Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

OpenAI Pitches AI-Native ChatGPT Ads That Open a Brand Chat, Not a Website

Digiday reports that OpenAI has introduced a new AI-native ad format to select clients, attaching a branded business agent to the ad so that a click opens a chat inside ChatGPT instead of sending users to the advertiser's site. Wayfair is trialing the format, and OpenAI CFO Sarah Friar has described today's response ads as only a "basic starting point."

Published

OpenAI has begun offering select clients a new kind of AI-native ad built specifically for its chatbot app, according to a Digiday report published on September 14. Instead of turning an ad click into a browser tab, the format turns it into a conversation with a brand's AI agent. Digiday says it viewed the format via video and verified it.

Advertisers can attach a branded business agent to an ad. When a user taps the call-to-action button — a message such as "Chat to us" — they land in a branded chat inside ChatGPT and interact with that brand's agent, rather than being sent to the advertiser's website.

The format is an early iteration of what OpenAI has described as AI-native advertising: ads designed around the way people interact with AI assistants. It goes beyond the basic units OpenAI currently serves beneath ChatGPT responses for users on its free and Go plans, because the ad itself becomes the entry point into a conversation.

Speaking at Goldman Sachs' Communacopia and Technology Conference in San Francisco on September 8, OpenAI CFO Sarah Friar described those response ads as "the basic starting point" and said the company is 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.

It is unclear how many advertisers have been given access. A screenshot of the ad viewed by Digiday shows Wayfair trialing it. Farhad Divecha, group CEO of Accuracast, compared it to Meta ads that send people to WhatsApp: a way to open lines of communication with prospects, then qualify the ones who need customer service or sales — or route them to the page where they can buy.

For OpenAI, the format could also address a practical problem. Seven months into ads in ChatGPT, participating advertisers still describe the platform as promising but unproven, and most say it is too early to judge its role in a media plan. Until returns improve, spending stays inside incremental or experimental budgets. Shamsul Chowdhury, svp of paid media at Zeno Group, put it bluntly: ChatGPT has been a hot new platform for advertisers, but "the ROI just hasn't been there for the most part."

The new format raises a trade-off of its own, particularly for retailers: is there a benefit to keeping a potential customer inside ChatGPT rather than sending them to their own site? Chowdhury said retailers want user data through their owned properties, but they also know the consumer journey can happen anywhere and that they have to vie for those eyeballs and dollars.

Timing matters too. With the holiday shopping season approaching and advertisers facing rising costs and flat returns on platforms such as Google and Meta, OpenAI could become an alternative channel for some of that spending — provided the agent experience is seamless. OpenAI did not respond to Digiday's request for comment.

Why it matters

It is OpenAI's first real step from response-adjacent ad units toward advertising that is native to conversation, and it forces a fresh negotiation over traffic, customer data and where a purchase actually happens.

OpenAIChatGPTAdvertising
Back to realtime news

Nearby Updates

All

09/14, 11:55

Chinese Models Lead Weekly Token Volume for 20th Straight Week as DeepSeek V4.1 Flash Hits No. 6

National Business Daily, using the latest OpenRouter data, calculates that global large-model token consumption reached 127 trillion tokens in the week of Sept. 7 to Sept. 13, with Chinese models at 61.17 trillion tokens, leading for a 20th consecutive week. DeepSeek V4.1 Flash, released Sept. 10, climbed to sixth place within three days, while Chinese models took four of the top five slots.

09/14, 11:01

DeepSeek-V4.1-Flash: 552B-Parameter MoE Built for Efficient Inference

A technical breakdown of the DeepSeek-V4.1-Flash model card describes a multimodal mixture-of-experts model with a 552B-parameter backbone that activates only about 8B parameters at a time. The weights ship under an MIT license with a 1M-token context window, but the model still trails the larger V4-Pro family on several reasoning benchmarks.

09/14, 09:51

DeepSeek Routes All V4-Pro API Traffic to V4.1-Flash at Flash Rates

DeepSeek now routes all API traffic aimed at V4-Pro to V4.1-Flash and bills it at Flash rates, according to Pandaily. Requests that need V4-Pro capability are therefore answered by V4.1-Flash, changing the cost picture for callers.

09/14, 08:30

Zhipu Teases GLM-6.0 in a Financial Filing, Revealing a Fully Self-Trained Approach

Zhipu has slipped the first details of its next flagship, GLM-6.0, into a financial filing rather than a technical blog or paper, according to QbitAI. The filing points to a fully self-trained approach, while the model itself and its accompanying paper have yet to be released.