E3

Walkthrough

One loan file,
from delivery to a signed-off report.

What a reviewer actually does, in the order they do it. Four findings, deliberately chosen: a real exception, a legitimate difference that must not become a defect, an extraction error the system catches on itself, and a formatting failure that belongs to someone else.


What you are looking at. These are reconstructions of the working application, not screenshots, and the loan is synthetic. The states, rule families, severities and flow are the ones the product uses. The rule logic behind them is not published. If you want the real screens, that is what a working session is for, and it takes an hour.

01 · Intake

The file arrives whole, not page by page

A loan package is delivered into a processing session. One PDF or many, in whatever order the file was assembled. Nothing is built on your side for this: delivery is how an evaluation starts, and connectors are on the roadmap rather than in place.

02 · Classify

The package is split into documents, and each is named

Segmentation identifies where one document ends and the next begins, then classifies each against the supported form set. Pages that cannot be classified are surfaced rather than silently dropped, because an unread page is exactly where a missed defect hides.

03 · Extract

Values are read, with confidence kept alongside them

Fields are extracted with a confidence score retained per value. Low-confidence reads are marked for correction before anything downstream runs. This is the layer a general-purpose model does well, and it is the layer where the errors that matter get introduced.

04 · Verify

A person fixes what the system is unsure about

Reviewers inspect and correct uncertain values first. Judgement stays where the system is least confident, and it stays before the rules run rather than after, so a misread does not propagate into a finding that then has to be argued about.

05 · Synthesise

Documents become one loan record, with conflicts visible

Borrower, property, income, asset and liability data are consolidated into a single structured record. Where two documents disagree, the conflict is carried forward as a conflict rather than resolved silently by whichever document was read last. Derived values show the formula that produced them.

06 · Validate

Rules run, and four outcomes are recorded, not two

Rule sets are selected by loan programme. Every rule returns one of four states, and the distinction between the last three is the point:

RUN · LOAN LN-0002841PROGRAMME: CONVENTIONAL 30YR FIXED · INVESTOR: FANNIE MAE
Passed712
Failed31
Data required118
Not applicable44

Data required means the rule could not be evaluated because a field it needs is empty. Not applicable means the rule did not apply to this loan. Neither is a pass. A system that reports only pass and fail turns both of these into silence, and silence reads as clean.

07 · Findings

Four results, and why each is a different kind of thing

The list below is what a reviewer opens. Note that only one of these is a defect.

FINDINGSSORTED: FAILED FIRST
XDC-INCOME-RECONCILE SIGNIFICANT
Qualifying income does not reconcile across sources
Paystub year-to-date annualises to $8,940/mo. W-2 gives $8,875/mo. The 1003 states $9,200/mo and the 1008 used $9,200 to qualify. Spread 3.7%, inside the 5% tolerance, but the 1008 used the stated figure rather than a derived one. Not a defect. Raised for a reviewer to accept the basis or request the written calculation.
XDC-TRID-CD-TIMING DATA REQUIRED
Closing Disclosure timing could not be established
The rule needs the borrower receipt date. The Closing Disclosure was read and classified, but the receipt date is not present on the document in the file. Recorded as data required, not as passed. A reviewer supplies the date from the disclosure platform, or the rule stays unevaluated and says so in the report.
DQ-1040-TOTALS WARNING
Extraction error caught by an internal consistency rule
On the 1040, total tax read as greater than taxable income. That cannot be true, so the value is wrong rather than the loan. The rule exists to catch a misread, not a defect. A reviewer corrects the value and the affected income rules re-run against the corrected figure.
COL-UAD-CONDITION DATA INTEGRITY
Appraisal condition rating outside the permitted set
The condition rating on the 1004 is not one of C1 to C6. A formatting failure on the appraisal rather than a valuation problem, and it is returned to the appraiser rather than resolved in review.
08 · Disposition

A named person decides, and that is recorded

The reviewer accepts, corrects or escalates each finding, and the loan-level disposition is captured with the deciding reviewer against it. E3 produces findings; a person decides. That is a design position rather than a gap, and it is not on a roadmap to be removed.

Queues, assignment and escalation are still being built. Today a reviewer works the list; the workflow around it is coming.

09 · Report

The record that has to survive being questioned

A quality control report is produced with export. Each finding carries the rule that produced it, its version, and the source document each value came from. Rules that did not run are in the report as rules that did not run.

EXPORTED RECORD · EXTRACTLN-0002841 · RUN #14 · 2026-09-08
FINDING · XDC-INCOME-RECONCILE SIGNIFICANT
Qualifying income does not reconcile across sources
Rule version 2.3.0, published 2026-07-14, active at run time.
Values used $8,940/mo derived from paystub 03/2026 · $8,875/mo derived from W-2 2025 · $9,200/mo stated on URLA 1003 · $9,200/mo used to qualify on 1008.
Source documents each value traced to the document it was read from.
Threshold 5%. Observed spread 3.7%.
Disposition Accepted by reviewer, 2026-09-08. Basis: underwriter used the stated figure with a written calculation on file. Reviewer recorded against the loan.
NOT EVALUATED · 118 RULES DATA REQUIRED
Rules that could not run, and the field each was waiting on
Listed individually in the export with the missing field named. These are reported as unevaluated rather than folded into the pass count, which is the difference between a report that says what was checked and one that implies it.

That last block is the part an examiner asks for eighteen months later. It is a by-product of doing the review, not a document assembled afterwards to satisfy them.


What this does not show

The limits of a walkthrough

A curated example proves a system can do a thing. It does not establish how often it does it correctly, and we would rather say so than let a tidy example imply an accuracy claim.

We do not publish an extraction accuracy percentage, because we have not measured one against a stated document set we could defend. What we can do is run your own de-identified files and let you check the findings against outcomes you already know. On three to five files that establishes whether it is worth going further. It does not establish portfolio accuracy, and nobody should tell you otherwise.

See this on a real file.

An hour, on a synthetic file or de-identified files of your own. You see the actual screens, not reconstructions.