For most of the past two years, buying ads on AI platforms meant getting on a waiting list. OpenAI, Google’s Gemini, Perplexity — each gated their ad inventory behind managed relationships, sales teams, and minimum spends that kept smaller advertisers out. That model is changing. OpenAI’s expansion of its self-serve ads manager to the United Kingdom — now the fifth market to get direct access — isn’t just a geographic rollout. It’s a signal that AI platforms are serious about building scalable, measurable ad businesses, and the cost-per-click model they’re deploying alongside it is designed to prove ROI fast enough to pull budget from established channels.
The UK launch follows earlier self-serve rollouts in the United States, Canada, Australia, and Japan. UK advertisers can now create campaigns, set budgets, pull reporting, and manage billing directly inside ChatGPT’s advertising interface — no intermediary required. The CPC option, which first launched in the US on May 5, 2026, is now available in the UK from day one of that market’s self-serve access.
CPC Changes the Buyer Conversation
Cost-per-click is not a new pricing model. But its arrival on ChatGPT matters because of what it replaces: CPM-based sponsorships where advertisers paid for impressions against an audience they couldn’t fully verify was engaged. CPC flips that logic. If a user doesn’t click, the advertiser doesn’t pay. That’s a fundamentally different risk profile, and it’s one that performance-oriented budget holders — the ones running search and social campaigns — are already comfortable with.
Advertisement
300 × 250
Dave Dugan, OpenAI’s head of go-to-market for advertising, confirmed at Cannes Lions 2026 that CPC is now the default recommendation for advertisers entering the self-serve platform. Industry benchmarks cited by analysts put ChatGPT CPC in the $3–$5 range — above Google Search averages but potentially justified if conversion rates follow from the intent-rich query context that distinguishes AI assistant sessions from passive search browsing.
“The move to self-serve CPC is OpenAI essentially saying: prove it yourself,” said Claire Holubowskyj, an analyst at Enders Analysis. “They’re handing advertisers the tools to run their own experiments, which accelerates learning but also shifts accountability. If the ROAS isn’t there, the platform gets the blame — which means OpenAI needs conversion quality to be real.”
The Self-Serve Infrastructure Shift
Self-serve access — campaign creation, billing, and reporting in a single interface — is the operational threshold that separates a media buy from an advertising platform. OpenAI crossed that threshold in the US earlier this year. The UK expansion extends the infrastructure to a market with a mature programmatic ecosystem and large brands already running AI-adjacent campaigns on Meta and Google.
Ashley Fletcher, VP of Marketing at Adthena, noted that the UK market’s appetite for performance accountability makes it a logical second major test of the CPC model. “UK performance teams are used to rigorous attribution. If ChatGPT ads can show up in attribution models the same way paid search does, that’s when budget starts moving.” The implication: the UK rollout is as much a measurement credibility test as a revenue play.
The cost-per-action (CPA) variant, meanwhile, remains US-only for now — a more ambitious model that requires enough conversion data to optimize against outcomes rather than clicks. Its absence in the UK launch suggests OpenAI is sequencing deliberately: establish click-based accountability first, then layer in outcome optimization once the data infrastructure is in place.
What This Means for Marketers Right Now
The practical question for marketers in the UK — and those watching from markets not yet in self-serve — is where ChatGPT ads fit in an existing channel mix. The honest answer is that it depends on query intent. ChatGPT users in a research or decision-making session are different from search users scanning ten blue links. The ad placement context is more conversational, which may suit considered purchases better than impulse buys.
The CPC model makes experimentation low-risk enough to justify a test budget. The bigger risk is organizational: measurement infrastructure that wasn’t built to capture AI-assisted conversions will undercount performance, making the channel look worse than it is. Marketers who want accurate read-through from ChatGPT spend will need to instrument their attribution before the campaign launches, not after.
The shift also carries implications for the broader AI platform landscape. As AI platforms including OpenAI formalize partnerships with major brands like L’Oréal to function as both content factory and ad channel, the self-serve layer is what converts those strategic relationships into a scalable business. Enterprise partnerships set the narrative; self-serve CPC is what turns it into recurring revenue from the thousands of brands that will never get a managed relationship.
The Gatekeeping Era Is Ending
The managed-only model that defined AI advertising in 2024 served a purpose: it let platforms control brand safety, test placements without public scrutiny, and build relationships with accounts that could move significant budget. But managed relationships don’t scale. Five markets into self-serve, OpenAI is showing it has enough confidence in its ad product — and its moderation — to let advertisers in without a sales team holding the door.
That’s the real signal from the UK launch. It’s not that OpenAI suddenly needs UK ad revenue. It’s that the platform has decided its ad product is ready to be judged on performance, in public, by performance marketers who will pull spend the moment the numbers don’t work. That’s a high bar. The CPC model means they’ve accepted it.
