CapSeal - Mortgage lender scenario

This is an illustrative scenario, not a customer. CapSeal is pre-launch. Every figure below is modelled from published industry data and stated assumptions, which are set out in full further down so you can disagree with them. We will replace these with named, measured references when we have them, and not before.

The situation

Where the organisation starts

A digital lender writes roughly 6,000 loans a year, fully online, with income documents uploaded as PDFs and screenshots.
Verification is a manual read plus a bank-feed check where the applicant consents. Consent rates are well below 100%, so documents carry the weight.
Editing a PDF payslip requires no skill. The lender's fraud team believes a material share of arrears traces back to income that was never real.

Modelled outcome

What the model produces

94%

of sealed income evidence decided without manual read

11 hrs

median time-to-conditional-approval, from 3 days

$4.2M

modelled exposure avoided annually

additional friction for consenting applicants

The assumptions behind those numbers

6,000 settlements a year, average loan $620,000.
Application fraud at 0.9% of originations, toward the lower end of published estimates for prime lending.
Average loss given default on fraudulently obtained lending of 12% of drawn balance.
Sealed capture used for income and identity documents only; bank-feed verification continues unchanged where consent is given.
Exposure avoided assumes detection at origination rather than recovery after default, which is where the value actually sits.

What changes

How the process differs

The applicant captures, rather than uploads

Payslips and identity documents are photographed through the lender's app. A real lens seeing a real document is measured on the device.

Serviceability runs on evidence with a basis

The decision engine reads the verdict basis, not a score. Proven and declared results proceed; anything else routes to a credit officer with the report attached.

Manual review gets its time back

The team stops reading thousands of genuine documents to find a handful of forgeries, and starts seeing only the cases the evidence could not settle.

Nothing changes for the honest applicant

A borrower who captures in the app is approved faster. One who emails a PDF is treated exactly as today, not worse.

Honest limits

What this model does not claim

A model is an argument, not evidence. These figures show what follows if the assumptions hold for your organisation, and the point of publishing them in full is so you can check whether they do.

The full model behind these figures, with every input cited to published research, is in the industry brief.

Sealing evidence at capture does not detect every fabrication, does not establish that a truthful-looking document is true, and does not determine intent. The Product Disclosure sets out the boundary precisely.

Run the model on your origination volume.

We will show you the assumptions and let you change them.

Book a pilot Check a file first Lending brief (PDF)