ChatGPT Ads creative enhancements review (2026): an honest verdict on OpenAI's default-on auto-rewrite — what it changes, the trade-off, and who should opt out.
Creative enhancements is a reasonable idea delivered on uncomfortable terms. Letting OpenAI tailor your headline, image, translation, and even landing page to each conversation can lift relevance — but it's on by default, runs variations you never approve, previews no version, and offers no report isolating its effect. For a tightly governed brand or anything with compliance-sensitive copy, opt out per campaign until you've tested it on your own terms. For a low-stakes direct-response advertiser with plenty of raw creative, it's worth a controlled trial.
This review scores a feature, not a platform: "creative enhancements," the setting OpenAI turned on by default for ChatGPT Ads at the end of October 2026. It is not a review of ChatGPT Ads as a channel — we cover that separately — but of the specific decision to let OpenAI's system automatically change a live ad to match the conversation it appears in.
The scores below rate the things that actually decide whether you leave this toggle on: how much relevance lift it can plausibly add, how much creative control you keep, how transparent the changes are, how clear the opt-out is, how well you can measure the effect, the brand-voice and compliance risk, the convenience it buys, and — because this matters — how OpenAI rolled it out. Where the feature is genuinely sensible (conversation-matched relevance is the whole promise of advertising inside a chat), it scores well. Where it's raw (default-on, no per-version preview, no isolating report), the scores say so plainly.
One honest caveat shapes the whole review: the feature is new, and OpenAI has published no performance data for it. So this is a verdict on the design and the terms, not on measured uplift. We don't rate what we can't verify, and nobody — including OpenAI, by its own measurement caveat — can yet prove how much it helps. Everything here reflects the feature's state as of 2026-10-31; confirm the toggle, dates, and reporting on OpenAI's advertising help pages.
Creative enhancements is an automatic optimization setting inside ChatGPT Ads. With it on, OpenAI's ad system can change a running ad to better fit the conversation where it surfaces, drawing on the conversation context or the user's interests plus your existing ad and landing-page content. A single toggle covers four kinds of change: rewriting or writing new headlines and descriptions, translating the ad text into another language, resizing your image to fit the format, and changing the destination URL to a more relevant page on your own site. Variations may be prepared in advance or generated in real time. OpenAI emailed advertisers on October 20, 2026 and set the feature on by default for new and existing campaigns, effective Friday, October 30, 2026 — its own help page showed shifting dates in the days beforehand before settling on that same date. The control lives in Ads Manager Settings at both the account level (the default) and the campaign level, with campaign settings overriding the account default and any prior opt-outs preserved. Two design facts define it: the variations run without each one being approved by the advertiser, and Ads Manager does not preview every version. OpenAI also cautions that changes in results over time do not, on their own, establish the feature's effect, and it describes no report that separates an enhanced version from your original.
Creative enhancements fits a direct-response advertiser who cares more about conversions than exact wording, has a flexible set of relevant landing pages, and can treat early spend as a test rather than a forecast. If your value proposition survives being paraphrased and your pages are all on-brand, letting the system match copy and destination to intent can help. It fits poorly for anyone whose ad copy is load-bearing: regulated industries with approved claims, brands with a precise voice, offers where a specific price or term must appear verbatim, or funnels built around one exact landing page. For those advertisers the lack of per-version preview and the absence of an isolating report turn a relevance feature into an uncontrolled variable — and the right move is to opt the campaign out and run only creative you approved.
| Dimension | Score | Why |
|---|---|---|
| Relevance-lift potential | 3.8 / 5 | Matching copy, language, and landing page to the live conversation is a sound premise for advertising inside a chat — though unproven in published numbers. |
| Creative control | 2.3 / 5 | The system can rewrite your words, swap your image size, and reassign your click destination; with it on, you cede the exact creative decision. |
| Transparency & preview | 2.0 / 5 | Ads Manager does not preview every version, so you cannot see each variation that ran against your brand before it ran. |
| Opt-out clarity | 3.2 / 5 | A single clear toggle at account and campaign level, and prior opt-outs were preserved — but campaign settings override the account default, so an account-level opt-out is not enough. |
| Measurement & attribution | 2.4 / 5 | OpenAI itself says results over time don't prove the effect, and there is no report isolating an enhanced version from the original. |
| Brand-voice & compliance risk | 2.2 / 5 | Auto-rewritten headlines and translated copy are a real hazard for regulated claims, precise pricing, or a controlled voice. |
| Convenience / automation | 4.0 / 5 | For advertisers short on creative, hands-off variation and translation is genuinely useful and removes a production step. |
| Rollout approach (default-on) | 2.5 / 5 | Switching an opt-out creative-rewrite on by default, with only days' notice, puts the burden on advertisers to protect their own ads. |
Creative enhancements carries no separate price — it's a setting inside a campaign you already pay for on a cost-per-click or reach basis. So the honest way to read its "cost" is not in dollars but in control and legibility. You pay by handing OpenAI's optimizer the authority to change your words, your image dimensions, and your click destination, and by accepting that you won't preview every version and can't isolate its effect in reporting.
Whether that trade is worth it depends entirely on how load-bearing your exact creative is. For a direct-response advertiser with interchangeable landing pages and copy that survives paraphrase, the "price" is low and the convenience is real. For a brand whose headline carries an approved claim, a regulated disclosure, or a specific price, the price is high and mostly hidden until something runs that shouldn't have. OpenAI's own measurement caveat compounds this: if you can't attribute a November lift to the enhancement rather than the season or a reassigned page, you're paying in control for a benefit you can't cleanly bank.
The fair verdict on terms: the feature isn't a cash cost, it's a governance cost, and defaulting it on shifted that cost onto advertisers who may not have read the email. Budget your attention, not your money — decide per campaign, and test against your own opted-out baseline before you trust it.
| Use case | Fit | Why |
|---|---|---|
| Direct-response with flexible copy and pages | Strong | Conversion matters more than exact wording, and on-brand interchangeable pages make a reassigned destination low-risk. |
| Reaching other-language ChatGPT users | OK | Automatic translation can extend reach, but you can't preview every translated version before it runs. |
| Advertisers short on creative production | Strong | Hands-off variation removes a real bottleneck — though the raw material is still only as good as the ad you supplied. |
| Regulated industries with approved claims | Weak | Auto-rewritten headlines and descriptions risk altering claims or disclosures that must read verbatim. |
| Brands with a precise, governed voice | Weak | Without per-version preview, you can't guarantee each running variation stays on-voice. |
| Offers tied to one exact landing page | Weak | A changed destination URL can override the page and break the funnel or tracking built around it. |
| Clean A/B testing of your own creative | OK | You can opt out per campaign and run only your approved assets — the feature just isn't the tool for that. |
| Hands-off advertisers who trust the optimizer | Strong | If you want the platform to handle variation and accept the opacity, the default-on behavior is what you want. |
Kompozy is not an ad feature and doesn't compete with creative enhancements head-to-head — it never touches the toggle. Where it matters is the question the feature sidesteps: who produces the variations, and are they on-brand? Creative enhancements exists partly because supplying enough relevant creative is hard, so letting OpenAI rewrite one ad into many is the shortcut. The alternative to the shortcut is having the variations already.
That's what Kompozy does. From one Persona Brief it generates a batch of on-brand assets — Photo Posts, Quote Graphics, brand-exact Carousel Posts, Persona Shorts and other short-form video — each held to the same voice and banned-word list, so a dozen variations stay recognizably yours and compliance-safe. Opt a campaign out of enhancements and you have the approved creative to test and improve by hand; leave it on, and the raw ad the system remixes is already strong and on-brand. The honest framing is complementary, not competitive: OpenAI's optimizer decides what runs; Kompozy decides what it has to work with — and produces the organic footprint across eight social platforms plus blog and email that earns a ChatGPT mention no ad setting controls.
It depends on how load-bearing your exact creative is. If your ad copy carries approved claims, a specific price, a regulated disclosure, or a precise brand voice, opt out per campaign until you've tested the feature on your own terms — the lack of per-version preview makes it risky for you. If you run direct response with flexible copy and interchangeable on-brand landing pages, a controlled trial is reasonable.
Four things, under one toggle: it can rewrite or write new headlines and descriptions, translate your ad text into another language, resize your image to fit the format, and change the destination URL to a more relevant page on your own site. It draws on the conversation context and your existing ad and landing-page content.
No. OpenAI's help documentation states that the experience does not preview every variation, and the variations can run without each one being reviewed or approved by the advertiser. That is the central control concern with the feature.
Not cleanly. OpenAI itself cautions that changes in results over time do not, on their own, establish the effect of creative enhancements, and it describes no report that isolates an enhanced version from your original. Adding UTM tags to destination URLs helps, but attribution remains fuzzy.
No. The setting exists at both account and campaign level, and campaign settings take priority. An account-level opt-out does not switch off a campaign where the feature is already on, so you have to check each campaign individually in Ads Manager Settings.
OpenAI emailed advertisers on October 20, 2026, naming Friday, October 30, 2026 as the effective date. OpenAI's own help page showed shifting dates in the days before that, which it later corrected to match the email. Opting out before October 30 avoided any ambiguity.
The earlier option generated ad copy that you still had to approve before it ran. Creative enhancements removes that step — variations can run without per-ad approval and without previewing every version — and it also covers images, translation, and destination URLs, not just text.
You test your own variations. A content engine like Kompozy generates on-brand image and video assets from a single Persona Brief, so you can run and compare many approved creatives yourself instead of relying on an auto-rewrite you can't preview.
See ChatGPT Ads Creative Enhancements (OpenAI) vs Kompozy comparison → · Get Started →