💼 Toyota, NTT, Tokyu, Resona, Itoki. That list reads like a page from a stock index. It is also a list of Japanese employers that will help pay for their single staff to find a partner, through an app whose AI picks who you meet, coaches your small talk, and shows you a number for how much the other person seems to like you. Roughly 1,600 companies have signed up. The question worth asking is why so many Japanese firms decided their employees' love lives were a business problem.

The dating app your employer has to unlock

The service is called Aill goen, built by a Tokyo startup called Aill, founded in 2016 and selling the app to employers since 2021. ITmedia Business Online, which reported on the company on 27 August 2026, put adoption at more than 1,600 companies. Nikkei counted over 1,400 in July 2025. Aill's own homepage still advertises about 1,200, so the exact count depends on where you look.

Signing up starts the way any dating app does, with photos, age, where you live. Then it gets specific in a way most apps never attempt. You declare whether your job might move you across the country, whether you work remotely or in an office, what you want your career to look like, where you stand on marriage and children, and, as a percentage, how much of the housework and childcare you are prepared to carry.

An AI reads all of that and introduces one person per day. Never anyone from your own company. The pool is other companies' employees, which is the whole trick: the app pairs someone facing nationwide transfers with someone who can work from anywhere, or a woman who wants roughly 70 percent work and 30 percent home life with a man who wants the reverse.

The user base clusters at the ages when Japanese people actually marry. Forty percent are in their twenties, another 40 percent in their thirties, 20 percent are forty or older. Asked how they want to work after marriage, 43 percent of male members wanted a dual-income household and 45 percent had no strong preference. Among women, 61 percent wanted dual income and 32 percent had no preference.

Once two people match, the AI keeps working. If the woman tells it she would like to get a meal together soon, it will suggest to the man that this might be a good moment to ask her out. It also puts a visible score on how warmly the other person seems to be responding, which is meant to take some of the guesswork out of making a move. Aill's founder and chief executive, Toyoshima, told ITmedia the design replaces the job a mutual friend used to do, and that an app is harder than real-world dating precisely because no mutual acquaintance exists.

The company's own promo video, in Japanese only, shows how it is pitched at home.

The privacy architecture is what lets this work as a benefit. Under the model ITmedia describes, the employer pays a monthly fee based on headcount and cannot see which of its people have signed up. Nobody is matched inside their own company. Users can exclude specific firms, a major client for instance, and choose whether group companies are in or out. Employment is re-verified every six months with a company ID or a payslip.

The design has two limits. Introductions are opposite-sex only, which the company states plainly on its own site. And there is no published flat price: what an individual pays, if anything, depends on which employer they work for.

The retention math underneath the romance

Ask why a car manufacturer is paying for this and the answer, as ITmedia lays it out, is unromantic.

Japanese companies have spent a decade building parental leave, shorter hours, flexible schedules. Those policies help an employee who already has a supportive partner. They do very little for an employee whose partner assumes she will be the one to cut back. Someone can want the management track badly and still end up scaling down or resigning because the domestic load landed entirely on one side of the household.

That is the gap. A company can redesign every policy it owns and still have no standing to renegotiate a marriage. So it moves upstream, to the moment before the marriage exists, and tries to raise the odds that its employee ends up with someone whose plans are compatible with staying employed. Fewer resignations for domestic reasons means less hiring and less retraining. Recruiters have also found the benefit lands well with candidates, because it answers a specific anxiety: can I still have a family if this job transfers me, can I keep working after I marry.

Put plainly, the company has concluded that who you marry is partly a retention issue.

Japanese companies used to do this without an app

The function is old. The software is the new part.

For a stretch of the postwar decades, the Japanese workplace was the country's dominant matchmaking institution. In the 1970s, roughly one marriage in three began at work. Bosses made introductions. Company clubs, company trips and company dorms did the rest.

A 2005 paper in The Japanese Journal of Labour Studies, by two researchers at the National Institute of Population and Social Security Research, took apart the decline in the first-marriage rate over the previous three decades to see what caused it. About half of the fall was the disappearance of arranged marriage, a category that included introductions by relatives and by superiors at work. Nearly another 40 percent was the decline of workplace marriage. Every other route into a marriage, meeting at school, through friends, out in the city, had barely moved in forty years. The authors described the corporate society of the 1960s and 70s as a matching-maker whose function was never handed on to anyone else as it shrank, and judged that reviving it was almost certainly impossible.

Two decades later, companies are trying to buy the function back as a monthly subscription.

The trend line has not reversed in the meantime. The 16th National Fertility Survey, conducted in June 2021, is the most recent published round; a 17th was carried out in 2026 but the results are not out yet. Within that survey, couples who married between July 2015 and June 2018 had met at work in 28.2 percent of cases. For couples who married between July 2018 and June 2021 the figure was 21.4 percent. Online routes, meaning social media and matching apps, moved the other way across the same two windows, from 6.0 percent to 13.6 percent. One caveat the institute itself flags: the later window covers the pandemic, when marriages fell sharply, so a rising share does not automatically mean a rising number.

The most recent official picture is starker. The Children and Families Agency surveyed 20,000 people aged 15 to 39 in 2024 and asked those who had married in the previous five years how they met their spouse. Matching apps came first at 25.1 percent, ahead of work, business contacts and part-time jobs at 20.5 percent, and school at 9.9 percent. The app overtook the office.

Why "benefit" is the right word here

To an outside reader, a company subsidising your dating life sounds like an overstep. Inside the Japanese system it reads as an extension of something old.

Japanese employers have subsidised private life for generations, and the accounting shows it. In Keidanren's survey of 608 large employers for fiscal 2019, published in November 2020 and so a pre-pandemic snapshot, discretionary benefits (everything the company chooses to spend on top of statutory social insurance) averaged ¥24,125 per employee per month, around $150 at the August 2026 rate of ¥159.4 to the dollar. Nearly half of that, ¥11,639 or about $73, went to housing: company flats, subsidised dorms, rent support. Meals, health programmes and congratulatory payments made up much of the rest.

If your employer already houses you, feeds you lunch and sends money when you get married, arranging for you to meet someone is not a leap in kind. It is one more line item.

Government is on the same road. Tokyo opened applications for its own AI matching service, TOKYO Enmusubi, on 20 September 2024. Joining requires an official certificate of single status, documentation of your income and an online interview. As of 15 August 2026 the city reported about 39,000 applications, roughly 16,000 people who completed registration, 816 committed relationships and 302 marriages. Tokyo has budgeted ¥205.4 million, about $1.29 million, for its marriage matching programmes in fiscal 2026, up from ¥127.3 million, about $800,000, the year before.

What Japanese readers are actually arguing about

The reaction at home was not a shrug. The ITmedia piece drew around 165 comments on Yahoo News within a day, and the sharpest ones were not about romance at all.

The most-upvoted comment argued that the clearest beneficiary is the vendor, not the employer: every large company that joins adds vetted members and credibility for everyone else's staff, so the product gets more valuable with each signing, which is not how an ordinary benefit works.

A second thread was about why this became necessary. Approaching a colleague has become risky as harassment norms tightened, and remote work removed a lot of the casual contact that used to lead somewhere. Against that, a pre-vetted pool of strangers from comparable employers looks less like corporate overreach and more like a patch for something that broke.

Then there was the fairness objection, and it has teeth. Aill only sells to employers it has screened: large firms, listed companies and their group subsidiaries, licensed professional bodies, and companies certified under government schemes for childcare support and women's advancement. Everyone else is outside the wall. Readers pointed out that dispatch workers, contract staff and people at small companies are structurally shut out, and one argued that a marriage market sorted by employer status edges uncomfortably close to a class divide. Another put it more bluntly: a pool stocked entirely with blue-chip employees will mostly produce couples who were already going to be fine.

The skeptics were pointed. Several readers said courtship is starting to resemble a job interview, with a values spec sheet where attraction used to be. One raised a genuine consent question about the mechanics: if one user confides something to the AI and the AI then shapes what it says to the other user, whose conversation is it. Others wanted outcome data that nobody publishes, not match rates but divorce rates, five years out.

The challenge is fair, because almost every published figure comes from the company. Aill's own site says AI support lifts the rate at which chats turn into an agreed date from 28 percent to 76 percent, and that 81 percent of its male members are willing to take at least half the housework and childcare. Both are self-reported, and the second measures a population that opted into an app built around shared-workload matching. Read them as marketing until someone independent checks.

It took fifty years for the Japanese office to stop being the place where people met their spouse. The remedy has arrived as a line in the HR budget. Where you are, does your employer touch your personal life at all, and if HR announced a company dating app tomorrow, would your colleagues sign up or file a complaint?

参照