Of all the new features OpenAI rolls out every day, the ones we tend to spend the least time on are those related to its advertising business. Honestly, there is a great deal of news to cover in our industry and, when it comes to choosing where to focus, I usually stick with the stories that directly affect our market. After all, GPT Ads are not yet available in Europe.
But that does not mean that, in the meantime, Sam Altman’s company has not been slowly building a powerful advertising platform. Today, we will review some of its most recent features. Who knows—if ChatGPT ads ever make it to our region, we will at least be up to speed on everything they can do 😉
Honestly, this is one of the features that has surprised me the most among the ones OpenAI has added recently, and it challenges some of the assumptions we might have had about its advertising platform.
This new feature, launched in June, follows the same philosophy found on traditional platforms such as Meta (Facebook) and Google Ads. As a result, you can now upload customer or prospect lists through CSV or TXT files (up to 500 MB or 5 million records) using email addresses or phone numbers.
Specifically, you can use these audiences to, logically, show ads only to people who already know your brand or who are in advanced stages of your sales funnel, but also to keep your current customers from seeing acquisition ads, helping you save budget.
In addition, it allows you to pay more or less for a click depending on whether the user belongs to your list: “This will allow you to bid more competitively for your high-value customers or more selectively for lower-priority audiences,” OpenAI explains in its documentation. This is managed through the use of multipliers, which make it possible to increase or decrease the maximum bid when the user belongs to a given custom audience. “For example, you can set a 2x multiplier for high-value customers, a 5x multiplier for people who have already interacted with your brand, or a 0.5x multiplier for lower-priority audiences,” they explain.
That said, to use this custom audience feature you will need a minimum number of contacts: at least 25,000 matched users before it can be used. “We recommend audiences of at least 100,000 users,” OpenAI notes.
Until now, the system had been heavily focused on traditional CPC. Now the Conversions objective has appeared, using an Optimized Cost Per Click (oCPC) model designed to show ads to users who are more likely to convert (make a purchase, sign up, submit a lead…), although the advertiser still pays per click.
AppsFlyer and Adjust are two mobile app attribution platforms that are widely used internationally. With ChatGPT’s new integration with them, an advertiser can gain more information about its campaigns, including app installs, and properly manage attribution, which is essential for many brands.
Meta has a similar feature, called Advanced Matching, and OpenAI was likely inspired by it. It involves sending encrypted data to improve conversion attribution, boosting the algorithm’s ability to understand which clicks end up converting.
Until now, the daily limit was fixed, but in the future it will work like Google Ads: if there are more conversion opportunities on one day, it will spend more budget, and on another day it will balance that out by spending less. The idea is that the budget cap will be calculated over a seven-day window.
ChatGPT has added more advertising features in recent weeks, but these are some of the most eye-catching. In any case, the documentation explaining how it works is public, and I encourage you to browse through it to stay on top of the latest developments. Keep in mind that advertising is currently focused on the U.S. market (although it is beginning to launch campaigns in other markets such as the United Kingdom, Canada, Australia, New Zealand, Japan, and South Korea), where users can already receive ads that appear beneath ChatGPT responses, looking something like this.
As you can see, the ads include classic elements of these types of campaigns, such as the advertiser’s name, logo, headline, description, landing page, and yes, a creative asset designed to grab attention.
For now, the moment when GPT Ads can compete with Google Ads or Meta Ads is still quite far off, and it is still missing common features such as lookalike audiences or audiences similar to your own, A/B testing, or, for that matter, the ability to launch video campaigns. As for pricing, right now ChatGPT Ads supports two payment models: CPC (cost per click), in which the advertiser pays when someone clicks on the ad and bidding is set through a maximum CPC; and CPM (cost per thousand impressions), in which the advertiser pays for every thousand times the ad is shown, even if it does not receive clicks. Let us remember that OpenAI began its advertising pilot using CPM exclusively, but in May 2026 it added CPC bidding.
In any case, I think what really matters is that OpenAI is already bringing together, piece by piece, the components a modern advertising platform needs: measurement, attribution, optimization, audiences, automation, and tools for eCommerce.
What additional features do you think OpenAI should include in this platform?
Image: ChatGPT
Your email address will not be published. Required fields are marked *
Δ