loading...
Millionways positions its AI as a self-understanding tool, but the real bet is B2B: selling organizational culture analytics to companies that want to align employee motivation with business goals.
ENTRY ANGLES
Specialized AI companion for specific mental health segments (students, couples, women, children, faith communities) · Vertical-specific AI model fine-tuned on emotional language and trust patterns of defined audience · AI companion as entry point to broader mental health service ecosystem for same audience
VERTICALS
CAPABILITIES
AI model fine-tuning for emotional language and context-specific responses, Mental health domain expertise for target segment, Go-to-market strategy for healthcare/wellness vertical
"We often don't understand what to do next or what to aim for," says a character on Millionways' homepage – and that framing is deliberate. The founders spent years researching exactly this problem before building a product around it.
The core argument runs like this: most people lack someone who genuinely listens to them. Without that, they don't understand themselves well. And without self-knowledge, they default to chasing other people's goals rather than their own.
Millionways' answer is what the company calls "the world's first AI with emotional intelligence." Emotional intelligence, in the technical sense, covers the ability to recognize emotions, understand motivations and desires – both one's own and others' – and manage emotional dynamics to achieve practical outcomes.
Building the AI took four years. The team conducted 7,000 conversations totaling over 500,000 minutes, analyzed the data across five million parameters, and trained the resulting model under the supervision of 25 professional psychologists.
The platform works through conversation. The AI listens, asks the right follow-up questions, and builds a psychological model of the person it's speaking with. Millionways claims this model can be constructed from as few as 500 words of text or audio – even from multiple separate conversations held at different times.
An additional mode is the "smart journal": a user shares what happened during their day and how it made them feel, and the AI identifies patterns and emotions that the user may not have articulated or consciously recognized.
As an optional layer, the platform can connect users with strangers who have similar psychological profiles – people working through comparable goals and desires – for human-to-human peer support.
Individual pricing runs from €4.99 to €24.99 per month. The entry tier is AI-only. The €14.99 plan includes access to a nine-month structured group program. The top tier adds an individual mentor and direct access to the support team.
Millionways recently announced a $10M funding round. The company operated as a nonprofit research organization until recently and has now converted to a commercial entity – a trajectory with a notable precedent in OpenAI, which followed the same path from nonprofit to commercial startup.
Millionways' commercial bet is on B2B, not consumer subscriptions – and the individual AI companion is, by the founders' own account, primarily a data collection and model training vehicle.
The corporate product centers on organizational culture: using the AI to surface employee motivation and align it with company objectives (a retention play), to assess candidate motivation during hiring (a screening play), and to analyze team dynamics to reduce motivational conflict. A continuous monitoring capability tracks the psychological state of the workforce to inform internal communications and goal-setting.
At the B2B pricing level Millionways is targeting – roughly $10,000 per month per company plus $5 per individual analysis – the revenue model is fundamentally different from the consumer tier. The company is also licensing its AI for third-party integration, with at least one announced 50/50 joint venture with BCII.
The structural opportunity here mirrors what happened to ride-hailing economics: the unit economics of services that rely on credentialed human providers are constrained by the cost and availability of those providers. AI that can perform a meaningful portion of the psychological support function at a fraction of the cost per session has the potential to transform the economics of the mental health market, just as autonomous vehicles – if realized – would transform ride-hailing margins.
The mental health market is substantial and growing. Valued at $383M in 2020, it is projected to reach $538M by 2030. The addressable market for AI-powered support tools within that segment is still largely uncontested.
The most interesting structural observation is the segmentation already emerging across mental health startups. There are purpose-built platforms for [students](/review/a-na-urovne-jemocij-reshaetsja-i-prodajotsja), for [couples](/review/zdorovye-lichnye-otnoshenija), for [women](/review/i-rynok-bolshoj-i-sposob-horoshij), for [children](/review/rebjonok-chast-semi), and even for [specific faith communities](/review/bozhestvennoe-spokojstvie). Each segment speaks a different emotional language, responds to different framings, and has different trust thresholds for AI interaction.
That segmentation pressure applies directly to AI companions: a general-purpose model has to navigate too many different contexts to be deeply effective for any one of them. A model fine-tuned on the emotional landscape of, say, first-generation college students or recently divorced parents is likely to outperform a generalist on those specific use cases – and creates a clearer go-to-market story.
The most viable entry path may be to build a specialized AI companion for a well-defined audience first, then use that vertical as a foundation for a broader mental health service purpose-built for that same group. The companies that chase maximum generality from the start will spend more time building a universal product, compete with everyone else pursuing universality, and be less effective for any specific population than a specialized entrant. Vertical focus is the under-chosen option in this market.