Eric DeMatteis

Typing loan data into the LOS.

Your processors should be on the phone clearing conditions.

Talking with borrowers, agents, and underwriters, not reading every page of a title commitment. I connect your tools to your LOS so the data entry, filing, and first review happen on their own. It all runs in your own accounts, and your team learns to own it.

Book a 30-minute call See what's running

Built and running every day at a 16-state mortgage broker, on Arive.

Automation logTuesday morning
  1. Closing reminder sent to the team6 loans closing this week, open conditions and title contacts attached
    Sent
  2. Title review found a lien mismatchPayoff amount differs from the 1003 on a refinance
    Flagged for processor
  3. VOE filled out and emailedBuilt from the loan number alone, authorization form attached
    Done
  4. Purchase advice checkedInvestor figures match the loan, PDF filed to Arive
    Complete
  5. Wire expiring in 2 daysPost-closing board updated and the team alerted
    Needs attention
Example activity based on tools in production. No borrower data shown.

More than 35 of these automations run every business day at a mortgage broker licensed in 16 states.

I built them. Now I build them for other lenders.

Every document gets a first read before your team opens it.

AI reads the document and checks every claim against the loan file. Your processor opens a one-page summary with the mismatches at the top, not a 40-page PDF.

  • Title commitments
  • Purchase contracts
  • Homeowner's insurance
  • Appraisals and 1004Ds
  • Disclosures
  • Purchase advices
Title commitment reviewSample refinance · 6 checks
  • Payoff lender doesn't match the applicationSchedule B-I, page 4 lists a different lien holder than the 1003
  • Borrower names match the application
  • Property address and county match
  • Loan amount matches the commitment
  • No open judgments or tax liens
  • Vesting matches the borrowers on title
Uploaded to the loan in AriveExample, no borrower data shown

What I build for lenders

All aimed at the same thing: your people spend their time on borrowers, agents, and closings, not on screens.

Connect your tools to your LOS

Arive, Microsoft 365, pricing, title, your warehouse bank, accounting, email marketing. Data moves between them on its own, so nobody copies it by hand.

ExampleClosing disclosures are merged, matched to the right loan, and uploaded to Arive, without logging in to the doc vendor.

Document review before anyone opens the file

Title, contracts, insurance, and appraisals get checked against the loan first. Your team only looks at what doesn't match.

ExampleAn appraisal is sorted (report, 1004D, supplement), its dates and rental income pulled, the expiration set, and the file uploaded.

Purchase advice automation

When an investor buys a loan, the purchase advice is pulled from their email or portal, the figures go into your LOS, and the PDF is filed. Your team gets a Complete or Mismatch notice.

ExampleRunning today across nine investors, from email-based lenders to portal-only ones.

Sales automation

Your loan officers stay in front of past clients and partners without writing every email. Follow-ups drafted in their own voice, a client list they can search by rate or birthday, and an intro email the moment a new prospect comes in.

ExampleEvery funded borrower's rate is watched against the market. Twice a year they get a personal review, with exact savings when a refinance makes sense.

Automate the workflows that eat the day

The same twenty steps, done the same way on every file, are exactly what software should do. Your team just checks the result.

ExampleA verification-of-employment form filled out from the Arive loan number and emailed to the processor with the authorization attached.

Build what your vendors don't sell

A pricing engine on your website. A post-closing board for wires and expirations. The dashboard your LOS doesn't have. If you keep a spreadsheet to track it, it can be a tool.

ExampleA live pipeline dashboard refreshed every 10 minutes, without paying for Power BI.

Your LOS in the middle. Everything else plugged in.

Most lenders run a dozen systems that don't talk to each other. People fill the gaps. I close them.

  • Built on your LOS's own API wherever it has one
  • Secure automated sign-in where the API stops
  • No tools that depend on someone logging in every morning
Microsoft 365email, Teams, files Website pricingyour rate sheets Title & insurancereview and upload Warehouse bankwires, funding Accountinginvoices, entries Email marketingpast clients, partners Creditliabilities, reports Investorslocks, purchase advice Your LOSArive, and others

Ask your team what they'd stop doing.

That list is the build plan. Here is what processors, post-closing, and loan officers usually say, and where that time goes once it's automated.

What they'd stop doing

  • Typing employment details into a VOE form
  • Dragging documents into the LOS one at a time
  • Reading every title commitment and insurance policy line by line
  • Checking a spreadsheet for expiring locks, credit, and wires
  • Chasing the status of every file by email

What they do instead

  • On the phone clearing conditions with borrowers and underwriters
  • Opening a one-page review with the mismatches first
  • Working from documents already filed to the right loan
  • Getting an alert before anything expires
  • Starting the day with a closing list and open conditions

What they'd stop doing

  • Logging into investor portals to download purchase advices
  • Typing purchase figures into the LOS
  • Comparing the investor's numbers to the file by hand
  • Building an invoice every time a loan is purchased
  • Sorting closing documents into folders

What they do instead

  • Handling only the purchase advices that come back as a mismatch
  • Tracking wires and expirations on one board
  • Reviewing invoices and journal entries created at broker-check
  • Working from closing folders assembled as the documents arrive
  • Following up with investors on real exceptions

What they'd stop doing

  • Writing the same follow-up email one borrower at a time
  • Scrolling reports to figure out who to call
  • Building preapproval letters and findings by hand
  • Formatting call notes into the template
  • Screenshotting rate options into emails

What they do instead

  • Calling past clients whose rate is above today's market
  • Sending personalized follow-ups that are already drafted in Outlook
  • Sending a full preapproval package in one click
  • Talking with agents while call notes format themselves
  • Quoting from pricing that is already on their screen

Nothing sends or changes a loan without the rules your team sets. People stay in charge. The tools do the typing.

How working together goes

A three-month start, then you decide how much help you want.

1

Set up in your accounts

Weeks 1–2

A secure automation server in your Azure account, connected to your LOS and Microsoft 365. You own every login and every line of code.

2

Build from your team's list

Months 1–2

We start with what your people most want to stop doing. Document review, purchase advices, and follow-ups are the usual first picks.

3

Hand it to your team

Month 3

Weekly training with someone on your staff, so they can change, fix, and build tools after I step back.

You own everythingServers, code, accounts, and data live in your accounts from day one.
Borrower data stays homeNothing is copied out of your Microsoft and Azure environment.
Least accessA named account with MFA, admin rights only during setup, every grant logged.
Built by an operatorI run the same tools on a real pipeline, so I know where they break.

Ask your team what they'd stop doing. Then bring me the list.

In 30 minutes we'll go through it together: what can be automated, what can't, and what I'd build first.

Book a 30-minute call