Automated underwriting systems have become the central engine of the modern conforming mortgage market, transforming a process that once required days of manual review by human underwriters into a computational evaluation that can be completed in a matter of seconds. The two dominant systems in the conventional conforming market are Fannie Mae Desktop Underwriter, commonly referred to as DU, and Freddie Mac Loan Product Advisor, commonly referred to as LPA, both of which serve as the primary automated decisioning platforms for loans that are intended for sale to the government-sponsored enterprises. These systems ingest a standardized set of borrower and loan data, apply proprietary algorithms and risk models to that data, and produce a decision that tells the lender whether the loan meets the GSE eligibility requirements and what level of representation and warranty relief the lender can expect if the loan is sold to the GSE. Understanding how these systems work, what data they evaluate, and how their decisions should be interpreted is essential for mortgage professionals who originate conforming loans and for borrowers who want to understand the technology that determines whether their loan application is approved.
The history of automated underwriting systems in the mortgage industry traces back to the mid-1990s, when Fannie Mae introduced Desktop Underwriter as a tool to streamline the underwriting process and bring greater consistency and speed to the evaluation of conforming loan applications. Prior to the introduction of automated underwriting, every conforming loan was reviewed by a human underwriter who manually evaluated the borrower credit profile, income documentation, asset verification, and property appraisal against the GSEs published underwriting guidelines. The manual process was time-consuming, subject to interpretation differences among underwriters, and created a bottleneck in the loan origination pipeline that limited the volume of loans that could be processed within a given timeframe. The introduction of DU and, shortly thereafter, the introduction of Freddie Mac Loan Product Advisor, represented a fundamental shift in the mortgage industry, as the automated systems could evaluate a loan application in seconds and produce a decision that was consistent, objective, and based on the vast historical loan performance data that the GSEs had accumulated over decades of mortgage lending.
The data inputs that automated underwriting systems require to generate a decision are standardized across the conforming market, and they reflect the core components of mortgage underwriting that have always been central to the loan evaluation process. The primary data inputs include the borrower credit report, which provides the credit scores, credit account histories, and public record information that the AUS uses to assess the borrower credit profile; the loan application data, which includes the loan amount, loan term, loan type, property type, occupancy status, and loan purpose; the borrower income and employment data, which includes the monthly qualifying income, employment history, and income source type; the borrower asset data, which includes the funds available for down payment and closing costs and the reserves that the borrower holds; and the property data, which includes the estimated property value, the property type, and the appraisal information if an appraisal has been obtained. The AUS combines these inputs with its proprietary risk models to generate a decision that reflects the comprehensive risk profile of the loan.
The credit score is one of the most important inputs in the automated underwriting process, and both DU and LPA use the credit score as a primary driver of the risk assessment and the eligibility determination. The credit score that the AUS evaluates is typically the FICO score, which is a credit scoring model developed by the Fair Isaac Corporation that summarizes the borrower credit history into a three-digit score ranging from 300 to 850. The GSEs have established minimum credit score requirements for conforming loans, with the baseline minimum typically set at 620 for conventional conforming loans, though higher minimum scores may be required for certain loan types, higher loan-to-value ratios, or specific property types. The credit score influences not only the eligibility decision but also the pricing of the loan, as borrowers with higher credit scores typically receive more favorable interest rates and lower loan-level price adjustments, which are risk-based pricing adjustments that the GSEs apply to loans with higher risk characteristics.
The debt-to-income ratio is another critical input that the automated underwriting systems evaluate, and the DTI ratio serves as a key measure of the borrower capacity to repay the loan. The DTI ratio is calculated by dividing the borrower total monthly debt obligations, including the proposed mortgage payment of principal, interest, taxes, insurance, and any homeowner association dues, plus all other recurring monthly debt obligations such as auto loans, student loans, and credit card payments, by the borrower gross monthly qualifying income. The GSEs have established a maximum DTI ratio of 50 percent for conforming loans, though the automated underwriting systems may approve loans with DTI ratios above 45 percent only when compensating factors are present, such as strong reserves, excellent credit scores, or significant liquid assets. The AUS evaluates the DTI ratio in conjunction with other risk factors, and a high DTI ratio may result in a refer or ineligible decision if the overall risk profile does not meet the system risk thresholds.
The loan-to-value ratio is a fundamental input that the automated underwriting systems use to assess the collateral risk and the borrower equity position in the property. The LTV ratio is calculated by dividing the loan amount by the property value or the appraised value, and it represents the percentage of the property value that is financed by the loan. The GSEs allow LTV ratios up to 97 percent for certain conventional conforming loan programs, though loans with LTV ratios above 80 percent typically require private mortgage insurance, which protects the lender and the GSE against the risk of default on high-LTV loans. The AUS evaluates the LTV ratio in conjunction with the credit score, the DTI ratio, and other risk factors, and higher LTV ratios generally require stronger credit profiles and greater reserves to obtain an approve decision from the automated system.
The risk models that power Desktop Underwriter and Loan Product Advisor are proprietary algorithms that have been developed and refined by Fannie Mae and Freddie Mac over many years, using the vast loan performance databases that the GSEs maintain. These models analyze the relationships between borrower characteristics, loan characteristics, and property characteristics and the historical performance of loans with similar profiles, identifying the combinations of factors that are associated with higher or lower default risk. The risk models are continuously updated and refined as new loan performance data becomes available, and the GSEs periodically release new versions of their AUS platforms that incorporate updated models and revised eligibility criteria. The proprietary nature of the risk models means that the exact algorithms and weighting of factors are not publicly disclosed, but the GSEs publish their underwriting guidelines and eligibility requirements, which provide the framework within which the automated systems operate.
The decision outputs that DU and LPA produce are the primary results that lenders use to determine whether to proceed with a loan application, and the decisions are expressed in a standardized format that communicates the system recommendation and the conditions that must be met for the loan to be eligible for sale to the GSE. The primary decision outputs include Approve or Eligible, which indicates that the loan meets the AUS risk thresholds and is eligible for purchase by the GSE subject to the satisfaction of specified conditions; Refer or Ineligible, which indicates that the loan does not meet the AUS risk thresholds and requires manual underwriting review or does not meet the GSE eligibility requirements; and Refer with Caution, which is a DU-specific output that indicates the loan is borderline and requires human underwriter review to determine whether the loan can be approved. Each decision output is accompanied by a list of findings that specify the conditions, documentation requirements, and verification steps that the lender must complete.
The findings that accompany an AUS decision are detailed instructions that tell the lender what documentation must be collected, what verifications must be performed, and what conditions must be satisfied for the loan to be eligible for sale to the GSE. The findings are generated by the AUS based on the specific data inputs and the risk profile of the loan, and they vary from loan to loan depending on the borrower characteristics, the loan structure, and the property type. Common findings include requirements to verify the borrower income with pay stubs and W-2 forms, to verify the borrower assets with bank statements, to obtain a full appraisal or a property inspection waiver, to verify the borrower employment with a verbal verification of employment, and to document the source of any large deposits that appear in the borrower bank statements. The findings are binding conditions that the lender must satisfy, and the failure to satisfy a finding can result in the loan being ineligible for sale to the GSE or subject to a repurchase demand if the loan is sold and subsequently found to be non-compliant.
The representation and warranty relief that the AUS decision provides is a significant benefit for lenders, as it reduces the lender exposure to repurchase demands from the GSEs. When a lender sells a conforming loan to Fannie Mae or Freddie Mac, the lender makes certain representations and warranties about the loan, including that the loan was underwritten in compliance with the GSEs guidelines and that the borrower meets the eligibility requirements. If the GSE determines that a loan does not comply with the representations and warranties, the GSE can require the lender to repurchase the loan, which creates a financial loss for the lender. The AUS decision provides representation and warranty relief for the specific aspects of the loan that the AUS evaluated, meaning that the GSE will not require a repurchase based on those aspects if the lender complied with the AUS findings. The level of representation and warranty relief depends on the AUS decision and the data that was submitted, with Approve or Eligible decisions providing the broadest relief and Refer decisions providing limited or no relief.
The property inspection waiver is a feature of the automated underwriting systems that can eliminate the need for a full property appraisal, saving the borrower money and speeding up the loan process. Both DU and LPA can issue a property inspection waiver, which is a finding that indicates the AUS has sufficient confidence in the estimated property value based on its internal automated valuation model and the data available to the system, and that a full appraisal is not required for the loan to be eligible for sale to the GSE. The property inspection waiver is typically issued for loans with strong borrower profiles, low LTV ratios, and properties in areas where the AUS has reliable valuation data. The property inspection waiver is optional, and the borrower or the lender may choose to obtain a full appraisal even when a waiver is offered, particularly if there are concerns about the property condition or the estimated value.
The difference between the data that is submitted to the AUS and the data that is verified during the loan process is an important distinction that mortgage professionals must understand. When a loan application is submitted to the AUS, the lender enters the loan data based on the information provided by the borrower on the loan application, which may include estimated income, estimated assets, and estimated property value. The AUS generates a decision based on this submitted data, but the decision is contingent on the verification of the data during the loan process. If the verified data differs from the submitted data in a material way, the AUS decision may change, and the loan may need to be re-submitted to the AUS with the corrected data. For example, if the borrower estimated income on the application is $10,000 per month but the verified income based on pay stubs and W-2 forms is $8,000 per month, the DTI ratio may increase beyond the AUS threshold, and the loan may no longer be eligible for an Approve decision.
The resubmission process for the AUS is a common occurrence during the loan origination process, as the initial AUS run is typically based on application data that has not yet been verified, and changes to the data are frequently identified during the verification process. When the verified data differs from the submitted data, the lender updates the loan file and resubmits the loan to the AUS with the corrected data, and the AUS generates a new decision based on the updated information. The resubmission process ensures that the final AUS decision reflects the verified borrower and loan data, and it is a critical step in ensuring that the loan is eligible for sale to the GSE. Lenders should plan for the possibility of resubmission when establishing their loan timelines, as a resubmission can add time to the process, particularly if the corrected data results in a different decision that requires additional analysis or documentation.
The role of the human underwriter in the automated underwriting process is reduced but not eliminated, as the AUS handles the primary risk assessment and decisioning, but the human underwriter is responsible for reviewing the AUS findings, verifying that the documentation supports the data submitted to the AUS, and making final determination on loans that receive a Refer decision or that have complex or unusual circumstances that the AUS was not designed to evaluate. The human underwriter also plays a critical role in ensuring compliance with the AUS findings, as each finding must be satisfied with appropriate documentation, and the underwriter is responsible for reviewing the documentation and confirming that it meets the requirements specified in the finding. The collaboration between the AUS and the human underwriter combines the speed and consistency of algorithmic decisioning with the judgment and expertise of a human reviewer, creating a process that is both efficient and thorough.
In conclusion, automated underwriting systems have transformed the conforming mortgage market by bringing speed, consistency, and data-driven decisioning to the loan evaluation process, and an understanding of how these systems work is essential for mortgage professionals and borrowers alike. The DU and LPA platforms evaluate borrower credit, capacity, and collateral using sophisticated risk models that have been developed from decades of loan performance data, and their decisions determine whether a conforming loan is eligible for sale to the GSE and what conditions must be met for the loan to close. Mortgage professionals who understand the AUS inputs, decision outputs, and findings can structure loan applications that are more likely to receive approve decisions, can anticipate the documentation requirements that the AUS will impose, and can guide borrowers through the underwriting process with greater efficiency and confidence. Borrowers who understand the AUS process can prepare their financial documentation in advance, can address credit or DTI issues before applying, and can work with their loan officer to present the strongest possible application to the automated system. This article is for educational purposes only and does not constitute a loan offer or financial advice. Borrowers should consult a licensed mortgage professional for guidance specific to their individual situation.

Have Questions About Your Mortgage Options?
Speak with Ebonie Beaco, Senior Mortgage Loan Officer (NMLS #2389954).
Mortgage Financing & Mortgage Education
Explore Mortgage Financing Options
Visit MortgageStrategist.net for professional services and resources.
This is a promotional link to an affiliated business of the publisher, separate from the magazine's editorial content.
Editorial Disclosure: This article is published by Mortgage & Real Estate Strategist Magazine for informational and educational purposes. It does not constitute legal, tax, financial, or investment advice. Mortgage and real estate professionals should consult appropriate experts and verify current program guidelines before making decisions based on this content.
Sources & References
- Fannie Mae Desktop Underwriter User Guide and Reference Documentation
- Freddie Mac Loan Product Advisor User Guide and Technical Specifications
- Consumer Financial Protection Bureau Mortgage Underwriting Standards and Guidance
- U.S. Department of Housing and Urban Development FHA Automated Underwriting System Requirements
- Mortgage Bankers Association Automated Underwriting Best Practices Manual
Sources Last Verified: November 8, 2026

Publisher & Mortgage Strategist
Ebonie Beaco is the founder and publisher of Mortgage & Real Estate Strategist Magazine. She is a licensed mortgage loan officer (NMLS #2389954) affiliated with NEXA Lending (NMLS #1660690), providing lending services in multiple states. She covers mortgage strategy, lending programs, and real estate investing.
Read More Articles by Ebonie Beaco
