The Fastest-Growing AI Category Right Now Has 10-Person Teams

01 /The Concept

Juicebox built an AI platform that helps companies "win the war for talent" – meaning find and hire the people who have the specific skills they actually need.

The core of the platform is a candidate search system that "understands what kind of specialist a company is looking for." The key differentiator: it doesn't match on keywords that appear in a profile or resume – it understands the meaning of the search request. And it handles multi-criterion queries naturally.

Example: "Find senior engineers in the San Francisco Bay Area who have experience building infrastructure for search engines." A standard keyword-based search won't surface engineers who listed only specific platforms and tools – but it will surface people who have the phrase "search engine infrastructure" in their resume, some of whom turn out to be consultants or "systems architects" who haven't written code in years.

Juicebox has assembled a database of 800 million people across professions, drawing from more than 30 sources that cover different dimensions of each person's professional life – enabling it to surface a meaningful and accurate candidate list for even complex queries.

Once candidates are identified, outreach emails can be sent directly from the platform. The AI drafts individualized messages for each candidate, factoring in their background, interests, and current employer. According to Juicebox, these AI-personalized messages generate 3x the response rate of standard recruiter outreach.

Candidate profiles are continuously updated – with historical data preserved alongside current information. That lets HR teams and recruiters track changes in the talent landscape over time: supply shifts, salary movements, and other measurable signals – making hiring strategy and forecasting more informed.

The baseline workflow: a recruiter submits a search, reviews results, refines the query, reviews again, requests email drafts, reviews those, sends them, waits for replies, responds, and continues until interviews are scheduled.

Recently, Juicebox rolled out AI recruiters capable of handling that entire workflow autonomously – surfacing to human recruiters only the candidates who are already warm, fully meeting the defined criteria. Each AI recruiter works one role at a time, but does so 24/7 with no breaks for sleep, meals, or anything else.

A free tier is available with search and result volume limits. Paid plans for unlimited searching cost $119 or $179 per recruiter per month. The lower tier fits individual recruiters and founders; the higher tier is better suited to startups and recruiting agencies. Enterprise pricing is available on request.

AI recruiters are an add-on: $300/month for two concurrent AI agents, billed annually.

Juicebox went through Y Combinator in summer 2022 and launched its current platform – originally called PeopleGPT – in late 2023. The startup then went relatively quiet before re-emerging recently with an announcement: $30M in new funding and a disclosure that it had raised $6M in an earlier undisclosed round.

02 /Why It Matters

Juicebox kept building during its quiet period – with a small team. With four people, it reached $1M in annualized revenue. After expanding to twelve – likely on the $6M it hadn't publicly announced – it reached $10M ARR and 2,500 platform users.

The growth is driven by genuine customer value. One startup reports building its entire 20-person team through the platform. A large enterprise says that using Juicebox freed its recruiters from search grunt work – with 80% of their time now spent on meaningful candidate engagement rather than sourcing.

Perhaps the most notable part: despite that step-function growth from $1M to $10M ARR, Juicebox still has no dedicated sales function. Its growth has been entirely word-of-mouth; the team only handles inbound.

That word-of-mouth has apparently reached the startup world. This past summer, a startup called Clado ([related review](/review/kak-najti-idealnogo-sotrudnika)) – founded this year – went through Y Combinator building a very similar platform.

Clado's example searches are equally impressive: "find investment bankers from Goldman Sachs or JP Morgan who moved into venture capital in the last three years and are now at top-tier funds," or "find engineers who worked on Tesla Autopilot, are now in autonomous systems roles elsewhere, and have published academic papers on computer vision algorithms."

The AI hiring platform space has other strong players, many also showing excellent growth with lean teams – and also without dedicated sales.

Mercor ([related review](/review/a-ty-jetu-ofigennuju-vozmozhnost-mozhesh-razgljadet)) began growing at 51% month-over-month last fall with a team of around 30. By February, it was at $75M ARR and had raised $100M at a $2B valuation. It is now on track for $450M ARR and is reportedly in discussions for a new round at a $10B valuation.

Laborup ([related review](/review/starye-rabotnye-sajty-pora-vykidyvat-na-pomojku-istorii)), which built an AI platform for hiring industrial workers, piloted in a small US city. Within the first three months, one in five working-age residents in that city had registered on the platform – allowing local manufacturers to hire quickly through it. Laborup raised $5.8M in August to expand.

03 /Opportunities

The broad direction: AI hiring platforms. It's a massive, perennial market, and AI has introduced genuine new capabilities – making this one of the most durable areas for builders.

And there's enormous room beyond what Juicebox, Clado, Mercor, and Laborup are already doing.

RefAssured ([related review](/review/na-tom-chto-ljudi-vrut)), which raised $3.3M recently, built a platform whose AI solicits and cross-references candidate references. The clever angle: meaningful insights can emerge not just from consistency across references, but from inconsistency.

Roster ([related review](/review/a-takih-marketplejsov-poka-net)) launched late last year with an AI platform for creative industry hiring – designers, video scriptwriters, content editors. The platform doesn't just match on skills – it matches on working style. Candidates' portfolios are analyzed; recruiters provide examples of the aesthetic or stylistic direction they want, and the AI surfaces candidates who work in a similar vein.

The defining characteristic of today's strongest AI hiring platforms seems to be the ability to assemble diverse and unexpected data sets – enabling candidate discovery along diverse and unexpected criteria.

Roster's example shows that niche-specific platforms have a real advantage: by collecting data specific to a narrow domain, they can deliver results no general-purpose platform can match within that space.

So – in what domain would you build your AI hiring platform? What datasets would need to be assembled and analyzed to find the most qualified candidates in that space? Where do those datasets come from, and how would you analyze and rank against them?

There's still significant open territory here. And most importantly – that open territory maps directly onto a concrete opportunity space.