A regional insurance carrier came to Mantle with a familiar problem: the work of writing a policy hadn't gotten any harder, but the volume of it had outgrown the team handling it. New business requests arrived by email, by phone, and as scanned attachments, and every request needed the same manual relay — someone reads it, someone logs it, someone types up a proposal — before an underwriter ever saw it.
None of that relay work required judgment. It just required someone to move information from one format into another, correctly, every time. That's exactly the kind of work Mantle is built to absorb.
What the automated intake pipeline does
The carrier's workflow now starts the moment a request arrives, with no manual sorting required.
Mailbox monitoring watches the intake inbox continuously and picks up new submissions the instant they land — no one has to check a folder or forward an email to get a request moving.
Call logging and transcription cover the requests that come in by phone. Calls with agents and applicants are logged automatically and transcribed into text, so a phone conversation becomes searchable, reviewable data instead of a note someone might or might not have taken.
Document parsing handles everything that arrives as a file — applications, supporting forms, scanned attachments — extracting the specific fields underwriting actually needs (coverage type, applicant details, property or asset information) instead of leaving someone to retype them by hand.
Proposal generation takes the structured information gathered from all three sources — email, calls, and documents — and drafts the initial proposal automatically, formatted and ready for an underwriter to review.
At every step, the workflow keeps the tracking ID, retry path, and audit trail that let the carrier's compliance team see exactly which pieces of information a proposal was built from and when — and a human reviews and signs off on the finished proposal before anything goes to the applicant. AI does the assembly; a person makes the call.
What changed
Requests that used to sit in an inbox waiting for someone to have time now start moving the moment they arrive, regardless of whether they came in by email, phone, or paper. Underwriters spend their time evaluating proposals instead of assembling them. And because every step is logged automatically, showing a regulator or an auditor exactly how a given proposal came together is a lookup, not a research project.
None of the individual pieces here are exotic — monitoring a mailbox, transcribing a call, reading a form, drafting a document are all things a person can do. What changes when Mantle runs them is that they happen continuously, at whatever volume shows up, with a complete record attached, freeing the people who used to do that work to do the parts of it that actually need a person.
That pattern — route the intake work that doesn't need judgment, keep a human on the decisions that do — isn't specific to insurance. It's the same shape as workflows we've built for operations, support, and data teams across other industries facing the same underlying problem: more requests than the manual process was ever designed to handle.
KEY TAKEAWAY
Insurance Intake Case Study: An anonymized carrier automated the judgment-free parts of policy intake — mailbox monitoring, call logging/transcription, document parsing, and proposal drafting — so requests start moving the instant they arrive and underwriters spend their time deciding, not assembling, with a full audit trail for compliance.