ChatGPT Sponsored Agents: OpenAI's Ad Bet Explained
OpenAI is done tiptoeing around advertising. On Wednesday, the company said it has begun testing "Sponsored Agents" — advertiser-run AI agents that ChatGPT users can talk to after clicking an ad — while rolling out a suite of AI-powered tools that let brands build, run, and analyze campaigns inside ChatGPT from a natural-language prompt. HubSpot signed on as the first CRM partner for ChatGPT Ads, and Shopify as the first e-commerce partner, with Shopify's app going international on September 23.
It is the most aggressive step yet in a reversal that has been building for two years: from CEO Sam Altman saying he "hates" ads to OpenAI selling AI-native advertising to the biggest names in American retail.
How Sponsored Agents work
A Sponsored Agent is an advertiser-run chatbot that lives inside ChatGPT. You click an ad, a clearly labelled chat window opens, you ask the brand's agent questions about what you want, and only then click through to the brand's website. OpenAI says the agent stays separate from ChatGPT's own answers and from your original chat, and the test is running with select US advertisers.
From ad to brand agent
The Sponsored Agents feature, first reported by Reuters, lets users start a conversation with a business-sponsored agent and then visit the business's website after interacting with an ad in ChatGPT. The agents will be clearly labeled and, OpenAI says, will remain separate from ChatGPT's independent responses and the user's original chat. The test is running with select U.S. advertisers.
In practice, the ad itself becomes a doorway into a sales conversation. Users can spell out what matters to them, pose follow-up questions, and click through to the business's website when they are ready to move forward, according to OpenAI's announcement. The format builds on what the company showed select clients earlier this month: an AI-native ad whose call-to-action — a button reading "Chat with us" — opens a branded chat inside ChatGPT rather than kicking users out to a website, as reported by Digiday, which viewed video of the format and confirmed that Wayfair is trialing it.
The move telegraphs a deliberate strategy. At Goldman Sachs' Communacopia and Technology Conference on September 8, OpenAI CFO Sarah Friar called the current ads shown beneath ChatGPT responses the "basic starting point" and said the company was already seeing "early indications" of what a more AI-native format could look like. "This is all before we've really launched a format that feels truly endemic to AI," she said, per Digiday.
An Ads Manager that runs on prompts
Alongside the agent pilot, OpenAI introduced the ChatGPT Ads Manager plugin — effectively an AI media buyer. Advertisers can use natural-language prompts to create, update, and analyze campaigns, turn a website or campaign brief into a campaign, review performance, and receive recommendations. The manager can suggest copy and images based on an advertiser's landing page and campaign objective.
A second, optional feature adds AI text customization: it can tailor an advertiser's existing headlines and descriptions to fit the context of a conversation, and automatically translate ad copy into a user's preferred language.
Put together, OpenAI is deploying generative AI on both sides of the advertising transaction — one model deciding which ad a user sees and how it reads in context, another set of models helping the advertiser build and optimize the campaign in the first place. It is a vision of advertising with almost no human media buyer in the loop.
The first advertisers: Wayfair, Angi, and America's big-box heavyweights
The launch roster reads like a tour of American big-ticket retail. Wayfair is participating at limited scale, while Angi has launched a homeowner-facing agent, with Newegg, Best Buy, Lowe's, and VistaPrint also named as early advertisers.
Angi's bet is illustrative of how the format is supposed to work. The company's branded agent surfaces in ChatGPT conversations about home services and routes homeowners into its service-request flow. "Our AI-first strategy is simple: make it easier for homeowners to connect with and hire the right pro for their project," said Angi co-founder Angie Hicks in the announcement.
It is a shrewd beachhead for OpenAI. Home services, furniture, and electronics are high-consideration purchases where shoppers genuinely do research, compare, and ask follow-up questions — exactly the behavior ChatGPT already hosts. If the sponsored agent can shorten the path from "help me pick a contractor" to a booked job, advertisers may care far more about ChatGPT placement than about the impressions it generates.
HubSpot and Shopify: the plumbing of the ad machine
The announcement's least flashy details may be its most consequential. HubSpot is OpenAI's first CRM partner for ChatGPT Ads: businesses that manage customers in HubSpot can connect a ChatGPT Ads account and, without leaving HubSpot, create ads, monitor performance, and follow up on leads, powered by their existing HubSpot context. According to the app's marketplace listing, marketers can build creative, target relevant conversations using free-form "context hints" with location targeting, and set budgets and schedules, with pricing on CPM, CPC, or oCPC models.
Shopify is the first e-commerce partner, and the push into retail is moving fast. A free ChatGPT Ads app in the Shopify App Store — developed by OpenAI — lets U.S.-based merchants create and manage campaigns without leaving Shopify, with product inventory syncing automatically from their stores and products already integrated through the Shopify Catalog. Campaigns run on a Shopify pixel and conversion events, and merchants get analytics covering ad spend, click-through rates, conversion tracking, and cost per acquisition. OpenAI said the app will launch internationally in markets where ChatGPT Ads is available starting September 23.
Between the CRM hook and the storefront hook, OpenAI is assembling the unglamorous integrations that turn an ad experiment into an ad business: the lead goes to HubSpot, the sale comes through Shopify, and ChatGPT sits in the middle as the storefront.
From "I hate ads" to an AI-native ad business
Few reversals in tech have been as public as Altman's on advertising. In a 2024 interview, he said he "hates" ads and called combining them with AI "uniquely unsettling," though he allowed that "I'm not saying OpenAI would never consider ads," as CNN reported. Last year he softened further, saying he wasn't "totally against" ads in ChatGPT but that it would "take a lot of care to get right."
The care-taking began in earnest this year. When OpenAI announced in January that ads would arrive on the free and Go tiers, it teased the logical endpoint: "Soon you might see an ad and be able to directly ask the questions you need to make a purchase decision." The formal test started with logged-in adult users in the U.S. on the free and $8-a-month Go plans, under principles OpenAI laid out in a February blog post: ads would not influence ChatGPT's answers, conversations would stay private from advertisers, ads would not run on regulated topics such as health, mental health, or politics, users under 18 would see none, and free users could opt out in exchange for fewer daily messages.
The rollout drew fire immediately. Anthropic mocked the whole idea in February Super Bowl commercials showing AI chatbots derailed by clumsy ads — and Altman snapped back on X, calling the ads "dishonest" and Anthropic an "authoritarian company," as TechCrunch reported. OpenAI had also taken heat late last year for testing app suggestions that many users read as unwanted ads. It is the same rivalry playing out in product strategy: Anthropic's competing bet on the AI workday puts its agent inside work tools rather than inside ads.
Altman has since tried to draw a bright line. "We will not accept money to influence the answer ChatGPT gives you," he wrote on X, arguing that subscriptions alone cannot fund the scale of demand — "many people want to use a lot of AI and don't want to pay" — and that paid tiers would stay ad-free. Wednesday's announcement is the company betting that the line can hold while the ad product gets dramatically more intimate.
Why the money matters now
Advertising is no longer a side experiment for OpenAI; it is becoming load-bearing. Bloomberg reported in August that the company's annualized revenue run rate had topped $40 billion — roughly double where it ended 2025, when CFO Sarah Friar put the figure above $20 billion — with growth driven in part by AI coding software, subscriptions, and the nascent advertising business. Enterprise revenue has already overtaken consumer subscriptions as the company's largest line.
That financing push — see OpenAI's reported $1.2 trillion pre-IPO valuation — is unfolding against a race with Anthropic, which has filed confidential IPO paperwork of its own. And a source close to OpenAI told CNBC that ads are expected to account for less than half of revenue in the long run, per Engadget's reporting — a ceiling that suggests the company wants ads as a growth engine, not its identity.
The competitive pressure is real. Meta began using people's interactions with its Meta AI chatbot to target ads last December, and every major AI lab is now hunting for revenue beyond subscriptions. OpenAI's bet is that the conversation itself is the premium inventory — and that an agent you can interrogate beats a banner you ignore.
ChatGPT Ads goes global: the India expansion
ChatGPT Ads has already gone international. OpenAI launched the advertising program in India on August 27, 2026 — one of ChatGPT's largest markets — with WPP and Omnicom signed as the first agency partners and more than 50 brands expected to go live. A self-serve ChatGPT Ads Manager opened to Indian businesses from September 4, with daily budgets starting at ₹725 (about $9), signaling that OpenAI intends sponsored formats to scale in emerging markets, not just the United States.
The guardrails — and the questions still open
OpenAI's stated guardrails are extensive. Sponsored agents are labeled and visually separated from organic answers; the company insists ads do not influence ChatGPT's responses; conversations are not shared with or sold to advertisers; users can dismiss ads, view and delete their ad history and interest data, and toggle personalization off; and ads stay away from regulated topics and users under 18.
Can I opt out of ChatGPT ads? Yes — free users can opt out in exchange for fewer daily messages, and paid tiers stay ad-free.
But the announcement leaves consequential questions unanswered. OpenAI has not disclosed how frequently a sponsored agent will surface in a given conversation, or how the model decides which advertiser's agent to trigger — and that selection is itself a recommendation, made by an AI, on commercial grounds, inside a product people consult on consequential decisions. Given the lawsuits alleging ChatGPT encouraged users' suicides, the pressure to police which products and services the assistant effectively endorses will only intensify. The question sits inside the industry-wide AI safety debate over how much responsibility labs bear for what their models endorse.
The deeper question is one Altman himself raised in 2024: whether advertising is "uniquely unsettling" when fused with AI. Sponsored Agents are OpenAI's answer — not a banner beside the answer, but a brand inside the conversation. Whether users experience that as helpful or as the unsettling part will decide whether this becomes the ad format of the AI era, or a cautionary tale Anthropic keeps replaying at the Super Bowl.
Sources: Reuters (Sept. 16, 2026; reporting by Anzar Mehraj); OpenAI's "Testing ads in ChatGPT" blog post (originally published Feb. 9, 2026) and its September 16 advertiser announcement, as detailed by AI Weekly; Digiday (Sept. 14, 2026); TechCrunch (Feb. 9, 2026); CNN (Jan. 16, 2026); Search Engine Land (Feb. 2026); Engadget; eWeek; Bloomberg (Aug. 13, 2026). App-listing and integration details via Unite.ai and Securities.io summaries of the announcement.
Comments
Post a Comment