Be honest: when was the last time you went a full day without opening an AI app? For most of us, these platforms have basically become a new coworker. We use them to solve coding bugs or write that one tricky “per my last email” reply we’ve been putting off.
People are treating ChatGPT, Claude, Copilot, and Perplexity like the new Google. They’re ditching the search bar to ask these platforms everything from “best hiking boots for wide feet” to “local plumbers who work on weekends.” It’s a big shift in how we find things. And you know the golden rule of marketing: where the people go, the ads are sure to follow.
Earlier this year, OpenAI expanded ads to all U.S. free and low-cost “ChatGPT Go” users. This is a platform with 800 million weekly active users (and growing). With that kind of scale, it’s clear that AI advertising will eventually become a must-have on every company’s marketing plan. But the early returns are, well…not as clear. Between high entry costs and underdeveloped reporting tools, jumping in right now feels more like paying for a front-row seat to an experiment than buying actual results.
In this post, we’ll explain what the early data actually shows, why we haven’t recommended this channel to our clients just yet, and what we’re suggesting you do instead to stay ahead of the AI curve.
How Do ChatGPT Ads Work?
Running massive AI models is incredibly expensive. Showing ads to the free tier is how OpenAI plans to pay the bills, especially when hundreds of millions of people aren’t exactly rushing to shell out $20 a month for a subscription. This isn’t surprising. Any platform that hits this kind of scale eventually has to figure out how to turn that attention into revenue.
That also means this category isn’t going away. Over time, the tech will mature, the targeting will get smarter, and the buying process will get easier—whether that happens through ChatGPT or whatever platform wins the long game.
What You’re Actually Buying
Right now, OpenAI is setting the pace. They recently launched their first self-serve ads manager to help businesses manage these new placements. So, what does the actual product look like today?
- Ads appear at the bottom of answers and are clearly labeled as sponsored, separate from ChatGPT’s responses.
- Unlike a billboard on the highway that is shown to everyone who drives by, these ads are matched to the topic of a conversation. If you’re asking for a lasagna recipe, you might see a sponsored link for a grocery delivery service or a specific pasta brand right below the response.
- More than 100 individual brands jumped into the pilot. Retail and grocery were most represented, accounting for about 44% of ads, with travel and food delivery close behind.
The ad placements seem to make sense. They feel like the next logical step in a conversation. But there’s a massive difference between a platform that’s great at showing ads and one that’s actually ready for you to buy them.
Premium Pricing With Beta Reporting
When the ChatGPT ad pilot first launched earlier this year, there was no dashboard to log into and no automated way to buy space. Deals were happening the old-fashioned way: over long email threads, phone calls, and spreadsheets. For a platform built on the cutting edge of tech, the backend felt surprisingly manual.
Early agency partners told The Information that they couldn’t demonstrate measurable results for their clients. Brands were essentially paying for the “cool factor” of being first movers in the AI advertising space. But they didn’t have any way to track whether the money spent was worth it.
OpenAI’s official advice was to just supply more text and visual variations for the AI to test. That’s fair guidance for any ad platform, but it doesn’t solve the underlying problem: if you can’t see what’s happening, trying more things just means spending more money in the dark.
Walmart Walked Away
In November 2025, Walmart launched a pilot called “Instant Checkout” within ChatGPT. It allowed users to browse and buy 200,000 different products without ever leaving the chat window. In-chat purchases converted at one-third the rate of traditional click-out transactions, and Walmart’s EVP called the experience “unsatisfying.” They’re now moving away from the model entirely, embedding their own chatbot, Sparky, inside ChatGPT so purchases are completed within Walmart’s own system.
From Spreadsheets to Self-Serve
Earlier this month, OpenAI launched a self-serve Ads Manager — described by early users as similar to Google Ads — that lets a small group of pilot advertisers monitor performance and optimize campaigns directly, without waiting on a manual report from an account rep. And just this week, they added cost-per-click (CPC) campaigns, meaning advertisers can now choose to pay per view or per click and set their own maximum CPC bid. It’s still early and access is still limited, but the infrastructure is arriving quickly.
Who Should Be Running ChatGPT Ads Right Now?
If a retail giant like Walmart is walking away because the experience felt clunky, then who is the target audience for this right now?
Early on, joining the pilot required a massive investment—somewhere between $200,000 and $250,000. The threshold has since dropped to $50,000. That’s better, but still a pretty penny for most businesses.
Then there’s the CPM. The rate started at $60 at launch and has since dropped to as low as $25, with most buyers seeing somewhere in the $35–$45 range depending on how they’re buying. For context, that still puts the ChatGPT CPM average well above Meta’s average of $4.82 and even LinkedIn’s average of $39.
OpenAI is pricing these placements more like high-intent Google Search than a standard social media scroll. Their theory is that if you are asking an AI for the best solar company, you’re already halfway to a purchase. Whether that theory actually translates to sales at that price point is still the big unknown.
Right now, the only brands that can really justify this spend are the ones with deep pockets that can handle a test that might not pan out, big teams capable of managing a buying process that is being built on the fly, and flexible timelines so they can wait for better data to exist. These companies are really paying for market intelligence. If AI advertising becomes the next major marketing channel, they want to already be fluent in it while everyone else is paying to catch up.
For the rest of us non-Fortune 500 companies, the math just doesn’t add up yet. Being first isn’t always the way to win. Right now, being an early adopter of ChatGPT ads mostly means you are footing the bill for someone else’s learning curve.
The Ghost of Marketing Past
If this feels chaotic and half-finished, that’s because it is, but it’s also to be expected. Every ad platform you rely on today went through exactly this phase.
Google Ads launched in 2000 with a manual buying process that required calling a sales rep. Facebook’s self-serve ad platform didn’t arrive until 2007, years after the platform already had millions of users. (And if you’ve spent any time in Meta Business Manager recently, you know that “finished” is a generous description for a tool that’s been in some form of renovation for a decade.) Netflix launched its ad-supported tier in 2022 at CPMs of $55–$65. Within a year, they had fallen to $20–$30 as inventory scaled.
These platforms didn’t arrive polished and performance-ready. They were built and rebuilt until they became indispensable. The pattern is the same:
- The Audience: A promising new group of people shows up.
- The Product: A primitive, manual ad product is launched to test the waters.
- The Infrastructure: A slow, painful build toward the measurement and scale that professional marketers actually need.
OpenAI is following that same arc. They’re just doing it at “AI speed.” They have already projected that advertising could be worth $102 billion by 2030. To hit a number that big, they have to build a world-class ad engine. They have no choice.
That isn’t a reason to ignore this channel. It’s a reason to watch it closely and be ready to move the second the tools actually match the hype.
Why Organic Wins For Now
The fact that the ad product is still finding its footing doesn’t mean AI search is something you should ignore. It just means the real gold rush right now is happening in organic search, not the paid side.
People are already moving their old Google habits over to ChatGPT. They’re asking for product comparisons, local recommendations, and expert advice. That shift in behavior is happening every single day, regardless of whether the ad platform is ready for your credit card.
The question you should be asking isn’t how to buy an ad. It is whether your organization actually shows up when ChatGPT answers a question about what you do.
This is a content and SEO conversation, not a media buy. AI tools go straight for well-structured, authoritative content that directly answers a user’s specific question. If your website, blog, or resource library is already doing that, you’re building visibility in AI search right now without spending a single dollar.
We’ve written before about how to optimize your content for AI search engines, and the core strategy hasn’t changed: LLMs are looking for clarity and expertise. If your content isn’t showing up, that is the gap worth closing first. You want to be the authoritative answer now, before the paid channel matures and the competition gets a lot more expensive.
AI’ll Be Back (When the Tools Are Better)
We aren’t saying you should ignore ChatGPT ads forever. We’re just saying the time isn’t now. Once the measurement tools catch up, the entry costs settle into a more reasonable range, and the case for a solid ROI becomes clear, we will be the first to let you know it is time to jump in.
In the meantime, the smartest move is to focus on the fundamentals. Building a strong organic presence in AI search is an investment that starts paying off today without the high-priced entry fees.
If you want to talk about how your current content strategy is positioning you for this shift, reach out and let’s chat.