Intake starts with messy packets
Submissions arrive as emails, PDFs, applications, ACORD forms, loss runs, schedules, supplemental forms, and free-form attachments. The receiving system needs normalized data before underwriting can move efficiently.
An intake API should not force a new workflow. It should convert the packet into structured data that the customer's existing underwriting, broker, or MGA system can use.
Reuse the same rails
The intake extension reuses PolDex primitives: classification, schema contracts, evidence, abstention, webhooks, artifact delivery, and credit estimates.
That keeps the extension from becoming a separate product. It is a new packaging of the same insurance extraction infrastructure.
What makes it defensible
Intake gets better as schema memory grows. Every supported line, document family, field alias, correction, and benchmark case improves the quality of normalized submission output.
The defensibility is not a generic upload form. It is the accumulated insurance memory underneath the API.