# Indexing Agent

*For Lenders · Every page split, named and filed the moment it lands*

> An AI indexing desk that splits, classifies, names and files every mortgage document into the right LOS folder the moment it arrives, from any channel.

Splits, classifies and names every document that reaches the file — upload, email, fax or scanned package — and files each into the right LOS folder.

**Replaces** — The stacking queue, the misfiled paystub and the combined PDF someone splits by hand.

**Unlocks** — A loan file where every page is already in its folder, named and dated, the moment a processor opens it.

## What the Indexing Agent does

- Takes in documents from every channel — the POS upload, the borrower’s email, the fax, the title company’s closing package
- Splits a combined PDF into its documents and classifies each page, including the ones scanned sideways or photographed on a kitchen table
- Names each document to your convention, dates it, and files it into the right eFolder bucket in the LOS
- Keeps the latest version current — the newer paystub supersedes the older one rather than sitting beside it
- Stacks the file in the order underwriting, closing and the investor’s delivery each expect

## Where it stops

A page it cannot classify above its confidence floor is never guessed into a folder. It goes to your processor as a named exception, with the page, the likeliest document types and why it could not choose.

## No published study yet

Send us a batch of unindexed documents — the combined PDFs, the faxes, the phone photos. You will see every page split, named and filed against your own folder structure, and which pages it put in front of a person and why.

## What it runs on

- /platform/document-intelligence

---

_Source: [https://seiright.com/for-lenders/indexing-agent](https://seiright.com/for-lenders/indexing-agent) · Sei AI_