Known's AI interviews each user first, then matches on genuine compatibility signals – not curated self-presentation. $9.7M says the market agrees.
ENTRY ANGLES
AI interviewer extracting candid information through conversation for better matching · Style-based matching for creative professionals (designers, video editors) · Interest-based social connection (e.g., meme preference matching)
VERTICALS
CAPABILITIES
Conversational AI / natural language understanding, Matching/recommendation algorithms for fuzzy matching, Information extraction from unstructured user data
KNOWN FOUNDER
“the way dating should have worked from the start”
Known claims to have built a dating app "the way dating should have worked from the start" Profiles tell you almost nothing about a person. Swiping them left or right is about as meaningful as liking a post on Instagram.
The only thing that can actually change the equation, according to Known, is AI – an AI that can help you meet people you'll genuinely connect with.
At the core of Known is an AI that each new user talks to before anything else – so it can understand who this person actually is. The AI isn't shy about asking unexpected questions: for example, "if I asked your ex, what would they say it was like to date you?"
Interestingly, people turned out to be quite willing to open up this way. Average conversation length is currently 26 minutes, and some users spend upward of 90 minutes.
Only once the AI has a clear picture of someone does it start showing that person's profile to others – specifically to people whose profiles it also understands from similar conversations. Matching happens based on what the AI learned, not on what the profiles claim.
When a user sees someone else's profile, they can ask the AI – by voice or in chat – for more context: who is this person, what are they like, why were they suggested? Ideally, by the time a user says "I want to meet them" or "I'll pass," they already know enough to skip the awkward getting-to-know-you phase of a first date.
The mutual match rate is striking: Known users express mutual interest in 80% of introductions – far above the typical swipe-based app. AI-powered matching based on conversation depth actually works better.
When both sides say yes, the AI immediately suggests meeting options – specific venues and times drawn from both users' schedules, locations, and stated preferences.
The app also builds in a sense of urgency: users have 24 hours to respond to a suggested profile, and another 24 hours to schedule the date if interest is mutual. Miss the window and that person disappears from the feed.
After the scheduled date, the AI follows up with both users – checking whether the date happened and gathering impressions. This feedback feeds back into the matching algorithms, continuously refining them.
Known is currently running as a pilot in San Francisco while testing monetization models. One source reported a fee of $30 per scheduled date charged to each participant – which works out attractively for the company.
Despite its early stage, Known raised $9.7M last month.
Going on dates based purely on profiles is like conducting job interviews based purely on resumes. Most people have learned the hard way that both approaches are remarkably inefficient.
Pre-screening through text chat doesn't help much either – most questions get standard or surface-level answers. People only truly open up in conversation, and those conversations require skill to conduct.
It also helps when the conversation is run by a neutral third party – someone the other person doesn't feel pressure to impress or filter themselves for. This is why matchmakers have existed forever in dating, and why recruiters exist in hiring.
AI can now fill that third-party role. It can conduct sophisticated conversations, listen actively, and draw meaningful inferences. As a result, AI interviewers are finding their way into recruiting, social matching, professional networking, and a range of other applications.
Boardy ([related review](/review/produkt-kotoryj-sam-prinosit-investorov)) raised $11M for a professional networking app connecting entrepreneurs, founders, and investors – using the exact same playbook as Known. Boardy's AI holds voice conversations with new users, learns everything about them, then proposes introductions to others it has already spoken with. Matches are only made with mutual interest.
The same framework applied to recruiting can involve two AI agents – one focused on candidates, one on employers. Jack & Jill ([related review](/review/a-ty-jetu-ofigennuju-vozmozhnost-mozhesh-razgljadet)) operates this way and raised $20M in its first round this past October. Laborup raised $5.8M in August; Mercor has raised $483.6M total, including $350M this past August on a $10B valuation.
Your360.ai ([related review](/review/vtykaj-ii-mezhdu-ljudmi)) launched a workplace feedback platform in October using a similar principle: the AI conducts conversations with a person's colleagues, then synthesizes those opinions into an anonymized summary for the employee – along with a suggested development plan for strengthening their strengths and addressing their weaknesses.
There are also platforms that conduct user research conversations – gathering product feedback and improvement ideas at scale. Outset ([related review](/review/kak-poluchit-insajty-pro-svoj-produkt)) raised $30M a couple of weeks after its original review; Perspective raised $4M earlier this year.
The common thread across all of these startups isn't just the voice interface. It's that AI plays the role of an interviewer – drawing people into candid conversations and surfacing what lies beneath the surface, to use in their own interest.
Translated into product terms: through conversation, you can extract far more information from a person – including details they haven't consciously articulated – and use it to make far better matches: with other people, with employers, even with products on a marketplace. That is, in any context where today's experience relies on limited sets of filters and predefined search parameters.
That reframe suddenly expands the design space considerably. Because it encompasses platforms without voice interfaces at all – platforms solving analogous "fuzzy matching" problems.
Roster ([related review](/review/a-takih-marketplejsov-poka-net)) is one example: a platform for finding designers, video editors, and other creative professionals who work in the same style as you. Users upload samples of their work; seekers upload an example of what they're looking for; the platform finds freelancers who not only have the right skills but also work in the specified style.
And since this review started with a dating app, it's worth noting Schmooze ([related review](/review/o-chjom-s-toboj-trahatsja)), which raised $7.5M on the idea of connecting people based on the memes they find funny.
The founder's argument: for Gen Z, memes are a language – a direct window into someone's inner world. Matching on memes goes deeper than it might seem. The AI doesn't just find people who liked the same meme – it analyzes the full body of memes a person gravitates toward, infers their personality, and then suggests connections whose personalities appear compatible, even if their meme preferences are completely different.
The broad direction: building AI platforms for fuzzy matching, grounded in richer initial information gathering and smarter inference algorithms.
The number of possible applications is enormous – and all of them carry novelty. Because the underlying methods – AI-powered information collection and cross-referencing – are genuinely new.
So: in which domain could your version of this platform operate, and how?