to advertise; creative review checks whether the ad itself follows the rules.
You care because this review decides whether a campaign can run at all, so it gates every dollar you plan to spend inside ChatGPT. Ignore it and a single unsupported claim or broken landing page can get the entire ad rejected, stalling your launch while rivals keep serving.
Inside ChatGPT Ads, creative review is the stage where OpenAI judges your ad and its landing page against the OpenAI Ad Policies, and it is the middle of three documented review levels. It sits between advertiser eligibility and placement, and it decides whether the specific creative you uploaded is allowed to serve.
The review looks at your ad as one unit: the title, body copy, image, price text, and the destination landing page are checked together. OpenAI runs this mostly through automated machine learning systems, including large language models and classifiers, with human oversight. Borderline, high-severity, or restricted-category cases get escalated to human reviewers.
Advertiser review confirms your business is legitimate, and placement review weighs whether a given conversation suits your ad. Creative review governs the creative and its landing page. The outcome shows up as your Review Status: Approved or Rejected, or In Review until it resolves. AirOps helps you prepare creative and landing pages that clear this check the first time.
Resources: See how targeting, creative specs and ad policy shape ChatGPT ads
Creative review moves through a set sequence, from upload to an ongoing watch after approval. Here is the path your ad takes.
Submit: You upload the creative in ChatGPT Ads Manager: title, body copy, image, price text, and destination URL.
Automated screening: Machine learning systems, including large language models and classifiers, scan the creative, image, and landing page against policy.
Human escalation: Borderline, high-severity, or restricted-category cases route to human reviewers for a closer look.
Decision: OpenAI sets your Review Status to Approved or Rejected with a reason. It stays In Review until resolved.
Ongoing monitoring: OpenAI re-checks approved ads, and delivery can be limited, paused, or removed. Landing-page changes can trigger a fresh review.
A finished review tells you two things: the status (Approved or Rejected) and the reason behind it. It does not tell you the ad will appear beside any given ChatGPT conversation, because that call belongs to placement review.
Creative review is the gate between your budget and the conversation you paid to reach. Until an ad clears it, your money cannot buy a single impression in ChatGPT, so the review shapes how you plan timelines and spend. Treat it as a spend decision from the start.
It controls whether you can spend at all: No approval means no delivery, so a stuck review freezes the campaign and the budget sitting behind it until the issue is fixed.
A broken landing page sinks clean copy: OpenAI reviews the creative and its destination together, so a mismatched or broken landing page gets the whole ad rejected even when the ad text is clean.
Restricted categories need extra proof: Financial, health, and legal offers are limited to approved US advertisers case by case, so planning for substantiation up front keeps your launches on schedule.
Growth marketers use creative review to catch policy and landing-page problems before launch, so test campaigns start serving on the planned date.
SEO managers use creative review to confirm the destination page matches the ad claims, keeping the landing experience consistent with what ChatGPT users were promised.
Demand gen leads use creative review to forecast approval risk across a campaign set, so pipeline targets do not slip when an ad gets rejected.
OpenAI evaluates your title, body copy, image or media, price text, and destination landing page together as one connected unit, which means approved creative cannot link out to content that breaks policy, and a change to the landing page alone can send the whole ad back for another look.
Creative review runs mostly through automated machine learning systems, including large language models and classifiers, while human reviewers calibrate those models and step in on borderline, high-severity, or restricted-category cases that the automation flags for a closer decision.
OpenAI blocks whole categories such as adult, gambling, alcohol, tobacco, drugs, crypto, credit-repair, and political ads, and it limits financial, health, and legal offers to approved US advertisers that it reviews case by case.
Confirm your ChatGPT ad meets OpenAI policy before you commit budget to it.
Catch landing-page mismatches that would sink an otherwise clean ad.
Read your Review Status in ChatGPT Ads Manager to see Approved, Rejected, or In Review.
Shorten resubmission cycles by fixing the reason OpenAI gives on rejection.
Protect live campaigns by expecting re-checks after you edit a landing page.
Read the OpenAI Ad Policies before you design the ad, because the rules decide what your creative and landing page can say.
Match every claim in the ad to proof on the landing page, since reviewers check the creative and destination together.
Substantiate financial, health, and legal claims up front, because those categories are restricted to approved US advertisers reviewed case by case.
Keep the landing page live and consistent after approval, since page changes can trigger a fresh review and pause delivery.
Save the rejection reason and fix that exact issue before resubmitting, so you stop repeating the same violation.
Build the creative to stand on its own without imitating OpenAI's interface or implying an endorsement, because that alone draws a rejection.
Avoid treating approval as permanent. OpenAI keeps monitoring live ads after they clear, and it can throttle, pause, or pull delivery later, so build a habit of reviewing active creative and its landing pages on a regular schedule.
AirOps: AirOps partners with your team to prepare creative and landing pages that match OpenAI policy, so your ChatGPT ads clear creative review the first time.
ChatGPT Ads Manager: OpenAI's own interface is where you submit creative and read the Review Status that creative review produces.
Screaming Frog: This crawler checks your destination pages for broken links and mismatches before creative review flags them.
Read the policy: Open the OpenAI Ad Policies this week and note the prohibited categories and the claim rules. This costs nothing and sets the boundaries for everything you build.
Audit your claims: List every claim in your planned ad and mark which ones you can substantiate. Flag any financial, health, or legal language that needs approved-advertiser status.
Align creative and landing page: Make the title, body copy, image, and price text match the destination page, since OpenAI reviews them together. Remove anything that imitates OpenAI or implies an endorsement.
Submit in Ads Manager: Upload the finished creative in ChatGPT Ads Manager and watch the Review Status move from In Review to Approved or Rejected.
Build a resubmission loop: When a rejection lands, read the reason, fix that exact issue, and resubmit. Keep a short log so repeat problems get rarer over time.
Creative review is OpenAI's policy check on your ChatGPT ad and its landing page before the ad can serve.
It runs mostly on automated machine learning systems, with human reviewers calibrating them and handling the hard calls.
The creative and its destination landing page are judged as one unit, so both have to hold up.
A single weak claim or broken page can reject the whole ad and freeze your spend.
Your real leverage is preparation: policy-ready claims and a consistent landing page win approval faster and cut rework.
Advertiser review and creative review check different things, and ChatGPT Ads runs both. Advertiser review looks at your business: whether you are a legitimate, eligible advertiser allowed on the platform at all. Creative review looks at the specific ad you uploaded, including its title, body copy, image, price text, and the landing page it sends people to. One clears the company; the other clears the campaign asset. You can pass advertiser review and still have an individual ad rejected, because the creative broke a content rule or pointed to a non-compliant page. OpenAI also runs a third check, placement review, which decides whether a given conversation is a suitable spot for your ad. Knowing which level you are dealing with tells you where to fix a problem. A rejection tied to your copy or destination page is a creative issue, so that is where your edits should go first.
It varies, and OpenAI does not publish a fixed turnaround, so plan for a review window and keep buffer in your schedule. Automated systems handle most reviews and can return a decision quickly, but any ad routed to a human reviewer takes longer. Ads with borderline creative or restricted categories such as financial and health offers tend to add time, because those get a closer human look. Your ad stays In Review until the check resolves, then flips to Approved or Rejected. Because you cannot count on a set clock, submit well ahead of any launch date. Keep your claims easy to verify, and make sure the landing page matches the ad so nothing triggers manual escalation. If a deadline is tight, give yourself buffer days for a possible rejection and resubmission, since a second pass resets the wait.
Small differences that look trivial to you can cross a policy line, which is why two near-identical ads get different outcomes. Creative review weighs the whole ad plus its landing page, so a single changed word, a different image, or a destination URL that resolves differently can flip the result. A claim that reads as puffery in one ad might read as an unsubstantiated promise in another, and that triggers a rejection. The landing page matters just as much as the copy, so if one version points to a page with disallowed content or a broken link, that version fails while the other passes. Automated systems and human reviewers can also weigh context differently on borderline cases. When this happens, read the specific rejection reason instead of comparing the two ads side by side. The reason names what failed, and that is faster than guessing which detail made the difference.
You cannot pay to jump the queue or lobby a reviewer, but you can influence the outcome by how you build the ad. OpenAI runs creative review against published policies, so the most reliable lever is submitting creative that already meets those rules. Keep every claim truthful and easy to substantiate, drop any disparaging competitor comparison, and make sure nothing imitates OpenAI's interface or implies an endorsement. Align the ad with its landing page so the destination delivers exactly what the copy promises, because a mismatch is a common rejection cause. Clean, policy-ready submissions are also less likely to get routed to slower manual review. If an ad is rejected, the fastest path is to read the stated reason, fix that exact issue, and resubmit instead of tweaking at random. Treat your influence as preparation and precision that earns a cleaner pass.
There is no published benchmark pass rate for creative review, so chasing a specific percentage will mislead you more than it helps. OpenAI does not release figures scoped to creative review, and any number floating around without a clear source should be treated with caution. A more useful signal is your own trend over time. Track how many of your ads clear on the first submission, group the rejections by their stated reason, and watch whether that first-pass rate climbs as you tighten your process. If the same reason keeps appearing, fix it at the template level so it stops recurring. A healthy program is one where rejections stay rare and quick to resolve, because your claims are substantiated and your landing pages stay consistent. Measure yourself against your last quarter, since that comparison is real and within your control.