E3

Mortgage intelligence

Mortgage intelligence,
built for the depth of the industry.

Our aim is to bring expert mortgage understanding to every stage of the loan lifecycle. We are building specialised intelligence across mortgage documents, policies, calculations and exceptions with quality control as the first application, running on real files today.

See how to evaluate E3 Watch a file go through
EXAMPLE · ONE LOAN FILE, RULES RUNCONVENTIONAL 30YR FIXED
Passed712
Failed31
Data required118
Not applicable44

Look at the third number. Those rules could not run, because a field they needed was missing. E3 reports them separately from passes, because not checked and checked and fine are different answers, and a reviewer deciding where to spend an hour needs to tell them apart. See the whole file go through.

Where we are

The ambition, and what you can put a file through today

Two different things, and a buyer should not have to work out which is which.

THE VISION
A mortgage intelligence layer across the loan lifecycle
Expert understanding of mortgage documents, policies, calculations and exceptions, from origination through servicing and review.
FIRST APPLICATION
Quality control
QC is where that intelligence is applied first, because it is where the reasoning is hardest to fake: every finding has to carry the rule and the document behind it.
WHAT YOU CAN EVALUATE NOW
Complete files, processed end to end
Document classification, field extraction, a synthesised loan record, versioned rules, findings traced to the source document, and a reviewer decision recorded against them. On your own historical files. Set out in full, including what is not built.
WHAT COMES NEXT
Further applications, as they are proven
Each one earns its place by demonstrated capability on real files, not by appearing on a roadmap.

What we are building

Mortgage runs on documents. Understanding them is the whole problem.

Quality control is the first application, not the boundary of the ambition.

Whether to lend. Whether to buy the loan. What to service, and on what terms. Whether a file survives an examination in three years. Every one of those is a judgement about what a few hundred pages actually say, and every one is made today by someone reading a fraction of them.

That is the gap: not a workflow to automate, but a layer the industry has never had. Why we think it is the problem worth solving.

FIGURE Where E3 sits Quality control is not one moment. It is five, spread across the life of the loan. 01Pre-fundingBefore the loan closescure is cheapest here02Post-closingAfter close, before deliverythe sampled review most lenders run03Pre-purchaseAggregator diligencerisk on a loan you did not originate04Post-purchaseAfter the loan is boughtrepurchase exposure surfaces05Servicing & auditTransfers and examinationsthe file must speak for itselfThe same file is examined five times by five parties, each asking a different question of the same documents.E3 runs the same full-file check at every point, and every answer carries the rule version that produced it. Pre-funding, post-closing, pre-purchase, post-purchase, and servicing or audit QC.
The same file, examined five times by five parties, each asking something different.

The domain model

Reading a value and understanding a loan are different problems

Document reading is now a capability anyone can buy. What cannot be bought is the mortgage reasoning that sits on top of it. That is what a domain model is for, and this is the specific list of what it has to get better at.

TODAY: READING
Locate a value on a page
A model finds monthly income on a paystub and reports where. This works, improves industry-wide, and is not a differentiator.
THE HARD PART
Know which value the programme actually wants
Qualifying income is a calculation, not a field. A 24-month average, a year-to-date annualisation and a declining-income treatment can all be defensible for the same borrower, and the right one differs by programme.
THE HARD PART
Tell a legitimate difference from a defect
A paystub, a W-2 and a 1040 cover different periods. They are not expected to match. Telling an ordinary difference from a real problem is what separates a finding from noise.
THE HARD PART
Know what should be in the file and is not
Describing what a file contains is reading. Knowing a required document is absent depends on programme, property, occupancy and jurisdiction. A system with no applicable rule reports nothing, and nothing reads as clean.
THE HARD PART
Explain an exception well enough to act on
A reviewer facing a 203(k) binder or a DSCR file does not need a flag. They need the evidence assembled, the requirement named and the open question stated precisely.

Every row above marked the hard part is work in progress, and today is the first row plus a versioned rule engine and a human reviewer. What runs today, and what does not, set out in full. We will publish evaluations on representative, unseen mortgage cases as they exist: expert-reviewed accuracy, exception handling and the usefulness of the explanation. Those are the measures that would substantiate a claim of leadership. Coverage counts are not.

The first application

Quality control, running on real files today

The rest of this page is what E3 does now: the demonstrated capability the ambition has to be earned from.

What a miss costs
$32,288Estimated average cost per repurchase demand. National Mortgage News, drawing on STRATMOR Group analysis.
What the rate hides
1.50%Critical defect rate, CY2025, flat against 1.52%, while eligibility defects rose 291% and credit 166%. ACES Mortgage QC Industry Trends.
What a loan costs to produce
$11,109Fully loaded total production cost per loan, not QC cost. MBA Quarterly Performance Report, Q3 2025. No per-loan QC figure is published industry-wide.
What changes
Designed for every loan, not a sampleFull-file, cross-document checks rather than a sampled read, so depth stops being the variable that quietly gives. What is on offer today is a controlled evaluation: see capabilities.

How E3 works

Seven stages, and two boundaries

The model reads. Versioned rules decide. A person resolves the exception, and every step is kept with the finding.

How E3 works Seven stages. Intake, Classification and Extraction are performed by a vision-language model. Synthesis and Validation are deterministic and contain no model. Review is performed by a person. Evidence is the retained record. MODEL PROPOSES · PERCEPTION 01Intake Loan file receivedand queued. 02Classification Documents identifiedand segmented. 03Extraction Values located, keptwith page and source. 04Synthesis Reconciled across thefile; conflicts recorded. RULES DISPOSE · NO MODEL 05Validation Versioned rules run.Skips recorded too. 06Review Exception raised.A reviewer resolves it. 07Evidence Finding kept with therule version that made it. PEOPLE DECIDE · RECORDED A model reads. Versioned rules decide. A person resolves the exception, and every step is kept with the finding.
Reading, reconciling, deciding, resolving, and what is kept from each stage.

What a finding looks like

Most defects live between the documents

Experienced reviewers already do this work, and do it well. The constraint is not skill: the checks are numerous, the file is long, and the time is finite.

EXAMPLE FINDINGRULE XDC-INCOME-RECONCILE
QUALIFYING INCOME, FOUR SOURCES SIGNIFICANT
The documents disagree, and one of them is the one that qualified the loan
Paystub annualises to $8,940. W-2 gives $8,875. The 1003 states $9,200 and the 1008 used $9,200 to qualify. Spread 3.7%, inside the 5% tolerance. Not a defect — raised because the 1008 used the stated figure rather than a derived one, and a reviewer decides which basis stands.

No single document is wrong. The finding only exists when the file is read as one thing. Follow a whole file through.

The question you now have to answer

“You are using AI. Show us what it did, and how you knew it was right.”

Fannie Mae's LL-2026-04 took effect 6 August 2026; Freddie Mac's Bulletin 2025-16 has been live since 3 March. Both require a documented AI/ML governance programme, and both hold you to the same standard for the AI your vendors use.

For most lenders the honest answer is a description rather than a record. E3 is built so the answer already exists: every value carries the document it came from, every finding the numbered rule version that produced it.

FIGURE The questions you now have to answer Fannie Mae LL-2026-04, effective 6 August 2026. Freddie Mac Bulletin 2025-16, live since 3 March 2026. WHAT THE GSEs RESERVE THE RIGHT TO ASKWHAT E3 PRODUCESWhy is AI being used?To read documents: a perception task.Named and bounded per stage.For what purpose?Locating values, not deciding outcomes.The disposition comes from a stored rule.What safeguards exist?The deciding path contains no model.Structural, not procedural.How do you know it was right?Page, position and rule version, per finding.The review can be run again.The obligation extends to vendor AI, held to the same standard as your own. E3 produces the evidence; the lender holds the obligation. No approval or endorsement is implied.
The four questions, and what E3 produces in answer to each.

E3 produces the evidence; the lender holds the obligation. Nothing here implies approval or endorsement by Fannie Mae or Freddie Mac. How it holds up under each question.

Controls

The policy E3 enforces, encoded and versioned

Not a checklist held in someone's head. Each control is a stored rule, bound to a loan programme, and the record shows which ran and which were skipped.

TRID, ATR/QM, HOEPA, HMDA, RESPA and Regulation B, alongside the GSE selling guides and FHA, VA and USDA programme requirements. The full control set, and what each one checks.

Your data

We do not train on your borrower documents

The question every lender asks second, so it belongs on the front page rather than in a contract schedule.

TRAINING
Your documents are not training data
Not ours, not the model provider's. Any future domain model would use customer data only under a separate written agreement you sign knowingly.
PROCESSING
A third-party foundation model reads them
A general-purpose vision-language model from an established provider, running in our cloud environment. We say so because implying we built it would be false.
WHERE THIS GOES
Deployment inside your own AWS account
We are building a Terraform deployment that stands E3 up in your account, so files and the processing that reads them stay in your boundary. Not available yet.
RETENTION
Contractual, not a default
Retention, deletion on request and end-of-term destruction are set in the agreement. Ask and we send the terms before a pilot.

Because your obligations under LL-2026-04 extend to us, we expect a vendor assessment and will answer it in writing. Full data handling.

Who we serve

The same file. Different reasons for caring what is in it.

Eleven segments, from third-party QC firms to servicers. The mechanism does not change; which rules matter and which consequence lands first do.

The engine is programme-configurable rather than segment-specific, which is why it applies broadly. Segment-specific workflows, correspondent and wholesale, servicing review, housing finance programmes, are planned rather than built, and we say which is which. All eleven, and what differs in each.

See it run on a real file.

A working session on a synthetic file, or on de-identified files of your own. You see how each value was read, which rule was applied, and what the record looks like when someone asks in eighteen months.

Schedule a demo Email us