Solution

Income, Asset & Liability Verification Intelligence

For mortgage and consumer lenders — processors, underwriters and operations leaders: borrower income, asset and liability evidence extracted, cross-checked and cited before the file reaches underwriting.

PaystubsW-2sBank statements1040sCredit reports
Minutes from upload to verified figures100% of numbers cited to sourceEvery decision human-approved

The problem

Why this exists

12 docs

One file, a dozen layouts

Paystubs, W-2s, statements and returns each arrive in their own format — some scanned at an angle on a phone. A processor keys them all by hand.

Twice

Every figure gets checked twice

The processor keys the numbers, the underwriter re-checks them, and a single mismatch sends the file back to the start of the queue.

Late

Defects surface in underwriting

Expired documents, missing pages and wrong tax years hide until an underwriter opens the file — days after they could have been fixed at intake.

The product, not a promise

An evidence package you can interrogate

Income, Asset & Liability Verification Intelligence — workspace
Monthly qualifying incomecross-checked: paystub ↔ W-2cited
Liquid assetsverified against statement activitycited
Recurring obligationsmatched to credit reportcited
Program checklistcomplete — all documents currentcited
YTD income differs from W-2 run rate — both source pages side by side for reviewverify
HUMAN-APPROVED BEFORE IT POSTS

How it works

File in. Answer out.

  1. 1

    Intake

    Receive borrower documents from portal, email or LOS — paystubs, W-2s, statements, returns, in any format.

  2. 2

    Classify

    Identify each document type and match it to the verification checklist for the loan program.

  3. 3

    Extract

    Pull income figures, asset balances and liability obligations with the source page attached to each.

  4. 4

    Cross-check

    Reconcile figures across documents — paystub against W-2, stated assets against statements — and flag mismatches.

  5. 5

    Hand off

    Deliver a verified, cited income-asset-liability summary to the processor or underwriter queue.

Who it's for

Built for the people who own the outcome

Loan processor

Files arrive with the figures already keyed, checked and cited.

  • Zero rekeying from scans
  • Mismatches shown with both source pages
  • Missing and stale documents caught at intake

Underwriter

Review starts from verified numbers instead of raw PDFs.

  • Every figure clicks through to its source page
  • Cross-checks already run and logged
  • Time goes to judgment, not transcription

Head of lending operations

Throughput scales without stacking processors.

  • Upload to verified figures in minutes
  • Defects caught before underwriting
  • A file that stands up in QC and repurchase review
Mortgage lendingConsumer lendingAuto financeCredit unionsFintech lendersLoan servicing
Minutesfrom upload to verified figures
100%numbers cited to source
Any formatscanned or digital
Every decisionhuman-approved

Income verification is where loan files stall. A single application can carry a dozen evidence documents — paystubs, W-2s, bank statements, tax returns, credit reports — each in its own layout, some scanned at an angle on a phone. A processor keys the figures into the LOS by hand, an underwriter re-checks them, and any mismatch sends the file back to the start of the queue.

Botminds reads the whole evidence package in one pass. It classifies each document against the program’s verification checklist, extracts income, asset and liability figures regardless of layout or scan quality, and normalizes them into the structure your underwriting process expects — monthly qualifying income, liquid assets, recurring obligations.

Verified means cross-checked

Extraction alone verifies nothing. The platform cross-checks figures across the package: year-to-date paystub income against the W-2, stated deposits against bank statement activity, disclosed liabilities against the credit report. Consistent evidence flows straight through; discrepancies are flagged with both source pages side by side so a processor resolves them in seconds instead of hunting through the file. Stale documents, missing pages and wrong tax years are caught at intake, before an underwriter ever opens the loan.

Why governed matters here

A qualifying income number without a source is a finding waiting to happen — in QC, in a repurchase review, or in an exam. Every figure Botminds produces is cited to the exact page it came from, every cross-check is logged, and no verification result reaches the underwriter without passing the checks your policy defines. Underwriters approve decisions; the platform never does. The output is a loan file where each number can be defended months later with one click — which is what makes the speed usable, not just impressive.

Objections, answered

What teams ask us first

How accurate is extraction on bad scans?

It reads phone photos and skewed scans, and any value it is less confident about is flagged with the source page attached rather than passed through. A processor confirms flagged values in seconds; nothing low-confidence reaches the underwriter silently.

Does it follow our program guidelines and income calculations?

Yes — the verification checklist and income treatment are yours: qualifying income rules, asset seasoning, document freshness windows. The platform applies your policy on every file, consistently.

What happens in a QC audit or repurchase review?

Every figure is cited to the exact page it came from and every cross-check is logged, so any number in the file can be defended months later with one click. The trail includes who approved what, and when.

How does it fit our LOS?

Documents come in from portal, email or the LOS itself, and verified summaries post back to the processor or underwriter queue. It slots into the existing pipeline rather than replacing it.

Bring your messiest borrower file.

Watch a full evidence package become verified, cross-checked, cited figures — live in the demo.

Request a demo