Workflows

Mortgage workflow automation, one process at a time

Start with focused workflows that reduce repetitive manual work, then expand across processing, underwriting, closing, and custom operations.

How a workflow runs

From documents to reviewable output

A trigger fires, Loyola AI does the work, and you get a source-backed result you can check in seconds.

Trigger

New bank statement uploaded

checking_0492.pdf · 4 pages

Loyola AI workflow

  • Classify & extract
  • Parse accounts → 1003
  • Cross-check vs. LOS

Reviewable output

2 accounts added to 1003
!Large deposit flagged · $14,200

AI summary

Sample workflows

Workflows for the data-heavy work in every loan

From document intake to clear-to-close — each one configured around your process and the tools you already use, then extended by your team over time.

Documents

Classify & organize documents

As files arrive, Loyola AI recognizes each one and files it the way your team already works — your folders, your naming.

  • Automatic classification and stacking order
  • Keeps the latest version, clears duplicates
  • Flags missing documents at intake
🗂️Document intake
14 files

Paystub_Sep.pdf → Income

Classified

Chase_stmt_0492.pdf → Assets

Classified

URLA_1003.pdf → Application

Classified

W-2_2023 (copy).pdf → Duplicate

Cleared

W-2 · 2024

Missing

Auto-filed to your folders & naming — missing items flagged at intake.

Extraction

Extract exactly the data you need

Reads any mortgage document and returns the structured fields you ask for — not a fixed set of canned fields.

  • Custom extraction you define
  • Handles new and non-standard documents
  • Structured output built to your spec
🔍Data extraction
Pay stub · p.1

Borrower

Maria S. Alvarez

Employer

Northwind Health

Base pay

$6,250/mo

YTD gross

$58,900

Pay cycle

Bi-weekly

Custom fields you define — structured to your spec, no template.

Validation

Validate the file against its documents

Checks the data in your LOS and 1003 against the source documents — no more stare-and-compare. Mismatches raise an alert at intake.

  • Validates LOS / 1003 against source docs
  • Catches keying errors and stale data
  • Points to the exact document
Data validation
1003 ↔ source docs

Note rate

6.875%Matches

Loan amount

$420,000Matches

Subject address

Matches

Monthly income

1003 shows $7,450

$7,100Mismatch

Mismatches flagged at intake, pointing to the document.

Income

Build reviewable income worksheets

Reads pay stubs, W-2s, and returns, then itemizes each by earning type — with every figure traced back to its source.

  • Wage earner, self-employed (add-backs), and rental
  • Pay stub itemization: base, OT, bonus, YTD
  • Source-linked — review in seconds
💵Income worksheet
Wage earner

Base pay

Pay stub · p.1

$6,250

Overtime — 12-mo avg

Pay stubs · YTD

$540

Bonus — 2-yr avg

W-2 · 2023–24

$310
Monthly qualifying income$7,100

Every figure links to its source — review in seconds.

Assets

Review assets and source deposits

Reads each bank statement, parses out accounts and balances, adds them to the 1003, then surfaces items worth a look.

  • Accounts and balances parsed into the 1003
  • Large deposits flagged for sourcing
  • Asset summaries and missing-document checks
🏦Asset review
2-month statements

Chase checking ••4092

$18,430

Chase savings ••7741

$52,180

Large deposit · $14,200

needs sourcing

Review
Verified assets$70,610

Parsed into the 1003; large deposits flagged for sourcing.

Property

Read the purchase-contract package

Reads the contract, addenda, and counteroffers as one picture, extracts the terms, and keeps them current end-to-end.

  • Closing price, key dates, contingency dates
  • Concessions and cost allocations
  • Updates as terms are amended
📑Purchase contract
Contract + 2 addenda

Purchase price

$525,000

Seller credit

$7,500

Close of escrow

Oct 18, 2026

Appraisal contingency

Sep 29, 2026

Financing contingency

Removed

Kept current across addenda and counteroffers.

Underwriting

Turn any checklist into a workflow

Define a checklist and Loyola AI verifies each item against the file, attaches the evidence, and surfaces what needs a human.

  • Conditions (PTD/PTF), closing, investor, and QC
  • Cleared / needs-review / missing, with evidence
  • Catches gaps at intake, not at the closing table
☑️Prior-to-doc conditions
2 of 4 cleared

Verify employment (VOE)

VOE · received

Cleared

Source large deposit $14,200

Bank stmt · p.3

Cleared

Updated hazard policy

expires before close

Needs review

HOA master policy

not in file

Missing

Each item checked against the file, with the evidence attached.

Orchestration

Orchestrate the tools you already use

Loyola AI is the layer on top of your LOS that picks up your existing tools and puts them to work inside a task — the way a person would.

  • Runs on top of your system of record
  • Pulls property data, runs credit / AUS, orders verifications
  • Writes results back to the LOS, email, and reports
🛠️Tool orchestration
Runs on top of your LOS

Pull property data

Redfin · public records

Done

Flood determination (FEMA)

Zone X

Run AUS — DU / LPA

Approve/Eligible

Write back to LOS · email

Sent

Uses the tools your team already relies on.

Ready to automate your first mortgage workflow?

Start with a focused workflow, launch quickly, and expand as your team sees value.