A.I. is reshaping online dating, from recommending matches on Tinder to helping users break the ice on Hinge. Love-seekers now navigate platforms where algorithms increasingly guide their decisions. Done well, these tools can improve the experience and lead to better matches; done poorly, they risk drowning out authentic human communication.
At Match Group—the parent company of Tinder, Hinge, Match.com and several other dating apps—CEO Spencer Rascoff and CFO Steven Bailey are accelerating internal A.I. adoption. The company has given all 2,300 employees access to tools like Claude and hosted an “A.I. day,” where employees trained on the technology and built their own tools as part of a contest.
“We want to be an A.I.-native company as quickly as possible,” Bailey told Observer, adding that these initiatives have doubled Match Group’s annual A.I. budget to $10 million this year from last year’s $5 million.
Bailey, a longtime Match Group executive, has held leadership roles across the company since 2012, including senior vice president of financial planning and business operations and CFO of Match Group Americas.
Across the tech industry, companies are rapidly increasing A.I. spending. Executives are pushing engineers to fully utilize these tools, fueling a phenomenon known as “tokenmaxxing.” Uber, for example, reportedly burned through its A.I. budget in just four months, while Nvidia CEO Jensen Huang has said he wouldn’t be surprised if a senior engineer spends the equivalent of half their salary on tokens.
At Match, early A.I. use came without strict limits. Now, the company is entering what Bailey calls its “ROI phase,” focusing on efficiency. “We don’t want to constrain A.I. use, but make sure it’s being used in an efficient and effective way,” he said. “I’ve got to find a way to pay for all this.”
To manage costs, Match has introduced “speed bumps,” or token limits, across the organization. It has set one threshold for technical employees and a lower one for non-technical staff. Employees who hit the limit must request additional tokens from their managers.
Bailey said 100 percent of engineers now use A.I. to write code, with most code generated by A.I. and reviewed by humans. Engineers at Tinder are producing and shipping roughly twice as much code as they were a few quarters ago. The average engineer spends about $600 per month on A.I. tools, compared with roughly $50 for non-engineers, Bailey said. Top engineers, however, can spend as much as $3,000 monthly.
The spending comes as Match pushes A.I. to improve user experience and reduce “swipe fatigue.” Two-thirds of Tinder’s improvements this year have focused on its algorithms, while Hinge’s A.I.-driven matching system, rolled out a few quarters ago, has increased matches by 15 percent.
Hinge has also introduced “prompt feedback,” a feature that helps users improve responses to profile questions like “what’s your favorite beach?” or “What’s your favorite vacation?”
Bailey said men, in particular, often struggle to write prompt answers that are engaging or distinctive. The tool doesn’t generate responses outright but encourages users to expand and refine their answers.
Still, the shift to becoming A.I.-native comes with tradeoffs. To offset rising A.I. costs, Match has slowed hiring. Bailey said the move is intended both to fund A.I. investments and to reassess staffing needs as the company evolves.
“The roles we’re going to need are going to be different once we become A.I.-native than what we needed even just six months ago,” he said.
However, going A.I-native isn’t without its challenges. In the company’s effort to pay for its transformation, Match’s leadership has decided to slow hiring. Bailey adds the reason is to pay for the increased cost of A.I. tools and to get a better understanding of headcount needs as Match becomes an A.I.-enabled organization.