As third-party tracking disappeared, Checkmate built a profitable AI marketing business on its own shopper data. 700+ brands were already using it.

Checkmate, the consumer shopping platform backed by Google Ventures, Mantis VC (The Chainsmokers) and Paris Hilton, today revealed the business it has been quietly building for years: mate, an AI marketing platform now used by more than 700 brands including Everlane, Billabong, Brooklinen, Malbon and JD Sports. The business is profitable, on track to exit the year at roughly a $15 million run rate, and was built with zero marketing spend, according to company financials.

The reveal closes the gap between how Checkmate was understood publicly and what it was actually building. When it raised $15 million in 2023, it looked like a consumer story: a shopping app that hit #1 on the U.S. App Store by getting brands to approve real promo codes and cash back for shoppers. The business on the other side of those transactions went unannounced.

The timing is not an accident. The years Checkmate spent building on the consumer side are the same years the industry’s borrowed data disappeared. Apple’s tracking prompt cut off mobile signal. Safari and Firefox blocked third-party cookies by default. Google spent years promising a replacement inside Chrome and then scrapped the plan, leaving marketers with the signal loss and none of the new plumbing. Media costs kept climbing as more advertisers chased less signal.

Over the same stretch, AI made the software layer trivial. Anyone can generate a campaign now. It also started moving discovery itself, with AI traffic to U.S. retail sites up 393% year over year in Q1 2026, per Adobe Analytics. What looked like a consumer detour in 2023 turns out to be the one thing that got harder to buy: consented first-party data and a direct line to the shopper.

The consumer app was never just an app. Every offer redeemed, every price tracked, every purchase across app, email, SMS and desktop was teaching Checkmate how 100 million-plus shoppers actually decide. That is more than 12 billion live intent signals, and they are what mate’s agents run on. The shopping side earned the audience. The brand side quietly monetized the understanding. Marketing spend on the B2B business over that period: zero.

“We watched a generation of AI marketing tools launch as demos in search of data and distribution,” said Harry Dixon, co-founder and CEO. “We did it in the opposite order. We spent years earning the shoppers, the signals and the brand relationships, and only now are we putting a name on it. mate launched today, but it has been compounding for a long time.”

Everlane attributed $480,000 to mate in under 30 days across 2,450 orders, with a 30% lift in conversion. The roster added Quiksilver, Dickies and Eddie Bauer this month, and has begun expanding beyond ecommerce into travel, with marketplace RVshare, and into financial products.

“Before mate, we were considering turning off Meta ads altogether,” said Adam McAreavy, Director of eCommerce at PSA Skincare, which has seen a 150% lift in return on ad spend on the platform. “It surfaced shoppers we could never reach before.”

The team has stayed unusually lean for the scale: Dixon, an Australian who previously worked with architect Frank Gehry, and co-founder Rory Garton-Smith, a former iPhone engineering program manager at Apple.

“The industry keeps asking what happens to startups when AI makes software easy to copy,” said Garton-Smith. “Our answer is that the software was never the moat. The moat is a hundred million shoppers who show up for offers, and seven hundred brands whose campaigns get smarter because of each other. That took years of unglamorous work, and it is why the agents work.”

Checkmate keeps its name and its shoppers, and now operates as the distribution layer mate converts through. The company is applying the same technology to the shopper side of the network.