Eight stages, from brief to QA-cleared shortlist. Our agents find the executives who might be a good fit for your client. Deep research qualifies each one against your spec. Then every claim is challenged before you see it.
Role, must-have criteria, geography, disqualifiers. Every must-have becomes a rubric item Cerna evaluates each executive against.
→ Criteria too vague to check against public evidence — "visionary leader," "culture fit" — get flagged back before the search begins.
A wide first pass — typically 150–200 executives — designed for recall. Miss no one who belongs in the pool. Deep research runs only on the executives who survive this filter.
→ Surface-level filters: function, geography, approximate experience band.
Surface screening against must-haves. Clear disqualifications removed — wrong function, wrong geography, insufficient tenure — before spending on deep enrichment.
→ Output: ~50 executives worth deep research.
Each survivor is enriched with role history, dates, scope detail, and everything needed to evaluate the must-have criteria. Every claim recorded with its source.
→ Entity resolution happens here. Mis-attributed evidence is disqualifying for the specific claim.
Each executive evaluated against each must-have individually. Every criterion gets a state: confirmed, unconfirmed, or flagged with the specific reason.
→ Confirmed requires the source to directly support the claim — no reasonable-inference-from-a-title-alone.
Qualify pass: survivors resolve into Section A (on-spec) or Section B (laterals). Recall pass: a deliberate second look at near-misses — who belongs in B that a keyword filter would miss?
→ The recall pass runs every time. It's the pass that produces the laterals.
A separate adversarial pass reviews every claim before delivery. Clean context, no access to prior stages' assumptions. Looking explicitly for what the generation pipeline got wrong.
→ A shortlist that fails QA is not delivered. Recruiter receives only what passed.
Shared Slack Connect channel — one per client firm. Self-contained document: every confirmed claim has its source. No back-and-forth needed.
→ The recruiter can send the doc to a partner without re-qualifying names from scratch.
Section A + Section B is not a hedge. It's how a real research operation delivers. A keyword filter produces Section A. Only understanding what the role requires produces Section B.
Executives who satisfy every must-have criterion. Each entry: confirmed role history with sources, per-criterion rubric scores, source citations, contact info, research depth signals where relevant.
Executives who miss a criterion but share the operational substance. Each entry includes Section-A's fields plus the specific lateral reasoning: which criterion is missed, which are satisfied, why they're worth attention.
If a search returns no on-spec matches, Section A says so explicitly and states why. Cerna does not pad the list to reach a number.
Cerna does not rank executives by predicted fit. It does not recommend who to hire. It does not contact executives directly. It does not tell the recruiter who is right. The shortlist is an input to the recruiter's judgment, not a substitute for it.
The first search is on us. Same pipeline. Same sourcing discipline. Same QA bar as a paid one.
No credit card. $1,000 per search after that.
