OpenAI has added visual ad formats to ChatGPT. The platform previously supported text-only ads; advertisers can now use images and graphics.
For arbitrageurs, this opens a new placement with an audience arriving with specific intent. Users in chat are often at a decision point — seeking advice, comparing options, clarifying details. That's warm traffic.
Third-party measurement and tracking
OpenAI is integrating third-party measurement tools for campaign tracking. Advertisers will be able to connect external analytics systems and verify results independently of the platform's own data.
This matters. When a traffic source provides only its own statistics, there are always questions about objectivity. The ability to use third-party trackers reduces the risk of discrepancies and gives a clearer picture of conversions.
Brand suitability controls
Alongside visual formats, OpenAI has introduced brand suitability tools — controls over the content environment in which ads appear. Advertisers can configure where their creatives show up.
For verticals like finance, health or education, this is critical. Showing a credit product ad next to a discussion of bankruptcy, or nutra in the context of medical diagnoses, is a direct path to account suspension and compliance issues.
What this changes for arbitrage
ChatGPT as an ad platform differs from classic social networks and search engines. There is no feed, no infinite scroll. A user asks a question, receives an answer — and the ad appears in that context.
Intent is higher than in most other channels. The person is not scrolling passively; they are actively seeking a solution. If the offer lands at the moment it is needed, conversion can exceed warmed traffic from targeting.
Visual formats expand creative options. A text ad works when the offer is simple. But if you need to show a service interface, a comparison table or a before-and-after result, an image delivers more.
What to watch
It is unclear which GEOs will be available at launch and which verticals OpenAI will approve through moderation. Gambling, betting, crypto — these typically face restrictions first on new platforms.
Traffic cost is also an open question. If the platform starts with high rates, testing will be expensive. But if OpenAI wants to onboard advertisers quickly, there may be a period of suppressed CPMs — a window for those who enter early.
Third-party trackers solve the transparency problem but add a technical layer. You will need to configure integration, pass postbacks, monitor discrepancies between the platform and the tracker. For teams without experience, this can be a barrier.
Worth testing?
If you work with offers where warm intent matters — finance, SaaS, education, health — ChatGPT may deliver results. The audience there has purchasing power and specific queries.
For classic nutra or aggressive grey schemes, the platform is unlikely to fit. OpenAI moderates content strictly, and brand suitability tools exist precisely to filter out questionable verticals.
The first months after visual format launch are a testing window. Competition is still low, algorithms are learning, moderation may be softer than it will become later. If you have budget for experiments, it makes sense to enter now and collect data while traffic has not yet become expensive.
