Inside ChatGPT, a
brand's "sponsored conversation" has arrived.
On September 16, 2026, OpenAI announced it had begun testing "Sponsored Agents," a new ChatGPT ad format that lets users step from an ad straight into a conversation with a company's AI agent. Wayfair, Angi, Newegg, Best Buy, Lowe's, and VistaPrint are the six launch partners, and the alpha is limited to a subset of users in the United States.
Until last week, "no ads"
was ChatGPT's baseline
Until now, OpenAI's stated position was that ads would not be woven into ChatGPT conversations themselves. That changed when OpenAI announced on September 16, 2026 that it had added a new ChatGPT ad format called "Sponsored Agents" and begun testing it with a select group of advertisers. Click through an ad, and a user can now talk directly with that company's own AI representative (reported by GIGAZINE and The Register).
The named launch partners are Wayfair (rolling out at limited scale), home-services marketplace Angi (which launched a homeowner-facing agent), electronics retailers Newegg and Best Buy, home-improvement chain Lowe's, and print-and-marketing company VistaPrint — six companies in total. Alongside Sponsored Agents, OpenAI also announced natural-language campaign-building tools for advertisers and integrations with HubSpot and Shopify, framing this less as a one-off experiment than as part of a broader push to mature ChatGPT's ad business (as detailed by Innovatopia).
The test, by the numbers
An ad becomes
a conversation
OpenAI's own example: a shopper who spots a dining table in an ad.
You encounter an ad inside ChatGPT
A sponsored card from a business appears within the flow of a conversation. It's marked as "Sponsored" and visually set apart from ChatGPT's own answers.
You enter a clearly labeled conversation
If interested, the user can open a conversation with that company's agent directly from the ad. For the dining table, that might mean asking whether it fits the room, how many people it seats, or how to care for the finish — follow-up questions answered on the spot.
It never mixes with the main thread
The Sponsored Agent exchange is kept separate from both the user's original ChatGPT conversation and ChatGPT's own independent answers. Once satisfied, the user can click through to the company's website to buy or inquire further.
The advertiser's pitch no longer waits
behind a click — it lives inside the conversation itself.
Why this matters now
Search advertising has always meant "click a link, then go verify on someone else's site." Sponsored Agents pulls that verification step into the ChatGPT conversation itself. Users can resolve questions through dialogue before ever leaving ChatGPT for the brand's site — a shift from click-based search ads toward agent-based advertising. For OpenAI, it's also a way to diversify revenue beyond subscriptions alone, and a clear reversal of the earlier posture that ChatGPT conversations would stay ad-free.
Who it affects, and how
Business / leadership
If your company becomes an advertiser, a conversational storefront becomes one more channel alongside search and social ads. But the alpha is limited to a handful of US companies, and there's no public timeline yet for expansion to other regions or to smaller advertisers — this is still a wait-and-see stage.
Marketers
The natural-language campaign tools and HubSpot/Shopify integrations announced alongside Sponsored Agents directly touch ad-ops workflows. If you're evaluating ChatGPT ads, it's worth checking the Help Center specs and measurement approach now, and confirming whether your CRM stack is already supported.
Product managers
If your product embeds ChatGPT answers or the API, it's worth watching whether a similar ad layer eventually reaches your own answer surfaces. For now this is scoped to ChatGPT's own ad placements and does not affect API usage.
| ChatGPT before | After Sponsored Agents |
|---|---|
| No ads inside conversations, in principle | An ad card can open a conversation with a brand's agent |
| Reaching a company meant an outbound link | Ask in-conversation, then click through to the site |
| Revenue centered on subscriptions | Subscriptions plus a growing advertising revenue stream |
Counterpoints, risks, limits
OpenAI says ads "don't influence the answers ChatGPT gives you," and that whenever an ad appears it is "always clearly labeled" and visually separated from organic answers. Whether that promise holds up is a separate question. Since OpenAI itself designs the line between ad and answer, there's a structural concern that stays regardless of intent: how strictly that line is enforced depends entirely on OpenAI's own ongoing operational choices.
If ads feel "clumsy or opportunistic," users can switch to a rival chatbot with little friction, which gives OpenAI its own incentive toward restraint. Still, this is an early test: implementation quality of the disclosure labeling, how advertisers are vetted, and whether free and paid tiers see different treatment are all details the public record doesn't yet settle. Whether paid-plan users (Plus/Pro, etc.) will see these ads at all has not been stated, so how the free-tier experience actually changes will only become clear as the rollout widens.
What to watch next
Three things to track. First, how fast the region and advertiser list expands — from a US-only alpha to other markets and industries. Second, whether OpenAI ends up drawing a real distinction between free and paid tiers in ad exposure. Third, whether OpenAI clarifies how Sponsored Agent conversation data feeds back into ad targeting or ChatGPT's own future answers. Companies using ChatGPT for business purposes should keep an eye on the Help Center and terms of service until these three points are spelled out, given how sensitive some of those conversations can be.