OpenAI Just Launched Its Own Ads Manager. The $50K Entry Point Is the Smallest Part of the Story.

By Floodlight Editorial Team

As of April 10, OpenAI’s ChatGPT ad pilot is officially self-serve. That means advertisers can now track impressions and clicks in real time without routing through OpenAI’s sales team. Additionally, the minimum spend dropped from somewhere between $200,000 and $250,000 down to $50,000, making it much more accessible for smaller organizations. This price cut will get attention from anyone watching the channel, but for a performance marketer or DTC brand weighing whether to actually test this, the number that matters more is 1.5.

TL;DR: OpenAI’s new self-serve ads manager lowers the barrier to entry for ChatGPT advertising, but the pilot is still limited and access isn’t guaranteed. The more important question is whether the channel’s conversion data justifies getting in now, and for the right product categories, it absolutely does.

What the Ads Manager Actually Changes

Prior to April 10, buying ads on ChatGPT meant going through OpenAI’s sales team directly or working with a certified partner. Digiday reported that the new self-serve tool gives advertisers real-time visibility into impressions and clicks without that intermediary step. Advertiser sources have described the interface as being similar to Google Ads in terms of layout. That’s not an accident. It’s no surprise that familiar tools get adopted faster, and OpenAI needs adoption to build the revenue case ahead of its expected IPO later this year.

Currently, ads appear at the bottom of ChatGPT responses for free and Go tier users in the U.S., labeled as sponsored and visually separated from the organic answer. OpenAI has stated that ads do not influence what ChatGPT says and that conversation data is not shared with advertisers. The ad inventory itself has been live since February 2026, but as mentioned before, the minimum spend and barrier to entry were very high. What changed on April 10 is that measuring and managing those placements got significantly more accessible.

Why LLM-Referred Traffic Converts 50% Better

Criteo joined the ChatGPT ad pilot in March 2026, bringing roughly 17,000 commerce advertisers with it. Digiday named Criteo as OpenAI’s first certified ad tech partner. The aggregated U.S. client data Criteo has published shows that users arriving from an LLM referral convert at approximately 1.5 times the rate of users from other referral channels.

Take a typical Google search. Someone types “running shoes” or “dinner ideas” and gets back a page of sponsored listings, Google Shopping carousels, Business Profiles, and ten blue links all competing for the same click. The user will desperately tabs, compare pages, and try to concoct a trustworthy answer from sources that all have some reason to sell them something. It’s a process that asks the user to form their own opinion, and plenty of people give up or second-guess themselves before they ever buy.

On ChatGPT, instead of typing “running shoes,” the user types something along the lines of: “I have wide feet, my knees have been hurting on longer runs, and I mostly train on pavement. What running shoes should I look at?” ChatGPT doesn’t return ten links. It synthesizes what it knows about stability shoes, drop height, and wide-toe-box options and gives a direct answer built around that person’s specific situation. The user didn’t have to do the comparison work. ChatGPT did it for them. Users understand the power of ChatGPT and know that it can crawl data from across the internet to give them the “perfect” answer. It makes complete sense why LLM referral converts at ~1.5x the rate of other channels.

A 1.5x conversion lift at the referral stage changes the unit economics of a test campaign. Media cost per click is still uncertain and will likely vary by category and competition. But if a channel consistently delivers traffic that closes at that rate, it warrants a budget allocation even before the cost structure is fully understood.

What $50,000 Actually Gets You

While the spend minimum has come down substantially, brands still need to reach OpenAI’s sales team or a certified partner to participate. Meeting the spend threshold is a prerequisite. It does not guarantee access.

The most direct path in right now is through Criteo, which brought its existing commerce network into the pilot. For brands already running performance campaigns through Criteo’s platform, that’s worth a conversation with their account team.

This is also early enough that the Criteo conversion data, while directionally promising, reflects a thin sample. The 1.5x figure comes from aggregate behavior across a limited pilot. Category performance will vary. Some product types will outperform that benchmark, while others won’t get close.

The IPO Context

OpenAI’s IPO is expected later this year. The company is projecting $2.5B in ad revenue for 2026, with projections to reach past $100 billion by 2030. Those may prove to be lofty goals since ad industry forecasts at the early stage of a new channel have a consistent history of revision. The 2030 number in particular should be read as aspiration, not guarantee.

Regardless of whether those numbers hold, OpenAI is clearly building ads into its core business. That means more inventory, better tools, and a lot more competition over time. Early testers will have a head start when access opens up. The ones who wait will be learning the hard way while paying higher rates to do it.

How to Prepare Before the ChatGPT Ads Pilot Opens Further

Match Categories to the Format

ChatGPT ads are better suited to high-consideration purchases. Products that people research before buying, where the decision involves some complexity, and where a detailed ChatGPT answer might include a recommendation or a comparison. 

Categories like home improvement, skincare, supplements, and apparel where fit actually matters are a natural match. People research these before spending money, and ChatGPT is increasingly where that research happens. Impulse purchases don’t work as well here. If someone buys it without thinking twice, they’re probably not asking an AI about it first.

Rethink the Creative Brief

The ChatGPT interface isn’t a feed where people are browsing. Someone asked a question, got an answer, and your ad is the next thing they see. This is a completely different experience than an Instagram ad interrupting someone’s scroll, and the creative should reflect that.

Audit the Attribution Stack

Before spending anything, make sure ChatGPT traffic shows up as its own source in your analytics. If it gets lumped in with everything else under “referral,” you won’t be able to tell what’s working. The whole point of an early test is learning something. That only happens if the data is clean enough to read.

The brands that win here won’t be the ones who threw the most money at it first. They’ll be the ones who came in knowing what they were testing, picked the right products, and set up their tracking before they spent a dollar. None of that is complicated. It just has to happen before the channel gets crowded, and there’s still time to do it.