OpenAI’s ad platform is no longer a test. Since August 2026, ads inside ChatGPT have gone from a limited US pilot to a full rollout. For B2B marketers, that means a new channel is competing for budget, right alongside Google Ads and LinkedIn Ads.
The pitch is simple: your ad appears, clearly labelled as sponsored, underneath a relevant ChatGPT answer, matched to the conversation through a relevance-weighted auction. Someone clicks. They land on your site. And then, for most marketing teams, the trail goes cold.
Not because the click didn’t happen. Because proving which companies that click brought to your site is a different problem altogether, and it’s one that standard analytics wasn’t built to solve.
What ChatGPT ads actually offer today
OpenAI’s Ads Manager lets advertisers bid on a CPM or CPC basis, add UTM parameters to their landing pages, and see aggregate performance: impressions, clicks, click-through rate and spend.
But it stops at the click. Advertisers get no visibility into who clicked, only how many people did. For a B2B marketer used to asking “did this bring in a qualified account, or just traffic,” that’s a gap.
Why ChatGPT traffic is so easy to lose
Even with UTM parameters set up correctly, plenty of that traffic never reaches your analytics with its source intact. Links opened in ChatGPT’s in-app browser, or in a new tab on mobile, often strip the referrer header entirely. The result is a session that shows up as chatgpt.com / (not set), or worse, disappears into “direct” traffic with no source at all.
That’s not a small margin of error. It means a chunk of the budget you spend on ChatGPT ads shows up in your reporting as if it never happened, right when leadership is asking what your spend delivered.
Counting clicks isn’t the same as knowing who converts
Say your tracking is clean and every session keeps its UTM tag. You still only know that a visitor came from your ChatGPT ad. You don’t know if that visitor works at a 200-employee logistics company that fits your ICP, or if it’s a student testing prompts.
This is the same problem marketing teams already have with paid search and paid social, just newer and less mapped out. And it points to the same answer: stop measuring the channel by clicks, and start measuring it by the companies it actually brings through the door.
Where Leadinfo insights fit in
Leadinfo identifies the company behind a website visit directly, without relying on cookies or a completed form. Either the visitor arrived from organic search, LinkedIn, or a ChatGPT ad, it keeps working.
In practice, that means you can:
- Build a Segment for visits landing on your ChatGPT campaign’s UTM-tagged page, and see exactly which companies show up there, not just how many clicks arrived.
- Set a Trigger so sales gets an alert the moment a fit-profile account lands on that page, while the ad is still fresh in their mind.
- Feed what you learn back into targeting. The same principle that lets Leadinfo pass offline conversions into Google Ads, so it learns which companies actually convert, applies here: once you know which accounts your ChatGPT ads attract and close, that intelligence sharpens every future campaign, on ChatGPT or anywhere else.
The bigger picture
ChatGPT ads will keep evolving, and so will the ways to measure them. What won’t change is the underlying question every marketing manager has to answer: not “how many clicks,” but “which companies, and were they the right ones.”
