Analysis
OpenAI began testing "Sponsored Agents" with a small set of advertisers in the United States, letting ChatGPT users who click a relevant ad start a clearly labeled conversation with a business-run agent inside the app, according to ppc.land and OpenAI's own announcement.
How It Works
After seeing a relevant ad inside a ChatGPT conversation, a user can opt to continue chatting with an agent sponsored by that business -- explaining what they need, asking follow-up questions, and eventually following a link to the advertiser's site. OpenAI has designed the sponsored conversation to be visually and structurally distinct from ChatGPT's own independent answers, addressing an obvious concern about blurring paid content into the model's native responses.
“Roundel, Target's retail media arm, is also participating in a related ad-pilot program.”
The Launch Advertisers
Angi (formerly Angie's List), the home-services marketplace, is among the first publicly identified brands in the pilot -- an Angi-branded agent can appear in ChatGPT conversations about home repairs or renovation, continuing the discussion and gathering project details. Roundel, Target's retail media arm, is also participating in a related ad-pilot program. OpenAI is separately bringing ChatGPT Ads tooling into HubSpot and Shopify, giving small and midsize merchants self-serve access to build and run their own campaigns without an enterprise sales relationship.
Competitive Landscape
- Google -- the incumbent in search advertising and the most direct long-term competitor; Google has its own AI Overviews and Gemini-based ad experiments but has moved more cautiously on conversational ad formats than OpenAI's Sponsored Agents.
- Meta -- dominant in social advertising, has been building AI-driven ad-creative tools but has not launched an equivalent in-chat sponsored-agent format inside its own AI products.
- Perplexity -- has experimented with its own advertising formats inside its answer engine, a smaller-scale but conceptually similar approach to monetizing an AI-native interface.
Why This Matters For OpenAI's Business
OpenAI has historically monetized primarily through ChatGPT subscriptions and API access. Advertising represents a potentially much larger revenue pool -- Google's core search-ad business alone generates well over $100 billion annually -- but also carries reputational risk if users perceive sponsored agents as degrading ChatGPT's usefulness or trustworthiness as an assistant. OpenAI's decision to start with a narrow, high-intent category like home services, rather than broad brand advertising, suggests a deliberate attempt to prove ROI before scaling into categories where the line between "helpful agent" and "ad" is less obviously in the user's interest.
What To Watch
The test remains small and U.S.-only for now, and OpenAI has not disclosed pricing, targeting mechanics or how sponsored-agent conversations will be labeled at scale. Regulatory scrutiny is a real risk: the FTC and state attorneys general have shown increasing interest in how AI assistants disclose paid placements, and a chatbot agent that can converse at length on a sponsor's behalf sits in murkier disclosure territory than a traditional banner ad. How OpenAI handles that labeling as the pilot expands will likely determine how fast advertisers beyond Angi and Roundel sign on.