Judgment before
the search.

Why a contact-driven headhunter fails on AI roles

• No standard sample: on a new track there are not enough comparable incumbents to benchmark against.
• The JD describes tools, not problems: it lists frameworks, while the real question is which part of the business and organisation has to be solved.
• The best candidates are not in the active market: the more critical the person, the more passive they are, so résumé databases and inbound applications reach very few of them.
• Judgment depends on one person's experience: results are capped by whoever picks up the role, and cannot be reproduced.

The method: judgment before search

Four stages, each with a concrete output rather than a promise to find someone.

• 01 Business diagnostic: turn the JD back into a business problem and define what is worth hiring for. Output: the role standard, covering business stage, team structure, technical and product judgment, and organisational position.
• 02 Expert calibration: the business expert committee calibrates the dimensions in the language of peers, so screening is not driven by big-company pedigree alone. Output: judgment dimensions and weighting.
• 03 Confidential search: an AI talent map locates candidates across target companies and adjacent pools, with double-blind conversations and reasoning shared before résumés. Output: a shortlist with reasoning, first batch within 48 hours.
• 04 Onboarding support: offer negotiation, expectation management and landing follow-up, testing the judgment against results. Output: placement feedback.

The six judgment dimensions

The six dimensions describe a consultant's judgment style: insight (seeing through the business problem and the fit), empathy (reading motivation and what someone really wants), communication (making complex judgment clear), drive (owning pace and decisions), collaboration (working in step with the client and the committee) and poise (stability under pressure and confidentiality). Every consultant has their own peak, and clients see why this person fits this kind of role, not just a score.

Which roles fit this method

• 0-to-1 team building: several critical roles have to land at once, with no existing team to benchmark against.
• Business lead: the standard spans technology, product and commercialisation, so hiring experience alone is not enough.
• Confidential replacement: the incumbent cannot be exposed, so outreach must control pace and information boundaries.
• New business line: a new direction has no historical role samples, so talent has to migrate from adjacent pools.

How one project runs (anonymised)

• Week 1: a business diagnostic with the founder and CTO turns the hiring need into four weighted judgment criteria.
• Week 2: the expert committee calibrates, removing big-company bias and adding adjacent-industry samples.
• Weeks 3-4: the AI talent map locates three candidate profiles, double-blind conversations start, and the first shortlist arrives with reasoning.
• Month 2: critical interviews and offer negotiation, with expectation management in parallel.
• Month 3: onboarding and landing follow-up, testing the judgment against results.

Three anonymised cases

• Embodied AI 0-to-1 team build: a well-funded company filled 10+ critical roles across three functions within one month.
• Confidential search for an AI business lead at a Hangzhou top-tier startup: 8 months, 1,853 approaches, one placement, zero identity leaks.
• New business line for a listed company: 20+ critical offers in two months, with the core team in place.

How to evaluate a headhunter: five questions

If you are choosing a partner, these five questions separate the field quickly.

• What is your judgment standard? Answers that only mention resources or networks usually mean judgment has not been productised.
• How fast is the first shortlist, and does it include reasoning? A résumé-only shortlist hands the judgment back to the client.
• Who does the judging? Is the consultant's background comparable to the role?
• How is confidentiality guaranteed? When does a candidate learn who the client is?
• How is the judgment tested afterwards? Is there landing follow-up, or does the case close at offer?

Three common misunderstandings

• AI will replace headhunters: AI widens the search radius and speeds up validation, but the final judgment remains with consultants and the expert committee.
• A headhunter is a résumé channel: critical roles lack reasoning, not volume.
• Faster is always better: a 48-hour first shortlist exists to start judgment earlier, not to rush a candidate's decision.

FAQ