Automatic document relationships
Droqify links related documents after processing, such as bills to the claim they belong to, so a case reads as one connected file.
Documents in
- Claim records
- Medical bills
- Explanation of review documents
- Work status reports
- Independent medical examination reports
How it runs
-
Document processed
A document finishes the normal pipeline of text extraction, classification, schema matching and field extraction.
-
Linking job queued
After successful processing, the platform queues a document relationship job for that document.
-
Candidate links found
A handler for the document type compares extracted field values, such as provider and service date, and records listed in reports, with other documents in the same claim.
-
Confidence decides the link
High-confidence links are created automatically, medium-confidence links are created and marked as lower confidence, and weak matches are discarded.
-
Linked view
Related documents can then be read together, and people can also create or correct links manually via the API.
What you get
- Related documents connected without manual filing
- Links labelled with how and how confidently they were inferred
- Weak matches kept out rather than guessed
- Manual links available where the rules do not apply
How it works today
The relationship builder runs after processing and is rule-based, using extracted field values and the document type. The built-in handlers currently target claim-style document sets, such as workers’ compensation bills, review documents and examination reports. Other document sets need their own schemas and handlers.
Where review fits
Low-confidence links are marked as such so a person can check them. Manual links can be created through the API.