The field? ChatGPT is not like Google or social media. It’s a new kind of environment where advertisers reach users not through search results or interruptions in the news feed, but in the middle of a conversation. OpenAI is reportedly asking advertisers to commit nearly $1 million upfront for access, charging per view rather than per click. Ads will appear for both logged-in free users and subscribers of the new $8/month ChatGPT Go plan.
Behind this aggressive move lies a practical need: computing costs. OpenAI expects to burn at least $9 billion this year, and advertising is one way to offset that fire. But can this high-intention, conversation-driven model really deliver the returns marketers expect at this price?
The rise of the ‘intention economy’
To understand the use of OpenAI, you must understand the shift it represents. In the traditional search world, advertisers bid on keywords. They hope to intercept a user’s intent in the split second it comes to light. This intention is already known in ChatGPT.
“The fundamental difference is that ChatGPT ads work in a post-intent environment,” says Caroline Giegerich, VP AI at IAB. “With search, you bid on keywords to intercept the intent as users express it. With conversational AI, the intent has already been identified through dialogue. The AI not only knows what you are looking for, but also the context and limitations behind it.”
Dig deeper: AI is rebuilding the inbox and email marketers have no control over it
“Search advertising operates in the attention economy,” said Calvin Scharffs, VP Marketing at Orange 142. “Users type short, ambiguous queries, and advertisers infer intent from keywords, past behavior, and probability models. ChatGPT operates in the intent economy. Users express their needs directly and entirely in natural language. There is no guessing.”
That shift changes not only how ads are targeted, but also when and how they appear. “Rather than interrupting a search results page,” says Scharffs, “brands can appear at the exact moment when intent is fully formed, within a decision the user has already defined.”
Data precision versus data opacity
If this model is so accurate, why the hesitation? First, marketers don’t yet know what kind of data they’re getting. OpenAI likely uses call data internally to serve relevant ads, for example by showing travel ads when a user mentions a trip. But it’s unclear what parameters advertisers will see, what counts as ‘personalization’, or how detailed targeting actually is.
“Most likely, OpenAI uses conversation data internally to inform targeting (e.g., ‘this person discusses travel, shows travel ads’) without sharing raw conversation content with advertisers,” said Nicole Green, VP Analyst at Gartner. “But the vague language leaves critical questions unanswered: What targeting parameters do advertisers actually get? What does ‘personalization’ mean? What performance or audience data is provided?”
Dig deeper: why AI visibility is now a C-suite mandate
Ultimately, unless OpenAI clarifies what data and controls advertisers actually get in ChatGPT, it’s impossible to fairly compare its targeting capabilities to Search, which has had decades to build a mature system of audience insights, behavioral tracking, and attribution.
And then there’s the matter of attribution.
When a user sees a product recommended by the AI and also sees an ad for it, what led to that click? Was it the AI logic or the ad placement? That blurring of organic versus paid creates complexity that search advertising doesn’t have.
“Clear, consistent labeling is essential,” says Scharffs. “If users can’t immediately see what’s paid and organic, trust quickly erodes.”
Governance, risk and the hype gap
Groen sees governance as one of the greatest challenges ahead.
“We have never worked with any technology in this capacity before,” she said. “For marketing, this means we really need to raise our risk profile. Especially for regulated sectors, this means we need to bring compliance and legislation into the conversation much earlier.”
She also pointed out the hype gap: “AI’s biggest impact on marketing this year was the need to manage expectations. There’s so much energy around GenAI, AI agents, emotion AI, machine customers, but the execution maturity isn’t always there.”
What comes next is likely to be more radical. As AI agents begin transacting on behalf of humans, marketers will need to consider not only human journeys, but machine journeys as well.
Preparing for machine customers
“We predict that by 2030, 15% of sales will come from machine customers,” said Green. “Advanced AI agents will buy and sell just as businesses do today. As we look to the future of marketing, it will become even more difficult. We will need human journeys, machine journeys, and combinations of both.”
That changes the way marketers approach everything from messaging to UX and data structure. Caroline Giegerich advises that brands should now treat their reputation as a data set.
“Your product information must be AI-readable,” she said. “If an AI can’t easily understand what you’re selling, who it’s for, and why it’s differentiated, you won’t be recommended…Brand reputation becomes algorithmic input. Reviews, mentions, and third-party validation all become a signal for whether an AI recommends you.”
And in this new environment, direct brand relationships could experience a renaissance.
“When a user expresses clear intent in a conversational environment,” Scharffs said, “there is an opportunity for a brand to emerge immediately – with its own voice, terms and experience.”
Whether OpenAI’s $60 CPM demand is justified remains to be seen. But it’s clear that AI-powered environments are challenging long-held assumptions about how advertising works, what targeting means and what a customer actually is.
Energize yourself with free marketing insights.
MarTech is owned by Semrush. We remain committed to providing high-quality reporting on marketing topics. Unless otherwise stated, the content of this page was written by an employee or paid contractor of Semrush Inc.
#ChatGPT #Ads #Worth #Money #NFL #Level #MarTech