How Do You Hire Someone Who Isn't Looking?

01 /The Concept

Finding the right people – whether for a hire, a partnership, or an advisory role – is genuinely hard. Job boards are a poor answer: the right candidates rarely post there. They get headhunted from their current roles, or they respond to a posting that catches their attention.

But to headhunt someone, you first have to find them. Most professional platforms like LinkedIn offer blunt search tools. That forces recruiters to hunt across multiple sources, then manually stitch the picture together.

Clado offers an alternative: a people search engine that accepts natural language queries.

The platform supports simple searches by description, an "in-depth" mode with more complete candidate profiles, and a specialized alumni search by institution.

The queries can get sophisticated. For example: "Find me an AI engineer interested in biology currently based in Copenhagen who I can recruit for my pharmaceutical startup" or "Give me a list of all senior executives at US and UK accounting firms with 250 to 2,000 employees."

When a query comes in, Clado deploys 100,000 AI agents that fan out across data sources, cross-reference information, and rank candidates against the query parameters. The company describes it as having a recruiting firm with 100,000 staff on retainer.

Of course, finding someone is useless without a way to reach them. Clado also locates contact information – and claims strong performance here. On a test set of 100 random LinkedIn profiles, Clado found contact details for 17.6% more people than Clay, and three times more than Apollo.

Pricing starts at $25/month for unlimited searches. At $500/month, users get contact export and access to the in-depth search beta.

Clado is currently in Y Combinator. The platform launched on the YC website only yesterday.

02 /Why It Matters

A similar people-finding problem – in a narrower context – is being tackled by fellow Y Combinator graduate Verata ([related review](/review/chtoby-pobedit-konkurentov-nuzhno-znat-chto-u-nih-proishodit)). Verata focuses specifically on sourcing executive talent for private equity portfolio companies – what it calls "marrying LinkedIn and Pitchbook."

The key insight: Verata's AI doesn't just pull résumé data from LinkedIn. It evaluates how revenue and funding evolved at a candidate's company while they were at the helm. PE firms can screen candidates on actual financial impact rather than job titles – betting that past performance in that specific dimension is predictive.

HelloSky ([covered here](/review/vot-kak-teper-budut-iskat-pravilnyh-sotrudnikov)) raised $5.5 million last month for a broader executive search platform serving companies, PE funds, and recruiting firms. Its AI aggregates candidate and company data across sources, then produces detailed profiles that include an estimate of how likely the person is to be poachable.

Another YC graduate, Spott ([covered previously](/review/odnimi-rukami-95-000-000-ne-zarabotaesh)), built a similar AI recruiting tool for staffing agencies and closed $3.2 million two weeks after that review published.

All of this connects to a broader idea worth sitting with.

1. Prompt engineering isn't wordsmithing – it's management. You have to clarify what you want, provide the context needed to get it, and iterate until the output is usable.

2. But that requires knowing at least enough about the subject matter to frame the question and evaluate the answer. Without that, you can't catch hallucinations or assess whether the output is actually useful.

3. And "enough" turns out to be sufficient. Which opens up remarkable possibilities.

4. The old model said you had to become a deep expert in something. But experts are people who know more and more about less and less. Deep expertise narrows your aperture – which limits your ability to work at intersections.

5. Intersections are exactly where the boldest breakthroughs tend to happen. That's why "T-shaped" people – deep in one area, aware of many others – became so valued.

6. Now comes the "comb-shaped" person. No single area of deep expertise, but genuine surface-level familiarity with many domains. Enough to ask the right questions and evaluate the answers.

7. These people will drive breakthroughs precisely because they have the widest aperture. AI fills in the depth when needed.

8. This changes what skills matter – and it also changes what kind of people companies will need to hire.

Returning to Clado: if the "comb-shaped" model of talent becomes more valuable, hiring will require finding people with specific combinations of surface-level fluency across multiple domains. That kind of search – multi-parameter, qualitative, cross-source – is exactly what tools like Clado are built for.

Mark Andreessen recently made a similar point: the best future entrepreneurs will have competence across 6–8 domains. As AI handles depth, breadth becomes the differentiator. The ability to cross-apply patterns from different fields – to combine unlike things in new ways – is increasingly the skill that matters.

This shifts the hiring question from "how deep is this person?" to "how wide are they, and can they synthesize?"

03 /Opportunities

The broad direction: build platforms for intelligent people discovery – employees, freelancers, co-founders, advisors, partners.

The solution space is wider than the examples above suggest. A few more cases:

Roster ([related review](/review/a-takih-marketplejsov-poka-net)) helps companies find designers, videographers, and other creatives who work in a specific visual style. Show the platform a reference piece, and its AI analyzes the work uploaded by creatives in the system and surfaces matching candidates. The natural extension: scan portfolios across Behance, personal sites, and the open web, not just the platform's own database.

Boardy ([covered here](/review/produkt-kotoryj-sam-prinosit-investorov)) has raised $11 million for an AI that finds co-founders, investors, and collaborators through actual phone conversations – you tell the AI who you are and who you're looking for, and it matches you with someone it has already spoken with.

Series ([related review](/review/dlja-professionalnogo-netvorkinga-nuzhny-neobychnye-socseti)) raised $3.1 million for a text-based equivalent that started with student co-founders and is expanding to freelancers, mentors, and employers.

What all of these share: multi-source data, multi-parameter matching, and the ability to search on criteria that don't fit neatly into dropdown menus – things that emerge from natural descriptions or conversations.

Starting narrow and domain-specific – the way Verata and Roster did – may be smarter than building a universal people search from day one. A focused use case means a focused customer segment, deeper search capabilities for parameters that matter to that segment, and a faster path to meaningful business results.

So: whose people-finding problem are you solving? Start with the criteria current tools can't express – qualitative, multi-dimensional, contextual – and build from there.