AI Candidate Data Enrichment Agent

How Cassidy automates using ai

Step 1: Trigger on candidate event

The Workflow activates when a new candidate is added, a resume is uploaded, a CSV import completes, or a stage change occurs in your ATS—ensuring validation runs at the right moment without recruiter action.

Step 2: Parse and normalize candidate data

Cassidy extracts entities from resumes and profiles—name, contact info, work history, skills—then normalizes fields like phone numbers to E.164 format, standardizes job titles to your taxonomy, and maps skills to a controlled ontology.

Step 3: Validate contact information

The Workflow runs email validation (syntax, MX lookup, catch-all detection, disposable filtering) and phone validation (line type, carrier lookup, SMS deliverability) to assign status codes, confidence scores, and last-verified timestamps.

Step 4: Resolve identity and deduplicate

Cassidy matches candidates using deterministic keys (email, phone, LinkedIn URL) and probabilistic signals to create a unified candidate ID, flagging duplicates for merge while preserving field-level provenance and recruiter-entered data.

Step 5: Enrich and write back to ATS

Verified and enriched data—including contact status, compliance flags, and best-contact recommendations—writes back to your ATS with field-level update policies, while low-confidence matches route to an exception queue for coordinator review via Slack or email.