The automated underwriting system for non-agency and portfolio lending
Agency loans have DU and LPA. Everything else gets underwritten manually. AUS runs your credit policy against Non-QM, DSCR, business purpose, and portfolio book loans in real time.
A credit decision engine that can prove how it decided.
AUS is an automated underwriting system. It evaluates a complete loan file against the guidelines you published and returns approve, review, or rejected, with findings and conditions. AI drafts those rules from guideline text. Only the published rules decide the file.
The workflow
Bring a guideline. AI drafts the conditions. Your team reviews and publishes them. AUS then runs those rules. Same file, same answer, and every finding points at the guideline line.
Prove the decision
Every approve, review, or rejected traces to the rule that fired and the guideline behind it. Examiners, investors, fair lending, and warehouse QC can replay the file and see the same finding.
AI’s job
AI drafts underwriting rules from guideline text. It never runs on a loan file. Only the published rules decide.
Mitigate underwriting risk
A model returns a probability and can answer the same file two ways. A rule is a fixed condition. Keeping the model out of the decision cuts inconsistent decisions, unexplainable findings, and policy drift.
Your guidelines. Your credit box. No vendor release cycle.
Your team builds, edits, and deploys the rules in the platform. When an investor changes a guideline, you change the rule and publish it, the same day, without a ticket to a vendor who has to interpret it first. AI drafts rule conditions from your guideline text so a new program goes up faster. The rules alone make the decision.
What you can express as a rule
Non-QM documentation: bank statement, asset depletion, 1099, P&L
Business purpose: DSCR, fix and flip, bridge, construction
Agency and portfolio programs you keep in-house: conventional, HELOC, portfolio products
Other programs you can write as rules
Conditions tied to scenario data (income type, loan amount, LTV, investor preference)
Different rule sets by product, channel, investor, or user role
“Configurable in a way that other systems fall short. The tools are there to solve complex problems specific to our business.”
VP of Mortgage Operations, Verus Mortgage Capital
Non-agency and portfolio, on one automated underwriting system
Desktop Underwriter and Loan Product Advisor are the automated underwriting systems for agency-eligible loans. They are the standard. AUS does not replace them. It is the automated underwriting system for everything they will not underwrite.
Banks and Credit Unions
Specialty programs and portfolio products, with a rule-level audit trail.
Banks and Credit Unions
Non-QM Lenders
Bank statement, asset depletion, 1099, and P&L, decisioned automatically.
Non-QM Lending
TPO Lenders
Wholesale and correspondent files, decisioned against each investor’s overlays.
TPO Lenders
Business Purpose Lenders
DSCR, fix and flip, bridge, and construction, on your credit box.
Business Purpose Lending
The price is only good if the loan is approvable
A quoted rate that fails underwriting is not a price. AUS holds the credit policy; PricingEngine reads the same policy. PMIPricer quotes the premium from the facts already on the file. You can configure rules that restrict locks or pricing until the file is eligible.
“LoanPASS is faster, it’s cheaper, and it’s really just a much better tool for decisioning than any other PPE in the industry.”
Founder & CEO, Loan House
Fast enough to quote. Strict enough to audit.
Automated credit decisioning only earns its place if it is fast enough to sit in the quoting path and strict enough to survive an audit. AUS is built for both.
Under [X]
Underwriting results returned in real time, findings and conditions included.
[X] traceable
Every decision traces to the rule that produced it. Same input, same output.
AUS decides. The rest of the stack feeds it.
Feeds the engine
LOS
POS
Document AI
LOS, POS, and document tools send the file. Document AI reads and verifies. It does not replace an automated underwriting system.
Decides
AUS
The credit decision engine returns the determination.
AI writes the rules faster. Only the rules decide the loan.
AI works in two places: writing rules and answering questions about them. It never runs on the loan file.
Explore AI at LoanPASS
Configuration
Stand up new programs and encode guideline changes.
Support
Answers about your own rules and decisions, not a vendor queue.
Resiliency
Update once and reprice the shelf when rates move.
“We can add fields, update pricing, or adjust product guidelines on the fly, without waiting on development teams.”
EVP of Capital Markets, SmartFi
Implementation and partnership
Live in days. Supported by the people who build the engine.
Days to weeks, not quarters. Dashboard access on day one. You encode the first programs as rules alongside the people who build the platform.
Step 1
Load your guidelines.
Step 2
Connect your stack (dashboard, API, embeds).
Step 3
Validate recent files against your own manual decisions, then go live.
“The people at LoanPASS are as impressive as their technology. I don’t know if we are their most important client, but they consistently make me feel like I am.”
VP of Mortgage Operations, Verus Mortgage Capital
“After using LoanPASS to support select products for almost two years, the value was so clear that we didn’t hesitate to expand our engagement and sign a new three-year agreement.”
Managing Director of Correspondent Lending, Acra Lending
Decision, price, and MI on one file.
AUS decides the file. PricingEngine prices only the products that decision says the file fits. PMIPricer quotes MI from the LTV, FICO, and product already on the file. No re-keying.
AUS
Decision
AUS returns approve, review, or rejected against your credit policy.
Schedule a demo
PricingEngine
Price
PricingEngine returns rate, points, and adjustments for the products that decision says the file fits.
Explore PricingEngine
PMIPricer
MI
PMIPricer compares every carrier using the LTV, FICO, and product already on the file. No re-keying.
Explore PMIPricer
If you still have questions, ask us.
What is AUS?
AUS is LoanPASS’s automated underwriting system and credit decision engine. It evaluates a loan file against the guidelines you configured and returns an eligibility decision with findings and conditions.
What is automated underwriting software?
Automated underwriting software evaluates a loan file against guidelines and returns a decision without a manual first pass. Desktop Underwriter and Loan Product Advisor handle automated loan underwriting for agency loans. AUS does it for non-agency and portfolio loans, against your credit policy.
What is a credit decision engine?
A credit decision engine evaluates borrower and loan data against credit policy rules and returns a determination automatically. In LoanPASS, underwriting and pricing run on those rules, so a quoted rate is a rate the file is eligible for. PMIPricer quotes MI from the same file.
Can we build our own underwriting rules?
Yes. You build, manage, and deploy your own rules. You do not wait on a vendor release to change a guideline.
Does it work with Non-QM and business purpose loans?
Yes. Bank statement, DSCR, asset depletion, 1099, P&L, fix and flip, bridge, construction, and any other program you can express as rules.
Can we generate custom conditions?
Yes. You can define conditions from scenario data (income type, loan amount, LTV, investor preference) and return them with the decision.
How fast are underwriting decisions?
A full rules evaluation with findings and conditions, returned in real time.
How long does automated underwriting take?
The decision itself is returned in real time. Implementation is days to weeks, not quarters.
Is AUS a rules engine or a full automated underwriting system?
Both. It is a credit decision engine that powers real-time decisioning and returns conditions, eligibility, and findings in an auditable format.
Can underwriting decisions block pricing or locking?
Yes. You can configure rules that restrict locks or pricing until the loan is eligible.
How do we audit decisions or track rule changes?
Decisions and rule changes are logged. You can view historical decisions and export rule versions.
Can different products or channels use different underwriting logic?
Yes. You can assign different rule sets by product, channel, investor, or user role.
Should we buy an AUS or build one?
Most lenders outside the agency box still route files to a desk because DU and LPA will not take them. Building mortgage underwriting software in-house means owning a rules engine, a pricing sync problem, and a maintenance queue for every guideline change. AUS is the buy path: your guidelines, running as an automated underwriting system, without the multi-year build.
Does LoanPASS use AI to underwrite loans?
No, deliberately. Decisions come from deterministic rules you define. AI speeds up configuration, guideline changes, and answers about your rules.
What are AUS findings on a non-agency loan?
Findings are the decision plus the reasons behind it. DU returns Approve/Eligible or Refer; LPA returns Accept or Caution paired with Eligible or Ineligible. AUS returns approve, review, or rejected in your vocabulary, with each condition traced to the guideline rule that produced it.
Does this replace DU or LPA?
No. DU and LPA underwrite agency guidelines you cannot change. AUS underwrites yours.
Automated underwriting you can explain to an examiner
Bring a guideline set to the demo. We will show you how it runs as rules.
Schedule a Demo