AIM Media House

Target Among First to Test Ads Inside ChatGPT

Target Among First to Test Ads Inside ChatGPT

Early traffic data pushed the retailer to test ads where product discovery is starting to happen

Target Corp. says traffic from ChatGPT to its digital properties has been growing roughly 40% month over month, a data point that helps explain why the retailer is now among the first companies to test contextual advertising inside the AI chatbot.

The pilot, announced yesterday, pairs Target with OpenAI and Target’s retail media unit, Roundel, and introduces clearly labeled ads that appear alongside shopping-related conversations. Ads are triggered by keywords in a user’s prompt and shown separately from the AI’s responses, which Target says they do not influence, according to its corporate announcement.

Users were already using ChatGPT to research products, compare options, and plan purchases, and some of that activity was already sending traffic to retailers. Target’s decision to participate in the pilot treats conversational AI as another entry point into retail discovery, rather than a standalone experiment, the company said.

How ChatGPT Fits Into Retail Discovery

The test follows OpenAI’s decision to begin advertising trials inside ChatGPT for Free and Go tier users in the United States. OpenAI says ads are clearly labeled, visually separated from answers, and designed not to influence responses. Paid tiers, including Plus, Pro, Business, Enterprise, and Education, remain ad-free, according to OpenAI’s advertising policy overview.

Ads appear based on the context of a conversation. OpenAI says placement relies on the topic of the prompt and recent conversation context, rather than personal profiling, and that conversations are not sold to advertisers. Users can dismiss ads and view explanations for why an ad appears.

Conversational prompts differ from most existing advertising environments because interactions begin with a user-initiated question. Prompts are written in natural language and can contain specific product or comparison intent. That structure places conversational AI closer to search than to social or display advertising, where exposure often precedes intent.

Earlier attempts to turn assistants into shopping channels faced structural limits. Voice-based systems such as Amazon Alexa and Google Assistant struggled with limited visual context, weak attribution, and friction between intent and transaction. ChatGPT operates on screens, supports links, and allows users to move directly from research to external sites, addressing several of those constraints.

This Is a Retail Media Experiment First

Although the ads carry Target branding, the broader implications center on Roundel. Like other retail media networks, Roundel has grown by monetizing on-site placements tied to first-party data. It has also expanded off-site, supporting campaigns across connected TV, social platforms, and the open web. Advertising inside ChatGPT extends that expansion into an environment Target does not own or control.

The pilot relies on strict separation between advertising and AI responses. OpenAI has said ads are labeled, visually distinct, and prevented from shaping answers. These constraints align with established advertising disclosure standards and are intended to preserve neutrality in the assistant experience, according to OpenAI’s published advertising principles.

OpenAI has also invested in ad integrity infrastructure, according to reporting, signaling that governance is being addressed alongside monetization. Other AI companies have taken different positions. Anthropic, for example, has marketed its products as ad-free, highlighting diverging approaches to commercialization in conversational AI.

For Target and Roundel, the central operational challenge is measurement. Retail media networks rely on closed-loop attribution that links ad exposure to purchase behavior using first-party data. When discovery occurs outside owned platforms, attribution becomes more complex.

Identity resolution, conversion tracking, and incrementality measurement are harder when traffic originates in a third-party conversational interface. Target has not disclosed how it plans to address those issues.