Which brands are testing ads in ChatGPT?

Which brands are testing ads in ChatGPT?

Beyond understanding how ChatGPT ads work and how the platform decides which ones to show, now that the pilot is live, a more concrete question has come up: Which brands are already inside ChatGPT, and what can their early moves tell us?

This article from Floodlight, a programmatic ad solutions provider, sticks to what has been publicly confirmed, sourced from company announcements and verified reporting.

What OpenAI Has Confirmed About ChatGPT Ads

The pilot launched on February 9, 2026, having been delayed three days from its original start date. OpenAI confirmed that ads are visible to logged-in adult users on the free tier and the $8-per-month Go plan, while Plus, Pro, Business, Enterprise, and Education subscribers will not see them.

OpenAI has stated that ads do not influence ChatGPT’s answers, that conversations remain private from advertisers, and that placements are labeled and separated from the platform’s main response. Advertisers receive aggregate performance data (such as views and clicks) but not personal user information. There is no open self-serve platform. Access requires a direct relationship with OpenAI or participation through an agency group. OpenAI has confirmed it wants advertisers to commit at least $200,000 for the early ad test. The program is currently controlled, curated, and invitation-only.

Confirmed Brands in the ChatGPT Ad Pilot

Target

Target is among the first wave of brands to begin testing ads within ChatGPT. The company is promoting both its own business and select partners through its Roundel retail media network, with placements aligned to specific keywords in user prompts.

Target has previously noted that traffic to its site from ChatGPT is growing roughly 40% month over month. That trajectory made advertising a logical extension of a relationship already underway. The retailer launched a dedicated Target experience inside ChatGPT in November 2025 and has been building toward deeper AI commerce integration since.

In a theoretical example, a user asking for recommendations on countertop cooking appliances could be shown an ad for an air fryer brand that advertises through Roundel. The ad would appear separately from the answer, not embedded within it.

Williams-Sonoma

Williams-Sonoma, the parent company of Pottery Barn, announced its participation in the OpenAI Ad Pilot on February 11, making it among the first home goods companies to test ads on the platform.

The partnership is focused on reaching customers at decision-making moments and improving product discovery. Williams-Sonoma CEO Laura Alber described the effort as an opportunity to develop a new advertising approach, one designed to engage consumers contextually and in a manner aligned with how users expect to interact with information on the platform.

Albertsons

The grocery chain Albertsons also confirmed participation in the OpenAI advertising pilot in February 2026, according to MediaPost. Testing is expected to be contextual and tied to the kinds of questions shoppers ask most (meal planning, seasonal recipes, household staples). The category alignment is natural: grocery is among the most intent-rich and routine-driven verticals in consumer spending.

The Knot

The Knot Worldwide is among the early participants in the pilot. The company operates a wedding-planning marketplace connecting couples with more than 200,000 vendors, from venues and photographers to bridal boutiques and makeup artists.

Jenny Lewis, CMO of The Knot Worldwide, told Glossy that roughly 36% of couples are now using some form of AI in their wedding planning, which is nearly double the figure from the prior year. That shift in behavior is exactly what drove the decision to test advertising within ChatGPT: If couples are searching for wedding vendors inside AI chat tools instead of traditional search engines, the platform becomes a critical visibility channel.

The Knot also launched a dedicated ChatGPT app in early February, giving users access to its vendor marketplace directly within the chat interface, which surfaces imagery, reviews, and local options inside the conversation.

Agency-Driven Participation

Beyond the individually confirmed brands, several major agency groups have moved quickly to place clients inside the pilot.

The OpenAI Ad Pilot has drawn investment from Omnicom Media, WPP, and Dentsu. Omnicom Media confirmed that more than 30 of its clients have secured placements in the pilot. WPP Media has separately confirmed participation from clients including Adobe, Audible, Ford, and Mazda. Dentsu has also joined the effort.

These decisions reflect agency-level bets on a new channel with structured commitments made before performance data exists, based on strategic positioning rather than proven returns.

What the Early Participant List Reveals

Retail and commerce dominate the first wave

Every confirmed direct-to-brand participant (Target, Williams-Sonoma, Albertsons) operates in product-driven categories where purchase intent is closely tied to the kinds of questions people bring to ChatGPT. Someone asking about kitchen appliances, grocery staples, or home furnishings is already in a discovery mindset. These are not passive browsing moments. Instead, they are active, goal-directed conversations.

The categories span a wide intent range

The confirmed participants cover grocery, home goods, wedding planning, software, entertainment, and automotive. That said, OpenAI’s ad system appears to be drawing interest not from a single vertical but from any category where high-intent conversational queries are common.

Large, brand-safe advertisers are first

The $200,000 minimum commitment and invitation-only access structure ensures that the pilot skews toward established players with sizable budgets and legal teams capable of evaluating a new, largely undefined platform.

What Remains Unknown

Pricing structure is not publicly standardized. The $200,000 minimum commitment is known, and a $60 CPM figure has been reported by AdExchanger and Digiday, but how pricing ultimately scales (by context, category, query volume, or competitive bidding) has not been disclosed.

Measurement tools are underdeveloped. Target’s SVP of Roundel noted that the company will review aggregated performance metrics (impressions, clicks, and early engagement signals) but acknowledged that capabilities like first-party data integration, closed-loop measurement, and flexible buying models are not yet available. The infrastructure that makes other platforms reliable for performance measurement does not yet exist here.

Auction mechanics are not public. Whether ads are placed through a real-time auction, a fixed-rate placement, or some hybrid model has not been explained. This matters for brands trying to understand how competitive the environment will become as more advertisers enter.

ChatGPT Ads Are No Longer a Hypothetical

Real companies are spending real money to appear inside real conversations. The brands that have moved first are not doing so because the returns are proven; they are doing so because the format is new enough that early participation carries informational value that later entry cannot replicate.

As one Roundel executive put it, “The goal of the pilot is to take time to thoughtfully test and learn, while introducing new experiences in a measured way that prioritizes consumer relevance and trust.” That measured language reflects the stakes on both sides: OpenAI needs these pilots to succeed quietly, and the brands involved need the learnings more than the impressions.

For people who use ChatGPT regularly, the practical implication is that the brands already in the test have made a bet that conversational AI is where consumer attention is heading, and that showing up early, even imperfectly, is better than waiting for the rules to be written without them.

This story was produced by Floodlight and reviewed and distributed by Stacker.