Candidate Search AI connects directly to your ATS (such as Bullhorn) and transforms your existing candidate data into a smart, fully searchable talent network.
What makes it unique is not only what data we use — but how the AI understands that data in context.
Below is a breakdown of which ATS fields we use and why this makes your search far more accurate than traditional ATS search.
📄 Resume / CV
What we use
The complete resume stored in the ATS
All job descriptions
Skills mentioned anywhere in the document
Responsibilities, tools, achievements
Certifications and education
Soft skills and role summaries
Why this is unique
Traditional ATS search looks only for exact words.
AI Search reads the resume like a human recruiter:
Understands meaning
Interprets seniority
Understands related concepts
This allows the system to find matches even when
candidates describe their experience differently.
✍️ ATS Notes
What we use
Recruiter notes added in the ATS
Interview summaries
Behavioral notes
Availability updates
Qualification comments
Internal remarks about skills or fit
Why this is unique
Notes often contain the most accurate and up-to-date information.
Example: “Candidate recently completed AWS certification.”Traditional ATS search usually does not index notes deeply.
AI Search reads notes, understands their meaning, and connects them to:
Skills
Experience
Seniority
Candidate strengths
This makes your internal tribal knowledge searchable for the first time.
🛠 Skills & Experience
What we use
Structured skills from the ATS
Skills extracted from resumes through parsing
Years of experience
Seniority indicators
Skill variations and equivalents
Why this is unique
AI expands skills with synonyms and related terms
(e.g., React ↔ ReactJS, RN ↔ Registered Nurse).Mislabelled or inconsistently formatted skills still get matched.
Experience levels are interpreted contextually (e.g., “5+ years” or “Mid-level”).
📍 Location & Relocation Preferences
What we use
Candidate location
Region/city/state
Willingness to relocate
Acceptable locations
Remote preferences
Why this is unique
AI handles radius logic, relocation logic, and multi-location matches.
It understands variations (“Berlin”, “Berlin Area”, “Berlin, Germany”).
Traditional search ignores relocation almost entirely — AI doesn’t.
🧑💼 Job Titles
What we use
Current job title
Past job titles
Industry-specific titles
Seniority tags (Senior, Lead, Junior, etc.)
Why this is unique
AI understands similarities:
“Customer Support Rep” ≈ “Call Center Agent”
“Software Engineer” ≈ “Backend Developer”
“Charge Nurse” ≈ “Registered Nurse”
This makes title-based matching significantly more intelligent.
📊 Work Experience (Years, Timeline, Continuity)
What we use
Total years of experience
Role duration
Skill-specific experience
Gaps (if relevant)
Why this is unique
AI reads for experience details everywhere on the resume.
It can infer seniority even if the candidate doesn’t state it directly.
🏷 Custom Fields (Industry-Specific ATS Data)
What we use
Licenses
Certifications
Work eligibility
Clearance levels
Available start date
Compliance fields
Healthcare credentials
Shift preferences
Why this is unique
AI Search indexes custom ATS data structurally and contextually.
These fields influence ranking and matching.
Traditional search engines often ignore custom fields or treat them as plain text.
🧠 Parsing — How We Understand Candidate Data
Parsing is the process of converting messy or unstructured resume text into structured, searchable information.
What our parsing does
Identifies skills (even hidden in long text)
Detects experience levels
Extracts job titles and matches them to known roles
Identifies certifications and licenses
Understands tech stacks, tools, and responsibilities
Normalizes skills (e.g., “JS”, “JavaScript”, and “Javascript” all become the same skill)
Why parsing is essential
Resumes are inconsistent — everyone writes experience differently
Without parsing, searches miss 40–60% of relevant results
Parsing creates clean, structured data for AI Search
Combining parsing + semantic search gives unmatched accuracy
It improves ranking: strong matches go to the top automatically
Bad or incomplete ATS data becomes usable and searchable
Parsing is what makes “semantic search” possible — it’s the foundation.