OpenAI is testing Sponsored Agents with select US advertisers, letting users chat with business-sponsored agents and visit advertiser websites via ChatGPT ads
Context & Ripple Effects
OpenAI’s ad business moved from seeking large early advertiser commitments in February to a US self-serve Ads Manager beta priced on cost per click in May. Sponsored Agents add a post-click conversational layer to that buying stack.
The format also puts into practice the sponsored-content direction OpenAI staff had discussed in 2025, including surfacing commercial material for relevant queries. Public reaction framed the test as a shift from buying attention to supporting a buyer conversation inside ChatGPT.
First-order effects
- Select US advertisers can pair ChatGPT ads with branded agents that answer follow-up questions and direct interested users to the advertiser’s website.
- OpenAI expands its ad product from a click-based placement into an advertiser-supported interaction within ChatGPT.
Second-order effects
- Advertisers using OpenAI’s Ads Manager gain a new reason to measure performance beyond the initial click: whether an agent conversation advances a user to the brand’s site.
- Brands participating in the test must make their product information and handoff to their own websites useful inside a conversational flow, rather than relying only on ad copy and landing pages.
Third-order effects
- If conversational ad formats scale, ChatGPT can become an agent-distribution gatekeeper: it controls both the sponsored introduction and the interface through which users evaluate a brand.
- The model points toward permissioned agentic commerce, where paid discovery, product questions, and referral traffic are bundled into one platform-mediated journey.
The trend: AI advertising is evolving from sponsored placements toward sponsored agents that mediate commercial discovery and buyer follow-up within the assistant interface.