ChatGPT Visual Ads: What Agencies Should Know

OpenAI announced on October 5 that it is introducing a visual ad format inside ChatGPT, and both Unite.AI and Securities.io covered the rollout in detail. Testing starts later this month in the US with an initial group of advertisers, and it lands alongside a much wider measurement and attribution partner list than ChatGPT Ads has had before. For an agency placing or advising on paid media, this is the first real signal of what advertising inside a conversational AI product is going to look like in practice.

What Actually Shipped

Strip away the announcement language and three pieces matter for account work.

• A new image ad format. The ads show up during image generation inside ChatGPT, built around "product inspiration, product usage, or the experiences a product makes possible." They are clearly labeled and kept visually separate from whatever a user is generating themselves, and OpenAI has said the ads carry no influence over ChatGPT's actual answers.
• A bigger measurement bench. OpenAI added attribution partners including AppsFlyer, Triple Whale, Adjust, Northbeam, Branch, Singular and Kochava, plus advanced measurement partners like Fospha, Measured and INCRMNTAL, and data integrations through Hightouch, Tealium and LiveRamp. It also named DoubleVerify and Integral Ad Science as brand-safety pilot partners.
• Early performance numbers from pilot advertisers. OpenAI cited WeightWatchers seeing a 15.3% lower cost per acquisition versus search benchmarks, and Portland Leather reporting that 93% of visitors arriving from ChatGPT Ads were new to the brand. Both figures come from OpenAI's own announcement, not independent audits, so treat them as a vendor's best case rather than a guaranteed outcome.

Put together, this is OpenAI building out the plumbing — attribution, brand safety, geo-testing — that performance marketers expect before they commit real budget, not just adding a new ad unit for its own sake.

Why This Matters for Agency Work

ChatGPT now reaches roughly 1.2 billion weekly users, which is reach no agency planning paid media can ignore for long. A visual format that sits inside image generation is a genuinely new placement type, not a reskin of a search or social ad, and the fact that OpenAI is pairing it with real attribution partners instead of launching it measurement-blind is the detail worth flagging to clients who have been asking "should we be advertising on ChatGPT."

The honest caveat: this is a test, in the US, with an initial group of advertisers, in consumer verticals first. Financial services and healthcare access is rolling out more gradually, and the ad unit itself is scoped narrowly to the image-generation moment rather than every ChatGPT interaction. Clients who want in now may not get access yet, and the cost-per-acquisition figures being quoted are early pilot numbers from two specific brands, not a category benchmark.

Treat this as a format to understand and request access to, not a line item to budget against yet. The test group is still small.

What to Do This Month

If a client has budget allocated to emerging AI ad formats, this is worth a short brief: what the format looks like, who the measurement partners are, and that initial access is limited. For ecommerce and DTC clients in particular, the WeightWatchers and Portland Leather numbers are a useful talking point, with the caveat attached that they come from OpenAI's own case studies.

For agencies that already route attribution through one of the named partners — AppsFlyer, Triple Whale, Northbeam and the rest are common stacks — check whether that integration extends to ChatGPT Ads reporting, since that is the fastest way to get a real read on performance once a client account gets test access. And if a client asks whether they should wait for broader availability or request early access now, the honest answer depends on category: consumer brands with existing paid social budgets have more to gain from testing early than regulated categories still waiting on wider rollout.

None of this requires rebuilding a media plan around a format that is still in limited testing. It does reward agencies that can speak to it clearly before a client brings it up first. If you want a second opinion on where an emerging ad platform fits your client media strategy, book a free consultation and we will help you sort out what is worth testing now versus later.