Inside ChatGPT’s U.S. Product Ad Trial: What Marketers Need to Know

Ads have officially entered the Chat.

Ever since this year’s Super Bowl, marketers have been asking the same questions: How will ads show up in AI conversations? And how will users respond? Now that ads are being trialed in the U.S., the picture is starting to become clearer.

In our recent podcast, Zuber Alim, Senior Manager of Search & Social and Chris Khan, Director of Search & Social in the U.S., unpack the real impact of this shift. Especially how it could reshape user trust, measurement and how brands will start showing up in AI conversations.

Watch the full podcast here, and read on for highlights.

A Trusted Confidant Turns Advertiser?

Conversational AI has become a part of everyday life. On a daily basis we turn to it for answers, recommendations, and seek solutions to our problems. We carry them like a trusted confidant in our pockets. Now that these AI assistants are going to recommend brands to users, maintaining that sense of trust will matter more than ever.

“This is a totally new phase of advertising, one that is budding right before our eyes.”

Christopher Khan, Director of Search & Social – US, M+C Saatchi Performance

When search ads first launched, optimization meant keywords, CTR, and traffic. When social followed, the focus shifted toward engagement and reach, not just intent. The real disruption came when both channels became deeply measurable and truly performance-driven.

ChatGPT’s current model feels earlier in that cycle. It’s going to be CPM-based, with no auction dynamics and mainly focused on impressions rather than clicks. But at the same time, it sits closer to search than social due to high contextual relevance and intent signaling.

If decisions are being shaped inside the conversation itself, influence can happen without a click or a visit. For SEO and performance teams, that means the battle may be won or lost before a click even occurs.

What Do We Know So Far?

Ads will appear at the bottom of the ChatGPT interface as banners or swipeable product carousels taking up roughly 25–30% of the screen depending on query intent. Some formats will even allow users to engage directly with the brand via chat.

These ads are expected to remain separated from chat responses and not influence the answers themselves. Something Sam Altman emphasized earlier this year in response to Anthropic’s Super Bowl campaign.

And who are the recipients? Free and Go-tier users. As the vast majority of ChatGPT users are on the free tier, no doubt it will bring an additional revenue stream into the business, mirroring the Google and Meta ad playbook.

The Million-Dollar Question: Adoption

It may be a harder sell than search as users come here for trusted answers. Including ads to this dynamic could erode confidence. But with 800M users, the scale is undeniable. User adoption hinges on keeping ads non-intrusive and answers uncompromised.

For marketers, traditional attribution might not work for measurement and brands will need to rely on incrementality, lift studies, and media mix modelling.

What Should Marketers optimize in AI conversations?

Ad placement within an AI answer is very different from search or paid social. Instead of a single query, a series of prompts builds context over time and that evolving conversation becomes the basis for ad placement. So marketers need to:

• Be clear and authoritative about what they do.
• Provide clear, helpful answers to real customer questions.
• Make services and offerings easy to understand.

Let’s Talk Money

In the first half of 2025 alone, OpenAI generated around $4.3B in revenue while incurring roughly $7.8B in operating costs, which shows just how expensive large-scale AI infrastructure really is. Ads and subscriptions are the two clear levers for growth, with ambitions to reach $25B in revenue by 2029. Monetizing its large base of free users is one obvious way to get there.

If this channel scales, bigger brands will likely be the first to test it. Broader adoption will come once self-serve tools mature and measurement moves beyond simple impressions and clicks.

Testing Early? What Then?

Let’s be clear, this isn’t a light experiment. Entry starts at a minimum of $200K, with CPMs around $60, similar to early Netflix ad buys.

The sweet spot lies in high-intent moments, particularly decision-support and comparison queries where users are actively evaluating their options.

If you’re not a well-established brand, it’s worth proceeding with caution. AI environments bring greater scrutiny, and unclear positioning can quickly unravel in a conversational setting.

What Should Brands Do in the Next 90 Days?

Brands need to build a strong technical base for GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) and restructure their content in an answer-first format that’s clear, concise, and machine-readable. This is a crucial lever.

Its equally important to define your “one-sentence recommendation” which is a sharp articulation of who you are and what you solve. And move beyond click-based measurement by tracking assisted influence, sentiment, and consideration lift to understand how AI conversations are shaping demand.