How to hire people who aren't looking for a job

The best hire for a role is usually someone who isn't looking. LinkedIn's talent research puts roughly a quarter of professionals actively job-hunting and three-quarters passive, and most of that passive group is open to the right approach. They won't see your job post. They will take a look when someone they trust points them your way.

That's why referrals beat cold sourcing. Harvard Business Review finds referred candidates are higher quality, more likely to receive and accept an offer, and more likely to stay. But there's a catch worth building your whole strategy around: not all referrals are equal. A name passed along by someone who only knows the candidate through social media carries little of the trust that makes a referral work. A real relationship carries the signal; a thin connection doesn't.

Start by building a shortlist of people you'd love to hire, rather than posting a job and hoping. A few places to build it:

• LinkedIn Recruiter is the default for breadth: search the professional graph by skill, title, and company.
• Gem, hireEZ, SeekOut. Sourcing platforms that find, rank, and re-surface candidates, including specialized and technical roles.
• Your ATS (Greenhouse, Lever, Ashby), where inbound applicants and employee referrals flow through one pipeline.

These tell you who exists. The harder, higher-value question is who can get you a warm introduction to them.

A list of dream candidates is only half the job. The other half is the warmest way in: who on your team, or in your investors' networks, actually knows them well enough to vouch.

This is exactly where raw LinkedIn connections fail and relationship intelligence wins. Recall the catch. A thin, social-media-only tie is worthless as a referral, and what lands instead is the real relationship: a former colleague, someone a teammate actually worked with.

The Swarm and Village read your team's real relationships, former colleagues included, to surface those paths. With Village you describe the person you want ("Staff frontend engineer with React, in the Bay Area") and it ranks candidates by fit, maps the warmest path to each, scores how strong the relationship is, shows who should make the intro, and drafts it for you to approve. The trusted nudge that passive talent actually answers becomes a search you can run.

You can increasingly run sourcing and pipeline work from one AI assistant via the Model Context Protocol (MCP), Anthropic's open standard for connecting an assistant to your tools and data.

What's real today: several ATS platforms are exposing MCP servers, some official and some community-built, so an assistant can read your pipeline and candidates (Greenhouse, Ashby, and Lever are among those with servers; confirm official vs community on each vendor's docs). Pair that with Anthropic's reference servers for web fetching and you can research candidates and companies in one place.

Two honest caveats. LinkedIn has no official MCP server, and the community projects that exist aren't affiliated with LinkedIn and can run against its terms. The relationship-intelligence tools don't expose official servers yet either. So the connectable surface today is your ATS plus the open web, with the warm-path layer still to come.

Two asks matter here. The external one, getting introduced to a candidate, uses the same mechanics as any warm intro: the double opt-in, where your connector checks the candidate is open before connecting you, and a short forwardable note written to the candidate. Discretion matters more in hiring, because a passive candidate is doing their current employer no favors by being seen to take the meeting. That makes the double opt-in protective as well as polite. The fundraising guide covers the email anatomy in full.

The internal ask is the one teams skip. Employee referral programs are powerful but underused. Instead of waiting for submissions, ask your team specific, quality-seeking questions: "who's the best engineer you've ever worked with?" That surfaces the strong referrals that actually pay off, the ones HBR says programs need deliberate design to find.

Network-led hiring compounds. Every strong person you hire brings their own trusted former colleagues into your team's reachable network, so each good hire widens the warm path to the next one. Alumni and boomerang networks keep growing as careers progress, and a great team becomes a magnet for great people.

One qualifier carries all the weight: the flywheel only spins on real relationships, and amassed connections don't count. Knowing who actually knows whom, and how well, is what turns a growing team into a growing talent pipeline.

The loop: hire someone great → their trusted former colleagues come into reach → you make the next warm hire → who brings theirs. Each turn, your team's reach into the talent market widens.

Hiring through your network is one compounding system: find the people you want, find who can credibly vouch, ask in a way that protects a passive candidate, and let every great hire widen your reach.

The step teams do blind is finding who can vouch, because a recruiter sees their own connections and not the whole team's. Village maps those relationships, ranks the warmest path to any candidate, drafts the intro for you to approve, and keeps it current as people move. The idea is bigger than any tool: the person you want to hire is already one trusted introduction away, and your team just couldn't see who held it. (Raising instead? See fundraising through your network. Selling? See landing customers through your network.)