AI Underwriting

Risk Scoring for Real Estate Loans

Automated LTV, LTC, DSCR, and ARV analysis. AI calculates every metric your underwriters need and assigns risk scores based on your historical loan performance.

Instant
LTV/LTC/ARV Calculation
15+ Factors
Risk Assessment
90%+
Prediction Accuracy

Metrics That Matter

Every calculation private lenders need, automated

LTV (Loan-to-Value)

Calculates loan amount as percentage of property value. Flags deals above 75% LTV as high risk for fix & flip loans.

  • Automatic calculation from appraisal
  • Risk thresholds: >75% high, >70% medium
  • Compares to your historical norms
  • ARV-based LTV for flips

LTC (Loan-to-Cost)

Measures loan amount against total project cost (purchase + rehab). Critical for fix & flip deals with renovation budgets.

  • Includes purchase + rehab costs
  • Validates contractor bids
  • Flags >90% LTC as risky
  • Tracks cost overruns

ARV Ratio Analysis

Compares After Repair Value to purchase price. Identifies if flip has sufficient profit margin to justify the risk.

  • ARV ÷ Purchase Price ratio
  • >200% = too optimistic flag
  • <120% = minimal profit warning
  • Historical ARV accuracy tracking

DSCR (Debt Service Coverage)

For rental loans, calculates if monthly rent covers mortgage payment (PITI). DSCR <1.0 = negative cash flow, auto-decline.

  • Monthly rent ÷ PITI calculation
  • <1.0 = decline (negative cash flow)
  • 1.0-1.25 = conditional approval
  • >1.4 = strong cash flow

Borrower Experience Score

Weighs years in real estate investing, number of past flips completed, and credit score to assess borrower capability.

  • Years of investing experience
  • Number of completed projects
  • Credit score factor
  • 0-100 experience rating

Liquidity Assessment

Reviews bank balances to verify borrower has reserves for unexpected costs. Flags insufficient liquidity as major risk.

  • Liquid assets vs loan size
  • <10% of loan = high risk
  • <15% = medium risk flag
  • >25% = strong reserves

Property Market Risk

Analyzes location factors: days on market in area, comparable sales velocity, neighborhood appreciation trends.

  • Average days on market
  • Price trend analysis
  • Comp sales velocity
  • Neighborhood risk score

Composite Risk Score

Combines all factors into single 0-100 risk score. Machine learning weighs each metric based on your historical loan outcomes.

  • 0-30 = Low risk (auto-approve)
  • 31-70 = Medium risk (review)
  • 71-100 = High risk (decline)
  • Trained on your loan history

Risk Assessment Logic

How AI decides approve, conditional, or decline

Auto-Decline

Any ONE of these triggers immediate decline:

  • LTV >75%
  • LTC >95%
  • DSCR <1.0 (negative cash flow)
  • Credit score <580
  • Liquidity <10% of loan amount
  • ARV ratio >200% (too optimistic)

Conditional Approval

Two or more medium-risk flags:

  • LTV 70-75%
  • LTC 90-95%
  • DSCR 1.0-1.25
  • Experience score <50
  • Liquidity 10-15%
  • First-time flipper

Requires manual underwriter review with additional conditions (higher rate, more reserves, personal guarantee, etc.)

Auto-Approve

All of these must be true:

  • LTV ≤70%
  • LTC ≤90%
  • DSCR ≥1.25 (for rentals)
  • Credit score ≥650
  • Experience score ≥60
  • Liquidity ≥15%

Straight to closing, no underwriter review needed

Why Risk Scoring Matters

Better decisions, faster approvals, lower defaults

01

Consistent Decisions

Every loan judged by same criteria. No more "John approves looser than Sarah." Same LTV gets same risk score every time, regardless of who underwrites.

02

Spot Red Flags Instantly

AI highlights problems underwriters might miss—borrower has liquidity but DSCR is negative, ARV seems inflated compared to comps, LTC budget doesn't include permit costs.

03

Learn From History

ML models train on your past loans. If deals with >72% LTV historically defaulted more, AI weighs that factor heavier. Gets smarter with every funded loan.

04

Justify Pricing

Risk scores tie directly to pricing tiers. Show borrowers: "70% LTV = 9.5% rate. Your deal is 73% LTV = 10.5% rate." Data-driven, defensible pricing.

Risk Scoring – FAQ

How credit teams use the risk engine day to day

What is Mentyx Risk Scoring for real estate loans?

It is an AI underwriting layer that calculates all key metrics your team cares about—LTV, LTC, DSCR, ARV ratios, borrower experience, liquidity, and market risk—and rolls them into a single composite risk score. The engine then proposes approve, conditional, or decline recommendations based on your credit policy.

Can we customize the thresholds and override decisions?

Yes. All rules are configurable: you define your own LTV, LTC, DSCR, liquidity, and experience cutoffs. Underwriters can always override an AI recommendation with a reason code, and those overrides feed back into the model so it aligns more and more with how your team actually makes decisions.

Does this replace underwriters or support them?

It supports them. Mentyx Risk Scoring takes care of the calculations, flags, and consistency, while your underwriters focus on edge cases, structure, and borrower relationships. The result: fewer missed red flags, faster clear approvals, and a clear audit trail of how every decision was made.

See Risk Scoring in Action

Watch AI calculate LTV, LTC, DSCR, and assign risk scores in real-time