Candidate Search AI is the AI search layer on top of your ATS. It searches the candidates you already have in Bullhorn and ranks them by how well they fit the role you describe.
You describe the role in plain language, or open a job in Bullhorn and let Candidately read the job description. The AI turns that into a search, runs it across your whole database, and shows you a ranked list with a score, a short summary and a list of evaluations for every candidate.
What you can do with it
Search in natural language. "Senior Java backend engineer, 6+ years, Spring and microservices, Berlin or within 50 km." No keywords, no Boolean.
Match a Bullhorn job. Open the job in Bullhorn, open the Candidate Search tab. The search is built from the job description automatically.
Refine step by step. Type follow-ups into the search bar ("add Kubernetes", "remove the location"). Each message adjusts the same search instead of starting over.
Filter on hard facts. Experience range, required or excluded skills, location and radius, and any Bullhorn field (status, owner, custom fields).
Understand every result. Each candidate has a match score, an AI summary and evaluations that show which requirements they meet and which they don't, with the evidence from the resume on hover.
Act directly from results. Open the full profile, open the record in Bullhorn, submit to a job, add to a tearsheet, list or shortlist, or hide a candidate, one at a time or in bulk.
Where you use it
In Candidately: Candidates page, AI Search tab.
In Bullhorn: on a job record, the Candidate Search tab (or Actions → Candidately Search and Match).
What it searches
Everything Bullhorn has on the candidate: parsed resume, skills, job history, titles, education, notes and custom fields. Data is kept current by live sync, so new and updated candidates are searchable within minutes. See Which ATS data Candidate Search AI uses and Initial sync and live sync explained.