AI Underwriting Core Integration Agent

How Cassidy automates Underwriting Core Integration using AI

Step 1: Trigger on new submission

The Workflow activates when a submission arrives—whether from a broker email inbox, portal upload, ACORD XML feed, or API call—and classifies the document types automatically.

Step 2: Extract and normalize submission data

Cassidy uses AI to extract key fields from ACORD forms, loss runs, SOVs, and supplemental questionnaires, then normalizes addresses, class codes, and coverage details to match your underwriting schema.

Step 3: Validate and check completeness

The Workflow runs domain validations—required fields by line of business, limit/deductible ranges, territory eligibility, and compliance checks—capturing any gaps or exceptions with explicit rationales.

Step 4: Enrich with third-party data

Cassidy pulls firmographics, property characteristics, CAT risk scores, and other enrichment data to complete the risk picture and calculate appetite fit scores.

Step 5: Triage and route exceptions

Submissions are scored and triaged against appetite and authority rules. In-appetite risks flow straight through; exceptions trigger Human-in-the-Loop referrals with a complete dossier for senior underwriters.

Step 6: Push to underwriting core

Cassidy creates or updates submission records in your PAS or underwriting workbench, attaches normalized documents, calls rating engines for indications, and writes back statuses—all via API with full audit trails.