Title automation · Doma Technology
ParseIQ
One parsing call for every title document, with a confidence score on every field.
My contribution
I built ParseIQ as the single parsing service behind Doma's title automation.
30title document types behind one endpoint
The problem
Title automation touches dozens of document types. A separate endpoint for each one multiplies integrations, and automation needs to know which extracted values it can trust before it acts on them.
How it works
- Put PUDIQ, DeedIQ, and 28 other document types behind one parsing endpoint.
- Return a confidence score for every field it extracts.
- Let downstream automation use those scores to decide when a person needs to review.
What changed
Development is complete and ParseIQ is deploying to production. Downstream systems make one call instead of hitting a separate endpoint for each document type.