Case Studies / Finance & Mortgage
How a mortgage brokerage stopped manually keying loan documents into their CRM
Published
5 min read
Some client identities are withheld due to confidentiality and engagement agreements. References are available where appropriate.
20 min → <1 min
Per document processed
~180/week
Emailed documents handled
0 re-keying
Manual data entry eliminated
Challenge
The brokerage receives a steady stream of loan documents by email - payslips, bank statements, identification, signed agreements - each of which had to be manually opened, matched to the right borrower file, and keyed into the CRM by hand. At close to 180 documents a week, this created real backlog and left borrower files frequently out of date.
Solution
PreSorted reads each inbound email and its attachments together, identifies the borrower, loan file, and document type, and checks the loan's current status before proposing a CRM update for a processor to approve. Where documents are still outstanding, it drafts a follow-up request checked against the actual loan file.
Outcome
Document handling time dropped from around 20 minutes to under a minute per document. The processing team absorbed a 30% increase in loan volume over the following two quarters without hiring, and borrower files now reliably reflect what has actually been received.
This mortgage brokerage processes home loans for clients across greater Melbourne, which means a steady stream of email attachments: payslips, bank statements, identification, signed loan documents, and conditional approval letters from lenders. Every one of those documents needed to end up correctly logged against the right borrower's file in HubSpot before processing could move forward.
That logging work fell entirely on the loan processing team, and it was almost pure manual data entry - open the attachment, read the relevant details, find the borrower's record, type them in, then check where that loan actually stood.
A CRM that was always a step behind
A single document could take a processor 15 to 20 minutes to handle properly: opening the email, reading the attachment, identifying which loan file it belonged to, keying the relevant fields into HubSpot, and then checking the application's current status to see what, if anything, was still outstanding.
With close to 180 documents arriving by email in a typical week, that added up to real backlog. Borrower files often lagged behind what had actually been received, which meant processors sometimes chased documents that had already arrived, or missed following up on ones that hadn't.
Reading the document, not just the email
PreSorted now reads each inbound email and its attachments together, identifying the borrower, the loan file, and the document type - payslip, bank statement, ID, signed agreement. It matches that to the correct record in HubSpot and checks the loan's current status against what's still required before drafting an update.
The proposed CRM update goes to a processor for a quick approval rather than manual entry from scratch, and if something is still missing, PreSorted drafts a follow-up email to the borrower listing exactly what's outstanding - checked against the loan file, not a generic template.
- Reads the email and identifies the borrower and loan file
- Recognises the document type and extracts the relevant details
- Checks the loan's current status against requirements
- Proposes the CRM update and, where needed, a follow-up request for missing documents
From data entry to actual processing
Document handling time dropped from around 20 minutes to under a minute per document - mostly the time it takes a processor to glance at the proposed update and approve it. That freed enough capacity that the processing team absorbed a 30% increase in loan volume over the following two quarters without hiring.
The bigger change was less measurable but just as real: borrower files stayed current. Processors stopped discovering gaps during a status check and started trusting that what was in HubSpot matched what had actually arrived.
For a brokerage where every loan file depends on dozens of small, document-heavy touchpoints, removing the manual re-typing didn't just save time - it removed the main source of the errors and delays that used to slow loans down before they ever reached underwriting.
About Victorian Mortgage Brokerage
Location
Melbourne, VIC
Team
22 employees across broking and loan processing
Systems connected
HubSpot
Ready to get started?
Put PreSorted to work inside your actual stack
Connect the systems you already run and let PreSorted prepare every request - so your team decides, not digs.
Enterprise ready · SOC 2 · 50+ connectors