Waiting more than a month to close can test any borrower’s patience. ICE Mortgage Technology reported an average mortgage closing period of 38.2 days in 2026, and missing records or conflicting details can stretch it further.

An automated underwriting system (AUS) helps lenders spot those issues earlier by reviewing credit, income, assets, debts, and loan eligibility before formal underwriting.

This guide explains how AUS works, what its findings mean, and how lenders can prepare better-organized files for faster review.

TL;DR

• An automated underwriting system reviews mortgage applications and returns a risk and eligibility recommendation for lender review.

• DU and LPA apply Fannie Mae or Freddie Mac requirements to conventional submissions and produce different findings reports.

• AUS findings show whether a file can continue, needs correction, or requires another review path.

• Lenders track review time, resubmissions, file defects, and production volume to measure operational results.

• Addy turns AUS findings into processing tasks, validates borrower records, and requests missing documents before underwriting.

What Is an Automated Underwriting System?

An AUS is software that evaluates loan applications against established lending criteria. It returns a risk and eligibility recommendation based on the submitted file.

During automated loan underwriting, the system checks the borrower’s credit history, income, debts, available assets, employment details, loan terms, and property data.

It uses this information to calculate the debt-to-income ratio and check whether the transaction meets the selected program.

The findings may flag missing records, conflicting information, or other documentation requirements. However, the recommendation doesn’t equal loan approval.

How Does an Automated Underwriting System Work?

An AUS receives application details, checks them against the selected program, and returns findings for lender review. The AUS completes the review in five stages.

1. Loan Information Enters the AUS

Application details pass from the loan origination system (LOS) into the AUS. The system may also retrieve credit reports from approved sources to confirm recorded debts and payment activity.

Accurate entries matter from the first submission. Unsupported income can overstate qualifying capacity, while an omitted debt can make monthly obligations appear lower.

2. The AUS Reviews the File

Once the information enters the system, the AUS uses calculations and rules engines to test it against program requirements. It also checks qualifying ratios and related entries for conflicts.

This risk assessment reflects the file as submitted. Missing or unsupported information can affect the outcome until the processor corrects it.

3. The System Returns a Recommendation

After the review, the AUS assigns a recommendation based on the information submitted. The exact label depends on the system and mortgage program, such as:

• Approve/Eligible

• Accept

• Refer

• Caution

• Refer with Caution

• Ineligible

• Out of Scope

The result indicates whether the application meets automated requirements or needs closer evaluation. It also gives the lender an immediate view of how the file can proceed.

4. The Findings Set the Next Tasks

Along with the recommendation, the AUS provides messages that explain which items need attention. If it requests updated income records, the processor checks whether current documents support the amount entered.

The report may also request proof of funds or identify an eligibility issue. Processors can then address the specific requirement before sending the file for human review.

5. Verified Changes Trigger Resubmission

The processor updates entries only when source records confirm the change. The team then reruns the AUS when those revisions could affect the recommendation.

Each submission adds to the file’s audit trail by recording the information reviewed and the outcome returned. Accurate updates can contribute to faster processing by removing resolved findings from the latest report.

Key Factors a Mortgage AUS Evaluates

A mortgage AUS examines the risk factors that affect borrower eligibility and transaction risk. Each factor can carry different weight depending on the selected program.

Credit and Monthly Obligations

A credit score gives the AUS a summary of the borrower’s credit standing. However, payment history shows whether they’ve paid previous obligations as agreed.

Records from credit bureaus also show current balances, required payments, and recent inquiries. A new account can raise monthly debt before it changes the score.

This explains why borrowers with similar scores may receive different outcomes. Their current obligations, proposed payments, and reserves may differ.

Income, Employment, and Assets

The borrower’s employment history helps show whether earnings are likely to continue. Recent job changes, variable pay, or self-employment may require a longer record than fixed wages.

Financial statements and account records give the mortgage lender context about funds available for closing. They also show whether the borrower will retain reserves after the transaction.

Large deposits need a documented source when they affect qualifying assets. Stable income and sufficient reserves may also offset weaker credit factors under the program’s rules.

Property and Loan Details

Occupancy affects the AUS evaluation criteria because primary residences and investment properties carry different requirements. Property type can also affect whether the selected program accepts the transaction.

The loan amount and property value determine how much equity the borrower has at closing. Less equity can increase the lender’s exposure after a default.

The mortgage program also sets the investor and regulatory requirements for the file. A borrower may qualify financially but still receive an ineligible result when the transaction falls outside those rules.

These criteria feed the two leading AUS options used for conventional mortgages.

Desktop Underwriter vs Loan Product Advisor

Desktop Underwriter (DU) and Loan Product Advisor (LPA) are the main AUS options for conventional mortgages. DU applies Fannie Mae requirements, while LPA reviews files for Freddie Mac.

DU returns an Underwriting Findings report that organizes eligibility messages and required file actions. It can also validate eligible income, employment, and asset information through approved third-party reports.

LPA provides a Feedback Certificate that lists assessment messages and applicable documentation needs. Freddie Mac also offers feedback features that help users examine factors affecting the submission.

A lender may run either system when the file meets the applicable agency requirements. Some eligible mortgages may run through both.

Results can vary because each system follows its own agency rules and risk model. Neither option is always more flexible, so lenders still follow investor requirements and internal policies when making underwriting decisions.

What Do Automated Underwriting Findings Mean?

Automated underwriting findings summarize how the AUS classified the submission. Each label points to a different level of risk, eligibility, or review.

The report also shows which factors influenced the result. This helps the lender separate credit concerns from program conflicts or missing verification.

Approve/Eligible and Accept

DU uses Approve/Eligible, while LPA uses Accept. Both mean the file passed the applicable automated credit and eligibility checks based on the submitted details.

These results can lead to a favorable credit decision, but they don’t represent the final approval decision. The appraisal, required verifications, and outstanding conditions still need review.

Refer, Caution, and Refer With Caution

A Refer may send an eligible government-backed file through manual underwriting processes. The program’s rules determine whether that option applies.

Caution directs the lender to LPA’s feedback for the issues behind the result. Refer with Caution means DU didn’t accept the conventional submission under its automated requirements.

These outcomes don’t always point to missing paperwork. Credit concerns or program conflicts may also prevent an automated approval.

Ineligible and Out of Scope

Ineligible and Out of Scope usually mean the submission falls outside the selected program. The issue may involve occupancy, property type, loan amount, transaction structure, or product rules.

The findings report identifies the requirement behind the result. This keeps a program mismatch separate from a negative risk assessment or manual review.

How Can Lenders Respond to an AUS Refer?

An AUS Refer means underwriting automation couldn’t approve the file as submitted. Below are seven steps lenders can use to choose the next path in the loan process.

1. Confirm the application details: Compare the submitted financial data with the source records. This check can catch data entry errors that affected the result.

2. Read every AUS message: Review the reason attached to each finding before changing the file. This prevents the processor from correcting the wrong issue.

3. Complete targeted document collection: Request records that address the cited concern. For example, an unexplained deposit may need an account statement and proof of its source.

4. Resubmit supported changes: Update the application only when new records or revised terms support the correction. Then rerun the AUS to receive current findings.

5. Try the other government-sponsored enterprise (GSE) system: Submit to the other GSE system when the file qualifies under both agencies. Different agency requirements may produce another result.

6. Check manual underwriting rules: Some cases require human judgment, but the program and lender need to permit manual review.

7. Review another product: A government-backed or non-qualified mortgage (non-QM) option may suit a file outside the original program.

Addressing the cited concern first improves decision accuracy and lowers the chance of repeated findings.

What’s the Difference Between Manual and Automated Underwriting?

In mortgage lending, an automated system applies preset calculations and program rules to standard applications. It returns findings quickly but cannot assess circumstances outside those requirements.

Manual underwriting relies on a qualified professional to evaluate the complete file. The human underwriter can review compensating factors, complex income, and permitted exceptions that need individual consideration.

Many lenders and financial institutions use both methods. AUS handles repeatable assessments, while underwriters evaluate nonstandard cases and retain authority over final decisions.

That process works best when the underwriter receives an organized file with relevant findings and supporting records already identified.

Book a demo with Addy to see how AI agents prepare mortgage files for a faster underwriting review.

Benefits of Automated Underwriting for Lenders

Automated underwriting shortens review cycles and lowers processing costs by handling repetitive calculations. It also minimizes human errors from manual re-entry and uneven file checks.

Standard rules apply the same initial checks to comparable applications. Recorded results show audit and quality-control teams how the system assessed each submission.

During high-volume periods, lenders can handle more files without giving each one the same review time. Routine submissions need fewer touches, while experienced underwriters receive complex cases sooner.

Earlier findings also lower correction and resubmission rates. Borrowers receive requests sooner, so unresolved items are less likely to affect later stages.

Automated Underwriting KPIs to Track

Most lenders should measure speed, file quality, recommendation trends, compliance, and production. They can analyze historical data to see whether new automation software improved the full workflow or only one stage.

Processing Performance

Track the time from a completed application to the first AUS result. Underwriting checklist completion and processor review times show how much work remains after that result.

The time from the AUS result to formal underwriting review shows whether the prepared file reaches an underwriter sooner. The full closing cycle reveals whether those gains continue through the lending process.

File Quality and AUS Outcomes

Approve/Eligible, Accept, Refer, Caution, and Ineligible rates show the share of files in each category. Refer-to-approval conversion and manual underwriting rates show how often referred files later qualify.

Resubmissions, missing-document rates, corrections, and conditions per file measure intake quality. High rates may point to weak verification or incomplete application details.

Compliance and Production Results

Documentation defects, audit exceptions, post-close defects, and unresolved AUS messages show whether review quality remains steady. These measures also support model governance by revealing patterns that need investigation.

Applications per processor, loans per underwriter, borrower request time, and pull-through rate show how much work the operation completes.

Best Practices for Using a Mortgage AUS

Use these practices to improve submission quality and avoid repeated AUS runs.

• Check the first submission carefully: Compare basic loan application information with borrower records. Names, income, debts, occupancy, and transaction terms should match the source files.

• Gather records before review: Collect documents that verify every entered amount. Early preparation prevents avoidable conditions from appearing after processing begins.

• Read every finding: Review the full report, not only the recommendation label. Individual messages may identify verification needs or eligibility concerns.

• Organize findings by subject: Separate income, asset, credit, property, and eligibility items. This gives processors an easier way to assign tasks and prepare borrower requests.

• Keep proof for every update: Record why an entry changed and retain the source behind it. This creates a reliable record for later review.

• Rerun AUS after material changes: Updated income, debts, assets, occupancy, property details, or terms may change the result.

• Follow policy updates: Review current agency, investor, government program, and internal guidance before relying on earlier findings.

• Retain professional oversight: Underwriters should review exceptions and make formal decisions.

These habits improve file accuracy and make AUS findings easier to review.

Prepare Automated Underwriting Files Faster With Addy

Addy uses artificial intelligence to review AUS findings and prepare applications before formal underwriting. Its agents check records, flag missing items, and organize the remaining requirements.

The platform works alongside Desktop Underwriter, Loan Product Advisor, and licensed mortgage professionals. Underwriters keep final authority over every lending decision.

Pre-Underwrite Loans in Under 5 Minutes

Addy reviews borrower details, supporting records, and AUS findings before underwriting. Its ChatGPT app can return structured pre-underwriting findings in under five minutes.

The review identifies open conditions and information that needs clarification. Processors can resolve these issues earlier in the process.

Convert AUS Findings Into a Processing Checklist

Addy’s Processing Checklist compares supporting documents with AUS findings and lender guidelines. It turns unresolved requirements into product-specific tasks.

Each item appears as complete or pending, so processors can see what remains. The checklist covers income, employment, assets, credit, property, and other applicable conditions.

Extract and Validate Records With Document AI

Addy’s Document AI uses machine learning algorithms to read common records, including:

• 1003s

• Pay stubs

• W-2s and 1099s

• Tax returns

• Bank statements

It extracts borrower details, classifies each document, and compares related figures for inconsistencies. Addy then links the record with the correct application.

Documents can sync from the LOS or arrive through the borrower’s email inbox. This reduces repeated uploading and manual entry.

Automate Document Requests and Updates

Addy can request outstanding records by email, text, or phone. Each follow-up reflects the file status and the lender’s instructions.

The platform connects with loan origination, customer relationship management (CRM), and point-of-sale (POS) systems. It also works with Gmail, Outlook, Slack, Microsoft Teams, browser workflows, and ChatGPT.

These integrations keep document requests and file updates within the tools lenders already use for lending services.

With accurate inputs, current findings, and professional oversight, lenders can prepare more complete files for review.

Book a demo with Addy to explore a faster way to prepare mortgage files for underwriting.

FAQs About the Automated Underwriting System

What is Fannie Mae’s automated underwriting system called?

Fannie Mae’s AUS is Desktop Underwriter (DU). It assesses submitted mortgage information against Fannie Mae’s credit risk and eligibility requirements.

Can a lender run both DU and LPA?

Yes, a lender can run both systems when the mortgage qualifies under each agency’s submission rules. LPA replaced Freddie Mac’s former Loan Prospector system.

How long does automated underwriting take?

An AUS can return initial findings within minutes. Formal underwriting continues until the lender verifies documents, conditions, and property information.