Claims intake and triage
Turn incoming claim packets into structured claim records, with uncertain fields routed to a reviewer before anything is released.
Documents in
- Claim forms
- Medical records
- Medical bills
- Wage statements
- Policy documents
- Correspondence
How it runs
-
Intake
Claim packets are uploaded in the portal or sent through the API, and each file becomes a tracked processing job.
-
Text extraction and OCR
Text is read from PDFs and Word files, and pages with little text are sent to OCR so scanned pages are covered.
-
Classification and schema matching
Each document, or page range inside a combined packet, is classified and matched to an extraction schema. Documents with no match are flagged for review.
-
Field extraction
Fields defined by the matched schema are extracted with a confidence value for each field.
-
Human review
Low-confidence fields and unmatched documents pause in the review queue. A reviewer corrects them and approves, and processing resumes.
-
Summary and output
A document summary is generated, the record becomes searchable, and results are delivered via API and webhooks.
What you get
- Structured claim record with per-field confidence
- Mixed claim packets separated into individual documents by type
- Uncertain items held for a named reviewer instead of passing silently
- Searchable claim documents with a short summary for the adjuster
Where review fits
Claims teams decide how much review they want. Classification, schema matching and field extraction can each pause for a person when confidence is low or no schema matches. Reviewer corrections are carried into later extractions for the same schema.
Integration points
Documents enter by upload or API, and completed results are available through the API and the document completed webhook. Droqify runs as SaaS, single-tenant in the customer’s Azure, or managed. Industry packs provide ready schemas for workers’ compensation (US), health insurance and general insurance.