ChatGPT is putting competing brands in the same answer: Here’s how often and where

When OpenAI announced in June that it would test multi-advertiser placements, grouping several relevant ads into a single unit sold through a second-price auction, the industry reaction was mostly speculative. Then screenshots of two ads in one ChatGPT answer started circulating in late July, and the question became practical: is this a rare glitch or a real format?

It’s real. Adthena’s LLM advertising dataset covers close to 3 million ChatGPT responses in the 15 days to July 29, and the pattern is unambiguous. Before mid-July, responses carrying more than one advertiser were consistently under 0.5% of ad-bearing answers. Around July 16 to 18 that changed sharply, peaking at 3 to 4% of ad-bearing responses between July 18 and 21, then settling to a steady baseline of roughly 1.6 to 2.4% in the final week of the month.

Two percent sounds small. Two points matter alongside it. First, 98% of ad-bearing ChatGPT answers still carry a single brand, so this is an emerging format, not the norm. Second, a multi-avertiser format that sat at effectively zero a month ago now shows up thousands of times a day across five markets.

Inside the multi-brand slice, two advertisers is the standard configuration at 97.6% of multi-brand responses. Three brands account for 2.3%, and four-brand answers exist but are rare enough to count by hand.

Every multi-advertiser response in our dataset comes from five markets: the United States, the United Kingdom, Australia, New Zealand, and Canada. Canada has the highest rate, with 2.57% of its ad-bearing responses carrying multiple brands, a striking result for the smallest market in the set. The US and UK dominate on volume, together accounting for more than 70% of all multi-brand responses we observed.

Financial services brands are the most frequently paired category in every live market: Barclays and Barclaycard in the UK, Rocket Mortgage and Capital One in the US, NAB and Vanguard in Australia, BMO and Intuit QuickBooks in Canada. Telcos are close behind in the UK, where giffgaff, Vodafone, and Sky all sit in the top six. Booking.com is the most consistent cross-market player, appearing in the top ten of four of the five live markets.

Comparison sites and aggregators show up alongside the brands they compare: BestMoney leads the US table, and Finder AU leads Australia’s. If your category includes an aggregator with an ad budget, expect to share answers with it.

A multi-advertiser answer is a competitive surface, not just an ad slot. Your brand can now appear directly beside a rival inside the one answer a buyer asked for, on the prompts where they’re actively choosing. Share of voice inside a single AI response is about to become a real metric, and right now almost no brand knows whether it’s on that shelf, who it’s sharing it with, or which prompts put it there.

That’s a visibility gap ChatGPT’s own reporting doesn’t close. Ads Manager tells you what you served. It doesn’t tell you who you served next to, or how often your competitors are being stacked against you on the prompts that matter in your category. 

Adthena’s AI Ads Intelligence tracks sponsored placements across ChatGPT and emerging LLM ad platforms, where available, so you can see the multi-brand answers in your market, the rivals appearing in them, and the prompt patterns that trigger them.

The format is at 2% today. The advertisers in the charts above are finding out what it looks like at 20% before their competitors do. If you want to know who’s sharing your shelf, talk to us.

See your market’s AI ad landscape.