Predictika · conventional loan eligibility and credit fees: limits, LTV, underwriting and pricing

An AI sales agent's loan quote, checked before it goes out

An AI sales agent prepares a quote for a conventional mortgage for a fictional borrower, with a credit fee in dollars. Before the rate lock goes out, the agent sends it to Predictika’s validator, which checks it against eligibility, underwriting and credit-fee rules, compares the credit fee the agent states with the one it computes, and returns a verdict with every reason.

Scope: conventional conforming and high-balance mortgage quotes in the style of Freddie Mac: eligibility, underwriting limits and credit fees. See everything this validator covers

This validator models conventional conforming and high-balance purchase and refinance quotes, for three sample counties, and the Freddie Mac-style credit fee grid. It handles any combination of purpose, occupancy, property, program and underwriting choices the model supports, not just the scenarios shown below.

  • Eligibility — Maximum LTV by purpose, occupancy, unit count and property type, the Home Possible and HomeOne rules, cash-out seasoning, second-home and unit-count limits, and the county conforming and high-balance limits.
  • Underwriting — Credit score, DTI under automated and manual underwriting, reserves and mortgage insurance coverage, with an approval step where a person must confirm.
  • Credit fees — The score-by-LTV fee grid, attribute fees for occupancy, units, condo, manufactured home, subordinate financing, ARM and high balance, the dollar total, and a check that the credit fee the agent states matches the validator fee.

It does not cover FHA, VA or USDA loans, loans above the high-balance limit, other lenders’ or agencies’ rules, counties beyond the three modeled, real Freddie Mac fees or eligibility decisions, or any real borrower.

The AI sales agentBuilds the quote from the conversation and says it is ready.
Predictika validatorChecks the quote against the product and contract rules and returns every finding it has.

9 scenarios in 5 groups · each group runs from the agent’s first attempt to the outcome · click one to switch

What the agent proposes

The AI sales agent says

Illustrative agent statement. The validator receives the structured request below.

Proposing agent: AI sales agent (scripted)

Before sending, the agent must call Recorded run

Predictika validator


      

        
        
        
      

      
Waiting for the call.

      
    

  

  
    

    
Where rules-based CPQ gets it wrong
Click to expand
    
Our assessment of what a typical rules-based CPQ (product rules, price rules and validation rules, as in Salesforce CPQ or Revenue Cloud) does with the same cases. It is a judgment about effort and reliability, not a claim that nothing can be built.

    
What the validator catchesTypical rules-based CPQPredictika
Eligibility that depends on several choices togetherOnly with workarounds
Maximum LTV is a matrix of purpose, occupancy, units, property type and high balance, and the over-95 percent exception adds a second set of conditions; each combination tends to become a separate rule.		All conditions are evaluated together on every call, whatever order the choices arrive in.
County loan limitsOnly with care
Limits are a lookup by county and unit count that must be kept complete, and the baseline and high-balance tiers feed both eligibility and pricing.		The county row supplies the limit, and the same value drives the check and the high-balance fee.
The credit fee gridOnly with care
A fee table by score band and LTV bucket, with waivers for some programs, is manageable in a price rule but grows with every new row.		The score band row is selected by rule and the grid fee is computed once for every check to read.
Attribute fees added up in points and dollarsOnly with workarounds
Seven optional adjustments, each with its own waiver conditions, must be summed and converted to dollars in a defined order.		Each fee is computed once; the points, dollars and line items all come from the same figures.
Seasoning, DTI and approval stepsOnly with care
A cash-out ownership period, two DTI limits by underwriting type and an approval step need separate rules and a place to record the approval.		All thresholds are evaluated together and returned with the reason.
The fee the agent statesOnly with workarounds
Comparing a stated fee with a computed one means a separate rule per quote layout, and price rules rarely run on stated figures.		The stated credit fee is a field in the request and is checked against the computed fee.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as the quote is saved, so a rep often fixes one error and meets the next.		All findings come back together, whatever order the choices arrived in.

  

  

    
What this validator catches
Click to expand
    
Every rule is checked on every call, in whatever order the choices arrive, and one problem never hides another. Blocking findings stop the quote; approval findings need a person to sign off; everything else is computed.

    
RuleEffectWhat it checks
1computedThe county sets the conforming loan limit; loans above the baseline limit are high balance
2blockMaximum LTV depends on purpose, occupancy, unit count, property type and high-balance status
3blockLTV above 95 percent needs a fixed-rate 1-unit principal residence in a HomeOne first-time buyer purchase, a Home Possible purchase, or a no-cash-out refinance of an existing Freddie Mac loan
4blockHome Possible needs a principal residence and income at or below 80 percent of the area median
5blockThe loan amount must not exceed the county limit for the unit count
6blockA cash-out refinance needs at least 12 months of ownership
7blockA second home must be a 1-unit property
8blockCredit score must meet the minimum for the program
9approval, then blockDTI above 45 percent needs a person to confirm the automated underwriting finding; above 50 percent is blocked, and manually underwritten loans are blocked above 45
10needs approvalA manually underwritten loan needs an underwriter to sign off
11blockReserves in months must meet the minimum for occupancy and unit count
12blockMortgage insurance coverage must meet the requirement for the LTV
13computedThe credit fee grid is chosen by credit score band and LTV bucket, and is waived for Home Possible
14computedAttribute fees are added for occupancy, unit count, condo, manufactured home, subordinate financing, ARM and high balance
15computedThe credit fee in dollars is the total points times the loan amount
16blockThe credit fee the agent states must match the validator fee within $1

  

  

    
Scenarios and results
Click to expand
    
The recorded result of the validator for each scenario on this page: the request the AI agent sent, the verdict, and every finding, in the order they appear above.

    
1. LTV above 95 percent: Standard loan at 97 percent

    
The agent's request. A 30-year fixed conventional purchase loan of $350,000 at 97 percent LTV, single-family primary residence in Ada County, Idaho. Credit score 700, DTI 38 percent, 35 percent mortgage insurance coverage, standard program, not a first-time buyer. Credit fee $3,062.50. Ready to lock.

    
Verdict: Blocked.

    
    
      
  • LTV above 95 percent not allowed for this loan. LTV above 95 percent requires a fixed-rate 1-unit principal residence - a HomeOne purchase with a first-time homebuyer or a Home Possible purchase, or a no-cash-out refinance of an existing Freddie Mac loan.
    • 
    

    
2. LTV above 95 percent: First-time buyer, HomeOne

    
The agent's request. A 30-year fixed conventional purchase loan of $350,000 at 97 percent LTV, single-family primary residence in Ada County, Idaho. Credit score 700, DTI 38 percent, 35 percent mortgage insurance coverage, first-time homebuyer with income at 150 percent of the area median. Credit fee $3,062.50. Ready to lock.

    
Verdict: Allowed.

    
3. Loan size against the county limit: Loan above the county limit

    
The agent's request. A 30-year fixed conventional purchase loan of $900,000 at 80 percent LTV, single-family primary residence in Denver County, Colorado. Credit score 780, DTI 38 percent. Credit fee $12,375.00. Ready to lock.

    
Verdict: Blocked.

    
    
      
  • Loan amount above the county conforming limit. Loan amount $900000 exceeds the 1-unit conforming limit of $862500 in Denver County CO; this is a jumbo loan.
    • 
    

    
4. Loan size against the county limit: Loan reduced to $850,000

    
The agent's request. A 30-year fixed conventional purchase loan of $850,000 at 80 percent LTV, single-family primary residence in Denver County, Colorado. Credit score 780, DTI 38 percent. Credit fee $11,687.50. Ready to lock.

    
Verdict: Allowed.

    
5. Cash-out seasoning: Cash-out after 8 months

    
The agent's request. A 30-year fixed conventional cash-out refinance of $300,000 at 75 percent LTV, single-family primary residence in Dallas County, Texas. Credit score 700, DTI 38 percent, property owned 8 months. Credit fee $7,875.00. Ready to lock.

    
Verdict: Blocked.

    
    
      
  • Cash-out refinance before 12 months of ownership. A cash-out refinance needs at least 12 months of ownership; the property has been owned 8 months.
    • 
    

    
6. Cash-out seasoning: Cash-out after 14 months

    
The agent's request. A 30-year fixed conventional cash-out refinance of $300,000 at 75 percent LTV, single-family primary residence in Dallas County, Texas. Credit score 700, DTI 38 percent, property owned 14 months. Credit fee $7,875.00. Ready to lock.

    
Verdict: Allowed.

    
7. The credit fee the agent states: Stated fee too low

    
The agent's request. A 30-year fixed conventional purchase loan of $400,000 at 80 percent LTV, single-family primary residence in Dallas County, Texas. Credit score 760, DTI 38 percent. Credit fee $2,000.00. Ready to lock.

    
Verdict: Blocked.

    
    
      
  • Stated credit fee differs from the validator fee. The credit fee the agent states ($2000.0) differs from the credit fee the validator computes ($2500.0).
    • 
    

    
8. The credit fee the agent states: Stated fee corrected

    
The agent's request. A 30-year fixed conventional purchase loan of $400,000 at 80 percent LTV, single-family primary residence in Dallas County, Texas. Credit score 760, DTI 38 percent. Credit fee $2,500.00. Ready to lock.

    
Verdict: Allowed.

    
9. Approval and release: DTI of 48 percent, approved

    
The agent's request. A 30-year fixed conventional purchase loan of $300,000 at 75 percent LTV, single-family primary residence in Dallas County, Texas. Credit score 740, DTI 48 percent, automated underwriting finding LPA. Approval has been granted. Credit fee $1,125.00. Ready to lock.

    
Verdict: Needs approval.

    
    
      
  • DTI above 45 needs a person to confirm. DTI 48 percent is above 45; confirm the LPA Accept finding supports it.
    • 
    

  

  

    
Narration transcript
Click to expand
    
The text of the narrated walkthrough, step by step.

    
    
      
  1. Welcome to a demo of Predictika's Validator. An AI sales agent prepares a quote for a conventional mortgage for a fictional borrower, with a credit fee in dollars. Before the rate lock goes out, the agent sends it to Predictika’s validator, which checks it against eligibility, underwriting and credit-fee rules, compares the credit fee the agent states with the one it computes, and returns a verdict with every reason.
    2. 
      
  2. I will start with the first request by the agent.
    3. 
      
  3. The agent says. A 30-year fixed conventional purchase loan of $350,000 at 97 percent LTV, single-family primary residence in Ada County, Idaho. Credit score 700, DTI 38 percent, 35 percent mortgage insurance coverage, standard program, not a first-time buyer. The rest of the request is on screen. Credit fee $3,062.50.
    4. 
      
  4. Before it goes out, the agent calls the validator.
    5. 
      
  5. Blocked, with one finding, all in one answer. LTV above 95 percent requires a fixed-rate 1-unit principal residence - a HomeOne purchase with a first-time homebuyer or a Home Possible purchase, or a no-cash-out refinance of an existing Freddie Mac loan.
    6. 
      
  6. The validator returned one blocking finding in one answer. A check that reads one field at a time would pass most of this; it shows only when the choices in the quote are checked against each other.
    7. 
      
  7. What this shows: for county loan limits, a rules-based CPQ is error-prone for this.
    8. 
      
  8. Now the agent resubmits.
    9. 
      
  9. The corrected request. It sets the FTHB to yes and the Income Pct AMI to 150 percent.
    10. 
      
  10. It calls the validator again.
    11. 
      
  11. This time the validator clears it.
    12. 
      
  12. The validator also returns the line-item quote, on screen.
    13. 
      
  13. The earlier finding is cleared, and the validator's total, $3,062.50, matches the quote.
    14. 
      
  14. One problem caught in one call, fixed in one pass, and a clean quote. That is the value of Predictika's validator: catching mistakes before they reach the customer.
    15.