ChatGPT Ads Can Now Generate Your Ads: How the AI Ad Builder Works and How to Use It Safely
ChatGPT Ads can now generate ads for you from your website and campaign settings. Here is how OpenAI's new AI ad builder works and how to use it safely.
ChatGPT Ads Can Now Generate Your Ads: How the AI Ad Builder Works and How to Use It Safely
ChatGPT Ads can now generate ads for you. As of July 6, 2026, OpenAI's ad platform added a feature that builds an ad for you automatically, based on your website and your campaign settings. You find it under the "Add new ad" option, where a new prompt reads "generate ads for you." You let ChatGPT draft the ad, then review, edit, and approve it before it goes live. This is OpenAI using its own AI to lower the effort of creating ads, and it changes the first step of running a ChatGPT Ads campaign from writing an ad to editing one.
Here is exactly what launched, how the ad builder works, and how to use it without letting an autogenerated ad embarrass your brand.
What OpenAI actually shipped
The feature showed up inside the ChatGPT Ads platform and was first spotted publicly on July 6, 2026, when marketer Anthony Higman posted screenshots on X that Search Engine Land then reported.
It is a small interface change with a real workflow behind it.
When you go to add a new ad, the platform now offers to generate one. The screenshot shows a message that reads: "We generated an ad variation based on your website and campaign settings. Review, edit as needed, and activate when you're ready." From there you move to a "Review and create" step, where you can accept the draft, change it, or throw it out.
A second, related option appeared alongside it: a quick "duplicate ad" button that lets you copy an existing ad rather than start from scratch. Together, the two additions point at the same goal. OpenAI wants fewer blank screens between an advertiser signing up and an ad going live.
It generates from your site and your settings
The important detail is what the AI reads to build the draft. According to the in-product message, it uses your website and your campaign settings. That means the quality of the generated ad depends heavily on two things you control: how clearly your website communicates what you sell, and how specific your campaign settings are. A thin or vague landing page gives the model less to work with, and a wide-open campaign setup gives it fewer guardrails. Neither of those is new advice, but the ad builder makes them matter at the very first step.
You still approve everything
Nothing publishes on its own. The generated ad lands in a draft state and waits for you to review, edit, and activate it. That human approval step is the whole reason this feature is usable for real brands. It is also the step most likely to get skipped by advertisers in a hurry, which is where the risk lives.
Why this matters for advertisers
ChatGPT Ads has moved fast since it launched in February 2026. OpenAI's chief revenue officer said in June, at Cannes Lions, that users dismiss ChatGPT ads about half as often as they did when the format first launched, which the company frames as the format getting more relevant over time. And a Search Engine Land analysis of 6.77 million sessions published in early July found ChatGPT accounts for roughly 92% of AI referral traffic, which tells you where the attention is going when people search inside an assistant instead of a traditional search box.
Put those together and the picture is clear. The audience on ChatGPT is large and growing, the ad format is getting less annoying to users, and OpenAI just removed one of the last bits of friction for advertisers by writing the first draft of the ad for them. For anyone weighing a ChatGPT Ads test, the cost of entry in time just dropped.
The tradeoff: speed versus control
Every "AI writes it for you" feature in advertising carries the same tradeoff. You save time up front and you take on review work at the back. Google made this explicit when it updated its advertising terms on July 1, 2026, putting the responsibility for reviewing and approving automated ad output squarely on the advertiser. OpenAI's ad builder sits in the same category. The platform will happily generate an ad, but the brand, the claims, and the compliance are still yours.
That is not a reason to avoid the feature. It is a reason to build a review habit around it.
How to use the ChatGPT Ads generator without regretting it
The generated draft is a starting point, not a finished ad. Treat it that way and it saves you real time. Skip the review and it can put an off-brand or inaccurate ad in front of a large audience. A short, repeatable checklist keeps you on the right side of that line.
- Check every factual claim. The model builds from your website, and websites go stale. Confirm that prices, offers, guarantees, service areas, and product names in the generated ad are still true today. If a claim is not verifiable, cut it.
- Read it in your brand voice. Autogenerated copy tends toward generic enthusiasm. Rewrite anything that sounds like it could belong to any competitor. Your ad should sound like you, not like a template.
- Match the ad to the landing page. Make sure the promise in the ad is the promise the page delivers. A mismatch between ad and page wastes spend and frustrates the person who clicked.
- Watch for regulated language. If you work in health, finance, legal, or any category with disclosure rules, assume the generator does not know your rules. Add required disclaimers and remove any claim you cannot back up.
- Keep your own control ad. Do not let a generated ad replace a version you wrote and trust. Run the generated draft against your own copy so you can tell whether the AI version actually performs better.
Tighten your inputs before you generate
Because the builder reads your website and campaign settings, the fastest way to get better drafts is to give it better inputs. Make sure the landing page you are pointing to states plainly what you offer, who it is for, and what happens when someone clicks. Set your campaign targeting and goals with intent rather than leaving them broad. The more specific the inputs, the less editing the output needs.
Decide where the generator fits your workflow
For small advertisers with no creative team, the generator is a reasonable way to get a first campaign live and start collecting data. For agencies and larger brands, it is most useful as a drafting tool that a human then shapes, and as a fast way to spin up variations to test. Either way, the value is in the editing, not the generating. The teams that win with tools like this are the ones with a strong point of view about what a good ad says, using the AI to get to a draft faster.
What we do not know yet
This is an early sighting, not a formal product announcement with documentation, so a few things are still unclear. We do not know whether every ChatGPT Ads account has the feature yet or whether it is rolling out in stages. We do not know exactly which model generates the ads, how much control you get over tone or length, or whether OpenAI plans to add brand guidelines or a memory layer the way other platforms have. We also do not know how the generated ads perform against human-written ones, because the feature is too new for reliable data. Expect those details to firm up as OpenAI publishes more and as advertisers report results.
Frequently asked questions
Can ChatGPT Ads really write ads for you now?
Yes. As of July 6, 2026, the ChatGPT Ads platform includes a feature under "Add new ad" that generates an ad for you based on your website and campaign settings. It creates a draft that you then review, edit, and activate. The ad does not go live until you approve it.
How does the ChatGPT Ads generator work?
You choose to add a new ad, and the platform produces an ad variation using your website content and your campaign settings as inputs. It shows the draft with a message telling you to review and edit as needed, then lets you create the ad once you are satisfied. There is also a duplicate option for copying an existing ad quickly.
Is it safe to use AI-generated ads on ChatGPT?
It is safe if you review carefully. The generated ad is a draft, and you are responsible for what it says, so check that every claim is accurate, the copy matches your brand and landing page, and any required disclosures are present. Do not activate a generated ad you have not read line by line.
Does the ChatGPT ad generator publish ads automatically?
No. The feature generates a draft and waits for you to review, edit, and activate it. Nothing is published without your approval, which gives you the chance to fix or reject anything the AI got wrong.
Should small businesses use the ChatGPT Ads generator?
It can be a practical way for a small business with no creative team to get a first campaign live and start learning what works. The key is to edit the draft rather than run it as-is. Tighten your landing page and campaign settings first, because the generator builds from those inputs.
How is this different from ChatGPT Ads costs and pilots?
This feature is about how ads get created, not what they cost. It automates the drafting step. Whether ChatGPT Ads makes sense for your budget and goals is a separate question about pricing, audience, and expected return, which you should evaluate before you spend.
Bottom line
ChatGPT Ads can now generate ads for you, and as of July 6, 2026 the feature lives under the "Add new ad" option, drafting an ad from your website and campaign settings for you to review and approve. It lowers the effort of getting a ChatGPT Ads campaign live, which matters because the audience on ChatGPT is large and the format is getting more relevant. The catch is the same one that comes with every automated ad tool: the platform writes the draft, but you own the claims, the brand voice, and the compliance. Use it as a drafting assistant, tighten your inputs, review every line, and keep a human version to test against. Do that and the generator saves you time. Skip the review and it can cost you more than it saves.
*Want help testing ChatGPT Ads for your business without the guesswork? Book a call with us.*
Related articles
- Claude Opus 5 Explained: What Anthropic's New Model Does, What It Costs, and When to Use It
- GPT-5.6 Explained: Sol, Terra, Luna, Pricing, and When You Can Use It
- AI-Generated Ads and the EU AI Act: What Advertisers Must Disclose Before August 2, 2026
- Claude Managed Agents, Self-Hosted Sandboxes, and MCP Tunnels: What Anthropic Just Shipped
- Gemini CLI Is Dead: What to Use Instead and How the Antigravity Transition Works
- What ChatGPT Ads Actually Cost in 2026, and Which Businesses Should Run a Pilot
- All Kanopy resources
- SEO & web services
- Paid ads services
- AI outreach services