ChatGPT In-Conversation Ads: What AI Agents Mean for CX

What happened

OpenAI is testing an advertising format within ChatGPT that goes well beyond conventional display or search ads. Rather than simply serving a sponsored link, the experimental model reportedly triggers a business-specific AI agent when a user engages with an ad — one capable of answering product questions, making recommendations and capturing leads directly inside the conversation.

The format, reported by MarTech, represents a meaningful departure from the click-through paradigm that has defined digital advertising for three decades. Instead of routing a prospective customer to a brand's landing page, the interaction continues within ChatGPT itself, with the agent acting as an always-on, conversational front-of-house for the advertiser.

Why it matters

For customer experience and service design practitioners, this development is significant because it collapses the traditional post-click journey — the sequence of landing page, form fill, chatbot hand-off and follow-up email — into a single, uninterrupted conversational moment. The friction points that typically cause drop-off between ad exposure and conversion are, in principle, removed. From a behavioural economics standpoint, this matters enormously: every additional step in a customer journey is an opportunity for effort-aversion to kick in. An agent that resolves intent at the point of discovery removes those steps entirely.

It also shifts the competitive battleground for brands. If OpenAI's model gains traction, the quality of a brand's AI agent — its ability to answer accurately, recommend relevantly and handle objections gracefully — becomes a direct determinant of commercial performance, not just a back-office efficiency tool. Service design, in other words, moves upstream into media spend.

The Renascence take

Most commentary on this story will focus on the advertising and revenue implications for OpenAI. The more consequential question for operators is what it means when the customer's first substantive brand interaction is mediated entirely by an AI agent they did not choose and cannot fully control.

The real risk here is not that the ad format fails — it is that brands sleepwalk into it without designing the agent experience with the same rigour they would apply to a flagship store or a contact-centre script. Behavioural research consistently shows that a single friction-laden or tonally mismatched interaction disproportionately damages trust relative to the effort it took to build it. Operators who treat the in-conversation agent as a lead-capture mechanism rather than a genuine service moment will find that the format amplifies disappointment just as efficiently as it promises to amplify conversion. The discipline required is not prompt engineering — it is experience design: knowing what the customer actually needs at that moment, and building the agent around that need rather than around the funnel.