SearchResidential Mortgage & Home Equity Origination

Mortgage Loan Scenario Search

Mortgage loan scenario search in plain words: every program a borrower fits, and nearly fits

Every program lit green, amber or grey in seconds, with the change that qualifies.

See one case, screen by screen ↓
demo38sfor the six agents to test all 27 cells and rank the answer, every finding cited
demo27cellstested on every scenario — 9 programs × 3 down payments
demo$492a monthless for the Whitfields on VA than on conforming, found because Elena is a veteran
demo22minmedian time to a scenario desk answer, against a 30-minute target
In short

What is Mortgage Loan Scenario Search?

A program match solution for loan officers. You describe the borrower in plain words; six agents ask for the facts that are missing, work out qualifying income the way each program counts it, and test every program on your shelf at every down payment — green eligible, amber eligible with a stated change, grey not — with the guide section behind every finding.

How to choose AI solutions for lending →

Updated

The problem

Why a tricky scenario still waits on the desk until tomorrow

A borrower calls with a story that does not fit one matrix: self-employed with two years of Schedule C, a spouse who served in the Navy, a condo they want to keep. Before quoting anything, the loan officer has to work out which programs on the shelf they fit — conforming, FHA, VA, jumbo, portfolio, non-QM — reading each matrix, running the pricing engine one product at a time, and emailing the scenario desk when the guide leaves it to judgement.

So the loan officer quotes the program they know best. The veteran co-borrower whose entitlement beats conventional by hundreds a month, or the near-miss that one credit card paydown would fix, goes unseen — and the borrower goes to the lender who saw it. Worse, a borrower stretched into a program they do not really fit surfaces after closing as an eligibility or income defect.

typical7.43%of critical defects are borrower-eligibility errors (Q1 2026)
typical20.07%of critical defects come from income and employment (Q1 2026)
Where a non-standard scenario’s hours goestimated
By hand10.5 days
With the solution1 days
  • Turning the call into the facts each program needs0.5 → 0.1 d
  • Working out qualifying income program by program1 → 0.1 d
  • Reading matrices, overlays and loan limits1.5 → 0.1 d
  • Pricing payment, mortgage insurance and cash to close1 → 0.1 d
  • Waiting on the scenario desk6 → 0.5 d
  • Writing up the scenario for the lead0.5 → 0.1 d

Estimated split for one non-standard purchase scenario, in working hours, by hand and with the solution.

How it works

How a scenario moves

Six specialist agents read the scenario, light every program and down payment, and find the change behind each near-miss; the loan officer decides.

What comes in
Scenario inplain words · income, credit, property, other homes
Agents at work
Scenario readerfacts + questions
Then
Income calculator
Program matcher
Then
Payment and cash estimator
Near-miss finder
Then
Scenario note writercited note
A person decides
Loan officerdecides; desk confirms
What comes out
Program match matrix
Near-misses with the change
Cited note on the lead
One case, step by step

One borrower’s story, from the call to the scenario note

Marcus and Elena Whitfield want to buy an $865,000 house in Lakewood with 10% down. He is a self-employed designer; she is a school nurse and a Navy veteran; they are keeping their condo in Aurora. Here is what happens next, screen by screen, in the working solution.

  1. 0108:12

    The borrower’s story, in plain words

    Jordan Pike · Loan officer assistant

    Jordan takes the call and types what the Whitfields told him: the $865,000 house, 10% down, Marcus’s two years of Schedule C at $118,400 then $131,200, Elena’s $68,900 W-2, scores 694 and 741, debts of $1,056 a month, and the condo they are keeping. No form and no product codes — just the story, on lead L-26-10418 for Priya Raman.

    “The agents read the facts from your words and ask only for what changes an answer.”

  2. 02One click

    Eleven facts underlined, two questions asked

    Scenario reader

    Priya presses Match every program. The scenario reader underlines the 11 facts it read in place, then asks the two things the words left out, each with answers to tap: will the Aurora condo be rented with a signed lease, and is it financed with Elena’s VA loan? Yes to both — $2,050 a month from 1 Dec, and a VA loan from 2019 with $268,400 owed.

    It never assumes a fact that was not stated; it asks.

  3. 0338 seconds

    Every program, at every down payment, lit

    Program matcher

    Nine programs against 0%, 5% and 10% down: 5 cells eligible, 7 eligible with a change, 15 not eligible. At their plan of 10% down, 3 programs fit, and VA 30-year fixed comes first at $5,266 a month. Click the cell and the cell detail shows why, rule by rule: partial entitlement needs $86,062.50 down and they bring $86,500; funding fee 1.25%, financed; ratio 39.8%, under 41%; residual income $5,760 against $1,117 for a family of 4 in the West.

    The condo rent on VA is marked a judgement rule — the scenario desk confirms it.

  4. 04Next

    Amber cells say what would qualify

    Near-miss finder

    Every amber cell names its change: conforming at 0% is “Eligible with 5% down”, the Larkmoor Select 10/6 ARM is “Eligible if 720 score”, non-QM bank statement at 10% is “Eligible if 12 months reserves”. Click that one: above $750,000 the matrix needs 12 × $6,113 = $73,356 in reserve, and after closing they would have $45,420 — $27,936 short.

    “It is not needed: they qualify on full documentation at a lower rate.”

  5. 05What-if

    Switch on an honest change, and the matrix re-lights

    Priya Raman · Senior loan officer

    Under the matrix the near-miss finder has left two honest changes: pay Marcus’s card to 30%, or sell the Aurora condo before closing. Priya switches on the condo sale and 3 cells turn green, each marked What-if — VA at 0% down at $5,889 a month, VA at 5% down, and non-QM at 10%. She switches it off again; the Whitfields want to keep the condo.

    “A what-if is a plan to discuss, never a fact on the file.”

  6. 06Ranked

    A short answer, then the ten options ranked

    Scenario note writer

    At their 10% down, three programs fit. VA first: $5,266 a month with no mortgage insurance, $492 a month less than conforming. Three near-misses — the Select ARM needs a 720 score, VA with no money down needs the condo sold, non-QM needs 12 months of reserves — and four that do not fit, each with the reason and the rule behind it.

    3 fit · 3 nearly fit · 4 do not fit · 1 for the desk

  7. 07Side by side

    Payment, ratio and cash, from the 08:30 rate sheet

    Payment and cash estimator

    VA, conforming, FHA, non-QM bank statement and the Select ARM in columns: down payment, base loan, financed fee, rate, principal and interest, taxes and insurance, mortgage insurance, payment, qualifying income and ratio. Conforming carries $298 of mortgage insurance; FHA $454 of annual MIP. Every figure is labelled with the Larkmoor rate sheet time, Mon 12 Oct 2026 08:30.

    Estimates, not a lock — never a locked rate or a Loan Estimate.

  8. 08Plan B

    The near-miss plan, with cost and time

    Near-miss finder

    For the Select ARM at 5% down with no mortgage insurance: pay Marcus’s Cobalt Rewards card from $9,850 to $3,990 for $5,860, order a rapid rescore in 3–5 business days, and the bureau simulator estimates 722. That leaves $35,370 more in the bank than VA, at $133 a month more. Priya can draft the paydown note for Marcus or order the rescore from here.

    “An estimate, not a promise. The score that counts is the one on the rescored report.”

  9. 0908:59 → 09:04

    The judgement call goes to the desk, with the draft

    Tom Reyes · Underwriter, scenario desk

    Priya asks the desk whether the condo rent can count on VA. The question arrives with the agents’ draft and the rule: without landlord experience, VA needs three months of reserves on both homes — 3 × ($5,266 + $1,640) = $20,718 — and they will have $50,360 after closing. Tom Reyes reads it and confirms at 09:04.

    “Confirmed. The lease is signed and reserves cover three months on both homes. Count the rent on VA.”

  10. 1009:06 · saved

    The scenario note, saved to the lead

    Priya Raman · Senior loan officer

    The note carries the borrowers, the recommendation — VA 30-year fixed, 10% down, $778,500 plus the $9,731.25 funding fee at 5.875% — the alternatives that fit, the near-misses, Tom’s confirmation and every citation. A plain-language summary for the Whitfields is drafted beside it. Priya saves the note to the lead at 09:06; nothing goes to the borrowers until she sends it.

Who it’s for

Built for everyone who touches a scenario.

The same borrower, seen by the four people who carry the scenario — what their day looked like, and what it looks like now.

PR
Priya RamanSenior loan officer · Lakewood branch
Loan officer
Before
Reads matrices and runs the pricing engine product by product, then waits on the desk before she can call the borrower back.
Now
Sees every program at every down payment lit in 38 seconds, with the near-misses and the change behind each, before she picks up the phone.
JP
Jordan PikeLoan officer assistant · takes first calls
Assistant
Before
Writes down what the borrower said without knowing which facts the programs care about, so the loan officer calls back for more.
Now
Types the story as told; the scenario reader underlines what it used and asks only the questions that change an answer.
TR
Tom ReyesUnderwriter · scenario desk
Scenario desk
Before
Answers loan officer questions by email and chat all day, many of them things the guide already answers.
Now
Gets only the judgement calls, each with the agents’ draft answer and the guide section, and confirms or corrects it.
AB
Angela BrandtBranch manager · Lakewood
Manager
Before
Finds out a borrower was quoted the wrong program first when the deal walks.
Now
Sees every scenario in the branch with its best program, its near-misses and where it stands, re-checked each night against rule changes.
Built on the engine

6 agents. Each with one job, and hard limits.

Six specialist agents read the scenario, light every program and down payment, and find the change behind each near-miss; the loan officer decides.

Scenario reader

Turns the loan officer’s plain words into the facts the programs need — borrowers, income types, scores, property, price, occupancy, down payment, debts, assets and other homes — and asks for the ones the words leave out.

  • Never assumes a fact that was not stated
  • Asks only questions that change an answer, at most three
Income calculator

Works out qualifying income the way each program counts it: self-employment averages and trends, W-2, rent from a home being left at 75% less its payment, bank-statement income, and VA residual income.

  • Never counts income a program would not count
  • Shows the working for every figure
Program matcher

Tests every option on the program shelf: eligibility, loan limits, VA entitlement and funding fee, mortgage insurance, ratios, reserves and Larkmoor’s overlays, then ranks what fits by monthly cost, then cash to close.

  • Never marks a program as fitting without citing every rule it passed
  • Judgement rules always go to the scenario desk
Payment and cash estimator

Prices each option that fits or nearly fits from today’s rate sheet: payment with taxes, insurance and mortgage insurance, financed fees, cash to close and reserves after closing.

  • Never quotes a locked rate or a Loan Estimate
  • Every figure carries the rate sheet time
Near-miss finder

For each rule a program fails, finds the smallest honest change that would pass it — paying down a debt, a different down payment, selling a property, more months of documents — with what it costs and how long it takes.

  • Never suggests anything that misrepresents the borrower, income or occupancy
  • Score changes are estimates, never promises
Scenario note writer

Writes the scenario note to the lead — facts, recommendation, alternatives, near-misses, desk confirmations, the rules cited and the next documents — and drafts a plain-language summary for the borrower.

  • Never sends anything to the borrower; the loan officer does
  • Estimates are always labelled
Loan officer

Decides what to present; the scenario desk confirms judgement rules. The agents propose; a named person decides.

Ask in plain words

Ask about any program, rule or near-miss

The loan officer can ask about the scenario in plain words, or say what to do next. Every answer cites the rule it rests on.

Why does VA come before conventional for the Whitfields?

Mostly because VA has no monthly mortgage insurance and the lower rate: 5.875% against 6.375%, $5,266 a month against $5,758. That is $492 a month less. The catch is entitlement: Elena’s is partly in use on the condo, so VA needs $86,062.50 down, which their $86,500 covers.

What would get Marcus into the Select ARM?

One rule stands in the way: the Select 10/6 ARM needs a 720 score, and Marcus’s middle score is 694. His card is at 74% of its limit. Paying it from $9,850 to $3,990 costs $5,860, and the bureau simulator estimates 722 after a rapid rescore, in 3–5 business days. It is an estimate, not a promise.

Can we count the Aurora condo’s rent on VA?

Probably yes. VA allows rent from the home being left with a signed lease and either landlord experience or three months of reserves on both homes. They have no landlord history; reserves needed are 3 × ($5,266 + $1,640) = $20,718, and they will have $50,360 after closing. Larkmoor marks this as a judgement rule, so it goes to Tom Reyes on the scenario desk.

Why is non-QM bank statement amber?

Because of reserves, not income. Above $750,000 Larkmoor’s bank-statement matrix needs 12 months of the new payment in reserve: 12 × $6,113 = $73,356. After closing the Whitfields would have $45,420, so they are $27,936 short. It is also not needed: they qualify on full documentation at a lower rate, so VA stays first.

Every screen

The working solution, as it ships.

13 screens from the working solution, on its sample data. Pick one to see it large.

Program matchEvery program on the shelf at every down payment, lit green, amber or grey — with the cell detail and its cited rules beside it.
The scenario in plain wordsThe assistant’s call note, as typed, on the lead — before any program is tested.
Facts read, questions askedEleven facts underlined in place and the two missing ones asked as answers to tap.
An amber cell, explainedNon-QM bank statement at 10% down: eligible if 12 months of reserves, with the shortfall worked out.
Try a changeSelling the condo switched on: three cells turn green, each marked What-if, and the tallies update.
The ranked answerHow the answer was built, step by step, and the short answer with every finding cited.
Side by sideFive options in columns: loan, fees, rate, payment, mortgage insurance, qualifying income and ratio, from the 08:30 rate sheet.
Near-missesThe paydown and rescore plan for the ARM and the entitlement working for VA with no money down.
The scenario deskJudgement calls with the agents’ draft and its rule; the underwriter confirms or corrects.
The scenario noteBorrowers, recommendation, alternatives, near-misses and the desk’s confirmation, saved to the lead with a summary for the borrowers.
Program rulesThe guide sections, overlays and rate sheet lines the answers cite, with the passage applied to the scenario.
Every scenario in the branchEach borrower’s best program, near-misses, stage and loan officer, re-checked each night against rule changes.
Shelf and overlaysPrograms on or off, minimum scores and ratios, what goes to the scenario desk, the rate sheet and the people.
Governance

Built for loan quality: cited, confirmed, on the record.

Every finding cites its ruleEach green, amber or grey cell lists the rules it passed and failed, with the guide section, overlay or rate sheet line behind each — and the passage opens in place.
The loan officer decidesThe agents recommend; the loan officer chooses what to present to the borrower. Nothing goes to a borrower without the loan officer sending it.
Judgement goes to the deskRules the guides leave to judgement — rent from a home being left on VA or FHA, non-QM exceptions — go to a scenario-desk underwriter, whose confirmation is recorded with name and time.
Only honest changesThe near-miss finder suggests real changes the borrower controls — paying a debt, more down, selling a property, more documents — never anything that misrepresents income or occupancy.
Estimates, never a lockPayments and cash to close carry the rate sheet time. They are never shown as a locked rate or a Loan Estimate, and a what-if is never a fact on the file.
On the record, kept with the leadEvery search, what-if, desk confirmation and saved note is on the activity trail, and scenario notes are kept with the lead for 7 years.
Configuration

Your shelf, your overlays

Which programs are tested, how strict the overlays are and what goes to the desk are settings. Changes are versioned and apply to new searches.

SettingDefaultChoose from
Program shelf10 options on, Lakewood branchOn or off per program
Select 10/6 ARM · minimum score720Raised from 700 on 1 Oct
Non-QM · minimum score660660–699 limited to 90% of price
FHA · minimum score580The guide allows 500 with 10% down
Conforming · ratio for manual approval45%Automated approval allows 50%
Send to the scenario desk whenRent from a home being left on VA or FHA · self-employment under two years · any non-QM exceptionAlso offered, switched off: a program that fits by less than 1 point of ratio
Rate sheet refresh08:30 every business dayAnd whenever secondary posts a change
Show payments asEstimates with the rate sheet timeRequired — never as a lock
Connections

Works with the guides and systems you already use

Your leadsthe scenario in, the cited note saved back to the lead
Agency and government guidesSelling Guide, HUD Handbook 4000.1, VA Lender’s Handbook M26-7, USDA
Investor and non-QM matricesportfolio, jumbo, bank statement and DSCR, by revision
Lender overlaysminimum scores, maximum ratios and loan amounts per program
Daily rate sheetfrom secondary marketing, with its time on every figure
Credit what-if simulatorscore estimates for a paydown before a rapid rescore
What it changes

The difference, in numbers.

Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.

target
38s
from a plain-words scenario to ranked programs
By the deskhours to a day
With agents38 s
target
27of 27
program and down-payment cells tested on every scenario, each cited
27 of 27 cells
on every scenario
typical
7.43%
of critical defects are borrower-eligibility errors (Q1 2026)
GuessedTested

“demo” = seen in the working solution, on its sample scenario data · “estimated” = our estimate · “typical” = published figures (ACES via National Mortgage News, Q1 2026 critical defects: borrower eligibility 7.43%, income and employment 20.07%). Program rules in the sample case follow the VA Lender’s Handbook M26-7, HUD Handbook 4000.1 and the FHFA 2026 conforming loan limit of $832,750; Larkmoor’s overlays and rate sheet are sample data. People, lenders and borrowers named on this page are fictional — characters and sample data in the working solution.

Questions

What mortgage lenders ask us.

What is Mortgage Loan Scenario Search?

A program match solution for loan officers. You describe the borrower in plain words; six agents ask for the facts that are missing, work out qualifying income the way each program counts it, and test every program on your shelf at every down payment — green eligible, amber eligible with a stated change, grey not — with the guide section behind every finding.

How does it find near-miss loan programs?

For each rule a program fails, the near-miss finder works out the smallest honest change that would pass it — a credit card paydown and rapid rescore, more down, selling a property, more months of documents — with the cost and the time. Switch a change on and the matrix re-lights, marked What-if, so you can discuss it without it becoming a fact on the file.

Does it handle VA partial entitlement, self-employment and rent from a departing residence?

Yes. In the working solution it computes remaining entitlement against the conforming loan limit and the down payment VA needs, averages two years of Schedule C when income is stable or rising, and counts lease rent from the home being left at 75% less its payment. Rules the guides leave to judgement go to the scenario desk.

Are the payments a rate quote?

No. Payments, mortgage insurance and cash to close are estimates from your daily rate sheet, labelled with its time. The solution never quotes a locked rate or a Loan Estimate.

Do people stay in control?

Yes. The agents recommend; the loan officer decides what to present, and nothing goes to the borrower until the loan officer sends it. Judgement calls are confirmed or corrected by a scenario-desk underwriter, and their confirmation is recorded with name and time.

Can we use our own overlays and program shelf?

Yes. Which programs are on, minimum scores, maximum ratios, what goes to the scenario desk and the rate sheet source and refresh are settings. Changes are versioned and apply to new searches, and open scenarios are re-checked overnight when a rule changes.

Where is the scenario saved?

The scenario note — facts, recommendation, alternatives, near-misses, desk confirmations and citations — is saved to the lead, with a plain-language summary drafted for the borrower. Notes are kept with the lead for 7 years.

How long does it take to go live?

The Agentic Solution Engine builds and deploys it from your requirements — your program shelf, product matrices, overlays and rate sheet — and it goes live once every quality gate has passed. We will walk you through it on a sample of your own scenarios first.

See it on
your program shelf.

We’ll run Mortgage Loan Scenario Search on your own programs, overlays and a sample of your recent scenarios.