Paste a person or a company — a LinkedIn URL, an X handle, or a website. We work out which it is, then find 3–4 with the same DNA: for people, the trajectory and rare skill combinations rather than the job title; for companies, the wedge, the buyer and the technical bet rather than the industry label. Plus the research being published on what they do.
Career arc
Defining traits
Couldn't be dated
The same pipeline a human researcher runs — turned into an algorithm.
Reconstruct the full profile from the open web — education, every role, the company's actual product.
Distil 5–8 defining dimensions — career arc, domain pivots, rare skill combos — each weighted by how distinctive it is.
Turn the trait vector into multiple neural search queries. You can't find a lookalike with one search — you search across trait combinations.
Exa neural search surfaces people by career meaning, not keywords, then results are merged and de-duplicated.
Each candidate is scored on every trait axis (0–1), weighted, and ranked. The top 3–4 come back with a note on what matches — and what doesn't.