Master PRD Optimization Map • Decision Tree • Nova Conversation Logic BIX Technology Corp | Purlend / PreFi, Inc. | CONFIDENTIAL | v1.0
This document is the primary product specification for the Clarity Engine. It synthesizes three source frameworks — the Mortgage Optimization Map (150 transformation moves), the Algorithmic Decision Tree, and the Nova Conversation Logic — into a single engineering-ready PRD. BIX should treat this document as the system’s architectural constitution.
“Traditional lending systems ask: ‘What can this borrower get?’ The Clarity Engine asks: ‘What financial structure creates the best life outcome?’” — The Product Truth
The moat is not the mortgage math. Every LOS can calculate a payment. The moat is the optimization engine — which answers: ‘What combination of financial moves produces the best outcome for this borrower?’ That answer requires: a complete borrower model, a complete program library, a full transformation move library, a scenario engine that evaluates combinations, and a ranking engine that weights by borrower purpose. That is what BIX is building. That is what no one else has.
Engine Architecture: The Seven-Layer Stack BIX should understand the Clarity Engine as seven sequential layers, each feeding the next:
1
Borrower + property + deal profile ingestion from URLA 1003, credit report, and conversation 2
Run borrower through all eligible programs (VA, USDA, Conv, FHA, Non-QM, CRA, Portfolio) to find current qualification ceiling 3
Classify what is constraining the outcome: DTI, cash-to-close, credit, program, property, or timing 4
Search the transformation move library for all applicable levers given the detected blocker(s) 5
Generate and evaluate thousands of move combinations, each producing a distinct outcome scenario 6
Score and rank scenarios against borrower’s stated priorities (payment, cash, purchase power, speed, wealth) 7
Nova synthesizes top-ranked scenarios into clear borrower conversation with tradeoffs and next steps
The decision tree is the engine’s control flow. It determines which optimization paths are searched, in which order, and under what conditions.
Everything starts with purpose. The engine must classify the borrower’s primary intent before any calculation runs. Intent assignment controls which optimization weights are applied downstream.
Intent Category Sub-Intents Optimization Weights Applied
Max purchase power | Lowest payment | Least cash to close | Best wealth outcome | Fastest approval Payment comfort: HIGH | Purchase power: HIGH | Cash preservation: HIGH
Lower payment | Lower rate | Shorter term | Cash out | Debt consolidation | Payment reset Monthly savings: HIGH | Interest savings: MEDIUM | Cash access: VARIABLE
Access capital | Remove high-interest debt | Fund ADU/repairs | Preserve reserves Cash access: HIGH | Payment neutrality: HIGH | LTV safety: MEDIUM
‘What can I afford?’ | ‘What would improve my options?’ | ‘Should I buy now or wait?’ Run all scenarios | Present full ladder | Education mode: ON
Before the engine can optimize, it needs a complete model. The financial twin is the borrower’s full reality represented as structured data.
Before optimizing, the engine establishes what the borrower qualifies for right now, across all eligible programs. ‘Approvable’ is not the same as ‘optimal’ — the engine keeps searching even after finding a strong baseline.
Program Max Loan Max Price Cash to Close Monthly Pmt Confidence VA (if eligible) Calculate Calculate ~$0 Calculate
USDA (if area eligible) Calculate Calculate Minimal Calculate
Conventional (DU/LP) Calculate Calculate 3-20% Calculate
FHA (3.5% down) Calculate Calculate
Calculate
CRA / State Assist. Calculate Calculate Reduced Calculate
Portfolio / Bank Calculate Calculate Varies Calculate
Non-QM (if needed) Calculate Calculate 10-20% Calculate ALT
This is the engine’s diagnostic layer. Every constraint must be classified before the optimization search begins. The engine assigns a primary blocker and up to three secondary blockers.
A
Borrower can technically qualify, but the monthly payment exceeds their comfort or budget threshold. Engine Action: Compare PITI against stated target payment. Flag delta. Trigger payment reduction branch.
B
Debt-to-income ratio blocks full approval or reduces purchase power below target. Engine Action: Calculate exact DTI gap. Identify which debts, income sources, or loan structures close the gap.
C
Borrower qualifies on income and credit but lacks sufficient liquid funds for down payment and closing costs. Engine Action: Calculate exact cash shortfall. Trigger DPA search, gift fund, seller concession, and asset strategy branches.
D
Score, utilization, derogatory items, or thin file reduce program access or pricing tier. Engine Action: Identify specific score gap to next tier. Calculate fastest improvement path. Model rate savings.
E
Borrower or property falls outside current product guidelines (LTV, property type, occupancy, geography). Engine Action: Test all alternative programs. Identify overlap. Flag exception path if available.
F
Condo rules, multi-unit considerations, census tract, county loan limits, or HOA affect eligibility. Engine Action: Test property variables independently. Model value of property switch or location change.
G
Borrower could qualify with a short sequence of actions, but not today. Engine Action: Generate Action Plan: exact steps, milestones, expected qualification date. Activate monitoring.
H
Borrower is emotionally blocked: anxious, unsure, or lacks information to proceed. Engine Action: Trigger Nova education mode. Present simple scenario ladder. Reframe from ‘No’ to ‘Path.’
Based on the detected blocker, the engine searches the transformation move library in a defined sequence. Each branch below is a prioritized search path.
Branch A: DTI BLOCKER
Branch B: CASH-TO-CLOSE BLOCKER
Branch C: PAYMENT BLOCKER
Branch D: CREDIT BLOCKER
The engine generates a scenario ladder — not a single answer. Every scenario is a specific combination of transformation moves applied to the borrower’s financial twin. The engine evaluates hundreds to thousands of combinations and surfaces the three most relevant.
Scenario Tier Definition Ranking Criteria
Safe / Best Fit High confidence, balanced outcome with no significant borrower actions required. Approval confidence ≥ 85% | Payment within comfort range | Feasible cash to close | No timing dependency
Optimized Strongest balance of borrower goals and deal economics. May require 1-2 borrower actions. Best goal-weight score | Payment comfort met | Maximum purchase power or savings | Highest approval confidence given actions
Stretch / Transformational Materially better outcome. Requires a specific action sequence. Longer timeline. Highest total economic benefit | Clear action path with defined milestones | Borrower capacity to execute | Meaningful upgrade from Option 1
This is what Nova says instead of a single number:
‘You qualify for $450,000 today.’ ‘With two adjustments, you could qualify for $610,000.’ ‘Adjustment 1: Pay off one $3,000 credit card.’ ‘Adjustment 2: Apply for a $7,000 local housing grant that you qualify for.’
That is the moment where the Clarity Engine becomes the most powerful tool a borrower has ever seen.
The 50 core transformation moves across 10 domains. Each move is a discrete engine action. BIX must implement each as a calculable, testable function with defined inputs, outputs, and eligibility conditions.
Debt moves are the most common deal unlockers. They work on both DTI (by removing monthly obligations) and credit score (by reducing utilization).
Pay Off Micro-Debts | Domain: Debt Example: Pay off $2,000 credit card balance Impact: Lower DTI. Improve credit utilization. Potential +15-40 pt FICO improvement. Engine Logic: Calculate exact paydown needed to close DTI gap. Show monthly savings vs. cash deployed. Flag accounts with <10 payments remaining for potential exclusion.
Consolidate Credit Cards Into Mortgage | Domain: Debt Example: Roll $19,000 at 24% APR into mortgage at 6.5% Impact: Lower combined payment. Eliminate high-interest revolving debt. Improve FICO by reducing utilization to 0% on consolidated accounts. Engine Logic: Calculate: (1) new combined payment vs. old total obligation, (2) interest arbitrage savings (24% → 6.5%), (3) FICO impact of utilization elimination, (4) downstream rate improvement from score gain.
Exclude Debts With < 10 Payments Remaining | Domain: Debt Example: Auto loan with 9 payments left at $487/month Impact: Remove obligation from DTI calculation under Fannie Mae guidelines. Immediate qualification improvement without cash outlay. Engine Logic: Flag all installment debts. For each debt with ≤10 payments remaining: calculate DTI impact of exclusion. Present as zero-cost DTI reduction.
Exclude Co-Signed Debt With Third-Party Payment History | Domain: Debt Example: Co-signed student loan, paid by ex-spouse for 14 months Impact: Remove obligation from DTI if 12+ months of documented third-party payment. No cash required. Engine Logic: Prompt borrower for co-signed obligations. Verify payment history. If eligible: remove from DTI. Calculate qualification improvement.
Exclude Business Debt Documented on Business Tax Return | Domain: Debt Example: LLC credit line, paid from business account Impact: Remove personal DTI liability. Requires CPA letter confirming business pays the obligation. Engine Logic: Flag all business-related debts. Prompt for business tax return verification. Calculate DTI improvement and resulting qualification change.
Strategic Debt Payoff Sequencing | Domain: Debt Example: Pay lowest-balance cards first for maximum DTI impact per dollar Impact: Maximize DTI improvement for a fixed cash budget. Prioritize accounts that cross key thresholds. Engine Logic: Sort debts by ‘DTI improvement per dollar of payoff.’ Recommend optimal payoff sequence given available cash.
Credit moves affect both access (program eligibility) and economics (rate pricing tiers). Every point above a tier threshold has compounding value.
Reduce Revolving Utilization Below 30% Per Card | Domain: Credit Example: Pay Discover from $4,200 to $1,260 (30% of $4,200 limit) Impact: Projected +20-40 FICO points. Improves pricing tier. Lowers PMI cost. Engine Logic: For each revolving account: calculate paydown required to reach 30% utilization. Model score projection. Calculate rate/payment improvement at next tier.
Reduce Revolving Utilization Below 10% Per Card | Domain: Credit Example: Pay Chase from $1,800 to $600 (10% of $6,000 limit) Impact: Additional +10-20 FICO points above 30% threshold. Maximum utilization benefit. Engine Logic: Calculate incremental paydown from 30% to 10%. Compare additional score gain against cash cost.
Correct Credit Reporting Errors via Rapid Rescore | Domain: Credit Example: Remove duplicate collection account incorrectly reported twice Impact: Score improvement within 3-7 business days via lender-initiated Rapid Rescore. Immediate qualification benefit. Engine Logic: Flag disputed items with high removal probability. Order Rapid Rescore through lender credit vendor. Project post-correction score and resulting scenario improvement.
Authorized User Strategy | Domain: Credit Example: Add borrower to parent’s 15-year-old Visa with perfect payment history and $500 balance on $10,000 limit Impact: Immediate credit history depth improvement. Projected +15-30 FICO points for thin-file borrowers. Engine Logic: Identify thin-file borrowers (<3 tradelines or <2 years history). Prompt for family accounts with aged, positive history. Model score improvement.
Score-to-Tier Optimization Targeting | Domain: Credit Example: Borrower at 718 needs 720 for Fannie pricing tier. Needs $340 in paydown to cross threshold. Impact: Smallest possible action for maximum pricing impact. Rate reduction at tier crossing = significant lifetime savings. Engine Logic: Calculate exact score needed for each pricing tier. Calculate exact paydown or other action to reach each threshold. Present as: ‘$340 paydown = $87/month savings.’
Program Shift for Credit-Constrained Borrowers | Domain: Credit Example: Conventional (640 min) → FHA (580 min) for borrower at 612 Impact: Access to mortgage financing at lower score threshold. Higher MIP cost — engine must model net benefit. Engine Logic: When conventional not viable, test FHA, VA, USDA, Portfolio, Non-QM in sequence. Calculate total cost comparison including MIP.
Income moves increase the qualifying income denominator in the DTI calculation, directly expanding purchase power or clearing DTI constraints.
Include Bonus Income (2-Year Average) | Domain: Income Example: Annual bonus: $18K (Year 1), $22K (Year 2) = $1,667/month qualifying income added Impact: Expand GMI. Reduce DTI. Increase max loan amount. Engine Logic: Verify 2-year history and employer likelihood-to-continue. Calculate qualifying monthly amount. Model DTI and loan amount impact.
Include Overtime (2-Year Average, Stable) | Domain: Income Example: Overtime: $14,400/year average over 2 years = $1,200/month added Impact: Same as bonus income. Requires stability documentation. Engine Logic: Prompt for overtime frequency and consistency. Verify with VOE. Add to GMI if meets continuity standard.
Self-Employed Income Add-Back Analysis | Domain: Income Example: Schedule C net $52K + $18,400 depreciation + $4,200 home office + $6,100 non-recurring = $80,700 qualifying Impact: 55% income increase from same tax return. Major deal unlocker for small business owners. Engine Logic: Run full Schedule C/S-Corp/Partnership add-back algorithm. Present qualifying vs. reported income with full derivation.
Add Co-Borrower | Domain: Income Example: Spouse adds $72,000/year income and $280/month in new obligations → net DTI improvement Impact: Increase GMI. Must net positive after including co-borrower’s debts. Model combined vs. individual scenarios. Engine Logic: Run both solo and combined scenarios. Flag if co-borrower’s debts offset income gains. Show net DTI impact.
Use Rental Income for Multi-Unit Property | Domain: Income Example: Borrower buying 3-unit: two units rent at $1,400/month each → $2,100 qualifying rental income (75% vacancy factor) Impact: Add $2,100/month to GMI. Major purchase power expansion for multi-unit buyers. Engine Logic: Calculate 75% of gross rental income (Fannie vacancy factor). Require executed leases or appraiser market rent analysis. Model expanded qualification.
Non-Taxable Income Gross-Up | Domain: Income Example: Social Security of $2,400/month (non-taxable) → grossed up to $3,000/month (125%) Impact: 25% increase in qualifying income for non-taxable sources. Standard Fannie/FHA allowance. Engine Logic: Identify non-taxable income sources. Apply 25% gross-up factor. Recalculate GMI and DTI.
Asset Depletion Income | Domain: Income Example: Retiree with $1.2M in retirement accounts: $1.2M ÷ 360 months = $3,333/month qualifying income Impact: Convert illiquid assets to qualifying income. Major unlocker for asset-rich, income-poor borrowers (retirees, early-retired). Engine Logic: Calculate asset depletion income per program guidelines. Determine program with most favorable treatment. Model qualification impact.
Asset moves solve cash-to-close constraints without requiring earned income changes. The engine must search all possible asset sources and combinations.
Down Payment Assistance (DPA) Program Match | Domain: Asset Example: County program: $12,000 forgivable grant after 5 years of owner-occupancy Impact: Eliminate or reduce down payment cash requirement. Expand buyer pool. Preserve liquid reserves. Engine Logic: Maintain searchable DPA database by county and census tract. Auto-match borrower profile (income limits, property limits, first-time buyer status). Calculate exact cash-to-close reduction.
CRA Credit Capture | Domain: Asset Example: Bank’s Community Reinvestment Act program: $7,500 grant for purchase in designated census tract Impact: $5K-$10K grant — most common free money left on the table. Often stacked with DPA. Engine Logic: Flag census tract eligibility. Identify CRA-active lenders in market. Match borrower profile. Calculate stacking with DPA.
Gift Fund Strategy | Domain: Asset Example: Parent gifts $25,000. Conventional at >80% LTV: gift allowed for any amount. FHA: gift allowed for entire down payment. Impact: Eliminate cash-to-close barrier with family support. Fully documented, fully compliant. Engine Logic: Identify gift eligibility by program and LTV. Generate gift letter template. Document trail requirements: donor statement + transfer record.
401(k) as Reserve Documentation | Domain: Asset Example: Borrower has $85,000 in 401(k). 60% = $51,000 counted as liquid reserve without withdrawal required. Impact: Vault reserves requirement. No tax event. No income impact. Engine Logic: Calculate 60% of vested 401(k) balance. Compare against reserve requirement. Flag as reserve-eligible even without withdrawal.
Seller Concession Optimization | Domain: Asset Example: Seller agrees to $14,000 closing cost credit in lieu of price reduction. Borrower preserves $14,000 cash. Impact: Convert seller discount into borrower cash preservation. Same economics to seller; major liquidity improvement for buyer. Engine Logic: Calculate maximum allowable seller concession by program and LTV (3% at >90% LTV, 6% at 75-90%, 9% at <75%). Model cost credit vs. rate buydown comparison.
Seller-Funded 2-1 Buydown | Domain: Asset Example: Seller funds $9,900 to buy down borrower’s rate: Year 1 at 5.25% ($2,484/mo), Year 2 at 6.25%, Year 3+ at 7.25% Impact: Dramatically lower initial payment. Seller pays same effective economics. Borrower gains 2 years of payment relief. Engine Logic: Calculate buydown cost (typically ~2-2.5% of loan). Compare seller concession vs. price reduction economics. Present Year 1-3 payment schedule.
Program switches are one of the largest deal unlockers. Different programs have fundamentally different eligibility rules, down payment requirements, and pricing structures.
Conventional to VA (If Eligible) | Domain: Program Example: Borrower with 6 years active duty: 0% down, no PMI, competitive rate, no funding fee if disabled Impact: Eliminates down payment and PMI. Saves $200-400/month on most deals. Best loan program in existence for eligible borrowers. Engine Logic: Check DD-214, service length, discharge type, remaining entitlement. Calculate funding fee or exemption. Compare all-in payment vs. conventional.
Conventional to FHA | Domain: Program Example: Borrower at 612 FICO qualifies for FHA at 3.5% down but not conventional (minimum 620+) Impact: Access to financing at lower credit threshold. Higher MIP offset by access itself. Engine Logic: Calculate conventional vs. FHA MIP breakeven. If borrower cannot access conventional: FHA unlocks the deal. Model total cost of ownership.
FHA to Conventional | Domain: Program Example: Borrower at 740+ FICO: conventional PMI is cheaper than FHA MIP. FHA MIP is life-of-loan at >90% LTV. Impact: Lower total cost. PMI cancels at 78% LTV. FHA MIP is permanent for most borrowers. Engine Logic: Compare monthly FHA MIP vs. conventional PMI. Calculate time to PMI cancellation. Model 7-year total interest + MIP/PMI cost. Conventional wins for high-score borrowers.
Standard QM to Bank Statement (Non-QM) | Domain: Program Example: Self-employed borrower: tax return shows $52K net, but bank deposits show $140K/year. Bank statement qualifying income: $140K. Impact: Unlock true income for self-employed. Major purchase power expansion. Engine Logic: When tax return income is insufficient: test 12-month and 24-month bank statement programs. Calculate qualifying income from deposits. Compare against conventional add-back result.
Switch to USDA Rural Development | Domain: Program Example: Borrower buying in eligible rural area: 0% down, competitive rate, income limits apply Impact: Eliminate down payment. No monthly PMI (only upfront guarantee fee). Strong alternative to FHA for rural properties. Engine Logic: Check USDA property eligibility map. Verify household income against county limits. Compare total cost vs. FHA and conventional.
Structure moves change how the loan is configured — rate, term, payment schedule — without changing the borrower’s financial profile.
Permanent Discount Point Buydown | Domain: Structure Example: 1 point ($4,500) reduces rate from 7.25% to 7.00%. Saves $74/month. Breakeven: 61 months. Impact: Lower rate, lower payment, lower lifetime interest. Works when borrower plans to stay past breakeven. Engine Logic: Calculate breakeven for each 0.25% rate increment. Present: cost, monthly savings, breakeven months. Flag if breakeven exceeds expected tenure.
Temporary 2-1 Buydown (Seller-Funded) | Domain: Structure Example: Seller funds buydown: Year 1 rate = 5.25%, Year 2 = 6.25%, Year 3+ = 7.25%. Cost: $9,900 on $450K loan. Impact: Year 1 payment relief of $587/month. Seller-funded = no borrower cash out. Borrower qualifies at note rate. Engine Logic: Calculate full buydown cost. Model Year 1-3 payment schedule. Present to seller as equivalent to price reduction with better buyer benefit.
ARM: 5/1 or 7/1 Adjustable Rate Mortgage | Domain: Structure Example: 5/1 ARM at 6.25% vs. 30-yr fixed at 7.25%. Saves $296/month for first 5 years. Impact: Lower initial payment. Qualify at initial rate. Appropriate for borrowers with defined exit timeline. Engine Logic: Model ARM vs. fixed payment delta. Calculate total savings over expected tenure. Flag reset risk at adjustment date. Present hybrid fix scenario.
Term Reduction to Same Payment | Domain: Structure Example: Rate drops from 7.25% to 6.875%. Instead of lower payment, keep $2,528 and pay off 4.3 years early. Impact: Same cash flow. Earlier payoff. $89,400 lifetime interest savings. Engine Logic: Solve for n (term) given fixed PMT and new rate. Present payoff year. Model total interest savings.
Payment Neutral Cash-Out | Domain: Structure Example: Keep same $2,528 payment at new rate. Maximum loan size at that payment = $421,649. Cash available: $30,249. Impact: Access liquidity with zero payment increase. Powerful for home improvement, debt payoff, or investment. Engine Logic: Solve for PV (max loan) given fixed PMT and new rate. Calculate cash-out amount. Flag LTV and PMI threshold.
Switching from condo to SFR when condo docs fail warrantability review Impact: Eliminate condo approval risk. SFR qualifies under standard guidelines.
Purchasing 2-4 unit property and using rental income to qualify Impact: Expand purchase power via rental income offset. FHA allows 3.5% down on 2-4 unit owner-occupied.
Targeting lower price point to close cash-to-close or DTI gap Impact: Right-size purchase to financial reality now. Return for upgrade when optimization path is executed.
Switching county to access higher conforming loan limits Impact: 2025 limits: $806,500 base; up to $1,209,750 in high-cost counties. Can change product eligibility significantly.
Targeting census tracts with CRA incentive overlays Impact: Unlock lender-specific CRA pricing benefits ($5K-$10K grants, reduced rates). Often not advertised.
Most loan officers don’t search for DPA, CRA, or grant programs because their systems don’t surface them. The Clarity Engine must maintain a real-time, searchable incentive library indexed by county, census tract, income, first-time buyer status, and property type. This alone is worth tens of millions of dollars in unlocked purchase power for borrowers who would otherwise not buy.
State and county programs. Often $5K-$25K. Some forgivable. Engine must auto-match borrower eligibility by income, geography, and buyer status.
Community Reinvestment Act obligations make banks offer $5K-$10K grants in designated tracts. Not advertised. Requires lender-specific knowledge.
Local housing agencies. Sometimes stackable with DPA. Income-restricted. Engine must maintain real-time database.
Some large employers (hospitals, universities, tech companies) fund down payment assistance. Flag employer for EAH program lookup.
Targeted conventional programs with reduced pricing for eligible borrowers in underserved markets.
Seller pays buyer’s closing costs. Max: 3% (LTV>90%), 6% (75-90%), 9% (<75%). Preserves buyer cash.
Seller funds permanent or temporary rate buydown. Lower buyer payment. Same seller economics as price cut.
Increase price, increase concession proportionally. Buyer nets same price but converts cash to financed concession.
Same payment, lower rate = shorter term. $128,570 interest savings by keeping $2,528/month at 5.75%.
Convert all high-interest debt to single mortgage rate. Eliminate $570/month in CC minimums. +40-60 pt FICO boost.
Refinance into lower first mortgage. Open HELOC for flexible liquidity. Best of both: low fixed rate + accessible equity.
Nova is the Clarity Engine’s AI loan officer interface. The engine’s math only matters if the borrower understands it. Nova translates optimization output into clear, confident, human-sounding recommendations. These are the canonical conversation templates.
Nova never delivers a single number without a path to a better number. Nova never says ‘you don’t qualify’ — only ‘here is your path.’ Nova always presents: (1) what you qualify for today, (2) what becomes possible with specific actions, (3) exactly what those actions are. Every borrower leaves every conversation knowing their current state AND their next move.
Borrower Says “How much house can I afford?”
Nova Response Logic
Nova Output Example “Right now, based on your income and current debt profile, you’re positioned for a home around $485,000.
There are two adjustments that could move that closer to $565,000 within 60 days: 1. Paying off one revolving credit card (approximately $3,200) drops your debt ratio by 3.1 points and improves your pricing tier. 2. A local assistance program you appear to qualify for can reduce your cash needed to close by $9,500.
Would you like to see both paths — your options today, and what becomes available after those two moves?”
Borrower Says “The payment is more than I want to pay.”
Nova Response Logic
Nova Output Example “I understand. Let me show you exactly where the payment is coming from and three ways we can bring it down:
Option 1: A seller-funded rate buydown reduces your first-year payment from $2,847 to $2,260 — saving $587/month for the first 12 months. This is something you can negotiate into the purchase offer. Option 2: A different loan structure at 5.75% (versus the 6.25% you were quoted) reduces the payment by $182/month permanently. This requires a 22-point score improvement, which we can walk through. Option 3: A local housing grant reduces the loan amount by $8,500, dropping the payment by $54/month with no action required beyond an application.
Which of those directions feels right for your situation?”
Borrower Says “I was told I don’t qualify. My credit isn’t good enough.”
Nova Response Logic
Nova Output Example “Your FICO of 618 puts you below the conventional threshold, but you’re 22 points from FHA eligibility and 62 points from the pricing tier that saves $187/month over the life of your loan.
Here’s your 90-day path: 1. Pay your Discover card from $4,200 to $1,260. That’s a reduction to 30% utilization. Based on your profile, that single move projects a +28-point improvement. 2. Dispute the duplicate collection account showing on your Equifax report. Our analysis shows this is a data error. Rapid Rescore can have it corrected in 5-7 business days, projected +12 points.
Combined: you likely cross 640 within 45 days. At that score, I have your pre-approval ready within 24 hours. I’ll monitor your score and alert you the moment you cross the threshold.”
Borrower Says “I served 6 years in the Army. I’ve never used a VA loan.”
Nova Output Example “Your VA entitlement is one of the most powerful financial benefits you’ve earned — and it never expires.
With your 6 years of active duty service and honorable discharge, you qualify for a VA loan:
On a $400,000 purchase, that’s approximately $287/month less than a conventional loan with 5% down. Over 30 years: $103,320 in savings.
Let’s request your Certificate of Eligibility right now. It takes 3 minutes.”
Borrower Says “I want to pull cash out of my house to pay off my credit cards.”
Nova Response Logic
Nova Output Example “That’s a smart impulse. Converting 24% credit card debt to 6.5% mortgage debt is a strong financial move. Let me show you three ways to do it:
Option 1 — Cash-Out Refinance: Replace your entire mortgage at current rates. New loan: $421,649. Same payment: $2,528. Cash in hand: $30,249. Pro: one loan, clean structure. Con: restarts your clock at today’s rate on the full balance. Option 2 — HELOC: Keep your current first mortgage untouched. Open a line of credit against your equity. Pro: preserves your existing low rate on the first. Con: HELOC rate is variable and currently higher. Option 3 — Debt Consolidation Refi: Roll your $19,000 in credit cards directly into a new mortgage at 5.75%. New payment: $2,385. Your total monthly obligation drops from $3,098 to $2,385 — a savings of $713/month. Plus your credit score is projected to improve 40-60 points as the revolving accounts go to zero.
Which matters more right now: preserving your current rate, or maximizing your monthly cash flow improvement?”
The following is the structured build checklist organized by sprint priority. BIX should use this as the engineering task backlog.
# Deliverable Type Priority 1 PMT / PV / NPER / RATE / FV calculation library Core Math P0 2 GMI engine: all 8 income types with documentation flags Income P0 3 Schedule C/S-Corp/Partnership add-back algorithm Income P0 4 Front-end and back-end DTI calculator with all liability types DTI P0 5 PITI calculator (P&I + tax + insurance + HOA + MIP/PMI) DTI P0 6 LTV/CLTV/HCLTV engine with PMI threshold logic LTV P0 7 Credit score tier classification with rate adjustment matrix Credit P0 8 5-scenario parallel computation framework Scenario P0 9 Loan product waterfall (VA > USDA > Conv > FHA > Non-QM) Program P0 10 Conventional vs. FHA MIP breakeven calculator Program P0 11 Debt optimization: payoff sequencing + exclusion logic Debt P1 12 Rapid Rescore simulation: score improvement + payment delta Credit P1 13 Rate buydown / point analysis with breakeven months Structure P1
# Deliverable Type Priority 14 Blocker detection: classify primary + secondary constraints Decision Tree P0 15 Optimization branch routing: DTI / Cash / Payment / Credit / Program / Property Decision Tree P0 16 DPA / grant / CRA program database with geo-indexed search Incentive P0 17 Asset combination engine: DPA + seller concession + gift + 401(k) stacking Asset P1 18 Seller concession optimization by program and LTV Seller P1 19 2-1 buydown calculator with seller-funded cost estimate Structure P1 20 ARM vs. fixed analysis with reset risk modeling Structure P1 21 Co-borrower scenario: combined vs. solo qualification comparison Income P1 22 Asset depletion income calculator Income P2 23 Non-QM / bank statement income qualification module Program P2
# Deliverable Type Priority 24 URLA 1003 field ingestion into structured borrower object Intake P0 25 Always Approved state machine (3 states + transition logic) State Engine P0 26 Action Plan generator: steps, dates, milestones, expected outcome Path Engine P0 27 Scenario ranking engine: weight by borrower intent profile Ranking P0 28 Nova conversation template engine: 5 core paths Nova P0 29
Triggers P1 30 Refinance opportunity monitor: rate / equity / score / life event triggers Monitor P1 31 Fair lending disparate impact testing layer Compliance P0 32 MISMO 3.6 field mapping for lender API integration Integration P1 33 Data partner API connectors: Attom, Clear Capital, Revaluate Integration P1
The Clarity Engine should never behave like a product eligibility checker. It should behave like a financial outcome optimizer. That is the science. That is the moat. That is what BIX needs to build.
Document Version: 1.0 | Sources: Mortgage Optimization Map + Clarity Engine Decision Tree | Prepared for: BIX Technology Corp | Classification: CONFIDENTIAL — PreFi, Inc. / Purpose Technology, Inc. (d/b/a Purlend)