Hello Group's AI Coaching Fails to Halt User Exodus: A Trust Crisis Unfolds

Hello Group's bet on AI-powered conversation coaching was meant to solve its retention crisis. Instead, the Chinese dating conglomerate behind Momo and Tantan watched paid users collapse by 15% to 7.3 million in Q4 2024, even as it rolled out tools designed to help male users craft better opening lines and sustain chat momentum. The algorithmic assistance couldn't stop the bleeding.

The product logic appeared sound. Men struggle to write messages that get responses, so AI analyses successful conversations, suggests replies, and smooths out awkward silences. Management positioned the feature as engagement infrastructure across both platforms, with usage metrics described as 'encouraging' but crucially not quantified.

The business results tell a grimmer story. Revenue fell 2.6% to $186.6M, continuing a decline that stretches back four years from the company's 2020 peak of $2.4B.

Create a free account

Unlock unlimited access and get the weekly briefing delivered to your inbox.

The home market problem isn't new

Hello Group's decline stretches back to 2020, well before this AI gambit. Management attributed Q4 results to 'soft consumer sentiment' in China, the same explanation offered in previous quarters. That framing conveniently sidesteps harder questions about whether the product itself has lost relevance.

China's dating app market has faced structural headwinds for four years. Regulatory crackdowns tightened content moderation requirements and increased operating costs. Demographic shifts — a shrinking youth population and worsening gender ratios — have reduced the addressable market.

Competition from short-form video apps like Douyin (China's TikTok) has siphoned attention and created alternative spaces for singles to meet without the transactional friction of a dedicated dating app. Hello Group's response has been to look elsewhere, with Middle East revenue surging 73% year-on-year to represent 17% of total revenue.

That's a sensible geographic hedge, but it doesn't solve the China problem — which still accounts for the majority of the business.

Gender asymmetry and the coaching economy

The decision to target male users specifically for AI conversation assistance reflects both user behaviour patterns and monetisation strategy. Male users on Momo and Tantan significantly outnumber female users and convert to paid subscriptions at higher rates. They're also more likely to pay for features that promise a competitive edge.

AI conversation coaching fits neatly into that willingness to pay for advantage. The tools reportedly analyse message history, suggest contextually relevant replies, and offer real-time feedback on tone and engagement likelihood. It's a productised version of what dating coaches have sold for decades, now embedded directly into the app interface.

But the gender asymmetry creates a disclosure problem. Female users receiving these AI-assisted messages have no way to know whether they're chatting with the person or the algorithm. That's not a minor UX concern — it's a trust issue that cuts directly into the value proposition of online dating.

If the initial connection phase is mediated by AI, what does that mean for the authenticity of the relationship that follows?

Other platforms have grappled with this differently. Match Group has experimented with AI icebreakers on Tinder, but positioned them as mutual tools available to all users. Bumble has explored AI for photo verification and scam detection, but kept conversation assistance at arm's length, citing brand concerns around authenticity. Hello Group's approach — coaching one side without disclosure to the other — is more aggressive, and the market response has been unambiguous.

What the failure signals

If AI conversation tools were going to stabilise Hello Group's user base, Q4 was the quarter to show it. The features launched early enough to generate data, uptake was reportedly strong, and yet paid users continued to churn at double-digit rates.

That suggests the engagement problem isn't solvable with better chat prompts. Users are leaving not because they can't think of what to say, but because the platforms themselves no longer feel worth the effort. Perhaps the match quality has degraded as the user base has aged and skewed male.

Perhaps the rise of alternative social discovery platforms has made dedicated dating apps feel transactional by comparison. Perhaps the stigma around online dating has returned in certain cohorts, particularly in markets like China where social pressures around marriage and family formation remain acute.

COO Yue Li said on the earnings call that management expects AI tools to 'stabilise the user base and improve engagement metrics over the coming quarters'. That's aspiration, not evidence. The company has been predicting stabilisation for three years, and each quarter the guidance gets pushed forward.

The AI experiment at Hello Group matters because it's one of the clearest tests yet of whether algorithmic assistance can fix a structural engagement problem. The early read is no. Teaching users to perform better within a system they're increasingly disengaged from doesn't address why they're disengaged in the first place.

Operators watching this should note the implications. AI conversation tools may drive short-term engagement spikes — users trying the shiny new feature — but if they don't translate into retention, they're feature theatre. Worse, if they erode trust by introducing hidden asymmetries into the matching process, they risk accelerating the authenticity crisis that's already driving platform fatigue across the industry.

Hello Group's numbers suggest that risk is real, and the payoff isn't.