Mortgage automation in action: how Key Mortgage cut setup time by 67%

TL;DR: Key Mortgage Services, an independent family-owned lender, reduced document indexing time by up to 67% by deploying Ocrolus Classify within Encompass, cutting per-file setup from 20 to 30 minutes down to 5 to 10 minutes. The setup team now handles more loan volume without adding headcount, and manual stare-and-compare verification has been eliminated entirely. This post breaks down how mortgage automation removes one of the most persistent bottlenecks in loan setup.

For mortgage setup teams, document indexing is one of the most time-consuming tasks that adds the least underwriting value. A processor opens a new loan file and spends 20 to 30 minutes manually sorting documents into the correct folders before a single piece of income documentation has been reviewed. That time compounds with every file, and the only way to grow throughput is to add more people.

Key Mortgage Services faced that constraint directly. The independent, family-owned lender relied on a third-party vendor to access each loan file and manually drag documents into the correct folders. After deploying Ocrolus Classify within Encompass, Key Mortgage cut that 20-to-30-minute window to 5 to 10 minutes per file, a reduction of up to 67%, without adding a single person to its setup team.

The true cost of manual document indexing

At 20 to 30 minutes per file, document sorting alone consumes a significant share of setup capacity before any underwriting work begins. The time is predictable but not recoverable, and it scales directly with loan volume.

Beyond the time investment, manual indexing introduces a compounding problem: verification. Key Mortgage System Administrator Luba Mainz called it “stare and compare,” manually confirming that each verification of employment appeared consistently across the 1003, AUS and loan file. Every discrepancy required tracing back to source, adding review cycles that delayed the file. “Indexing was manual,” said Mainz. “We used a third party that would go into the file, basically drag and drop things where they go.”

For lenders trying to grow, this math gets worse with volume. The MBA’s Quarterly Mortgage Bankers Performance Report put total loan production expenses at $11,988 per loan in the first quarter of 2026, with a $2,500 per-loan spread between top- and bottom-performing lenders. Headcount-driven throughput has a ceiling that technology-driven throughput does not.

How automated classification works inside Encompass

Key Mortgage deployed Ocrolus Classify directly within its Encompass workflow. When a loan is submitted to fulfillment, Ocrolus classifies each document and places it in the correct location before the setup team begins its review. Staff now work through a short unknown folder, typically five documents or fewer per file, then move directly to ordering services.

The implementation was self-directed from day one. Ocrolus’s document mapping feature lets Mainz correct classifications in real time without routing change requests through a support team. “It was very easy,” said Mainz. “Right as we implemented, Ocrolus came out with the document mapping that we can do ourselves, which was way quicker than trying to reach out to people.”

The stare-and-compare step is now automated. Cross-checks that once required a staff member to manually verify each VOE across the 1003, AUS and loan file happen before a processor opens the file. What previously added friction to every loan is no longer a task at all. Ocrolus also runs income calculations on every file, giving Key Mortgage a foundation for the income automation work it plans to build next.

More loans, same team

With classification handled automatically, Key Mortgage’s setup department increased loan throughput without adding staff. “We’ve been able to increase how many loans go through our setup department without adding people,” said Mainz, “because it’s not a human that’s doing the indexing.”

That outcome reflects what happens when lenders remove the constraint between volume and headcount. Automated mortgage document processing allows throughput to scale without a corresponding increase in labor costs. Manual workflows do not offer that flexibility: the only way to sort more documents is to hire more people to sort them.

Trust in the platform has grown alongside results. “People can see that it’s working and that it’s faster than a human doing it alone,” said Mainz. For lenders evaluating AI, her advice is direct: vet the vendor, then commit to trusting the technology. “You have to vet your vendor, but then you have to trust the vendor and realize that yes, AI can make mistakes, but you need to trust that it’s doing what it’s supposed to be doing.”

Classification first, income next

Key Mortgage is now building toward broader use of Ocrolus income analysis, working through how to fit it into existing workflows. “Income’s definitely the next step,” said Mainz. “We’re working through just how we fit it into the workflow and continue to build trust with AI and automation.”

Document classification is the right automation starting point for most setup teams: high-frequency, rule-based and directly upstream of every other loan-setup task. A team not spending 20 to 30 minutes sorting documents per file has that capacity back for reviewing income, flagging conditions and moving loans forward. That reallocation is where throughput gains compound. Read the Key Mortgage customer story to see the Encompass workflow in practice.

Key takeaways

• Key Mortgage Services reduced document indexing time by up to 67% with Ocrolus Classify, cutting per-file setup from 20 to 30 minutes down to 5 to 10 minutes.
• Manual “stare and compare” verification (cross-checking VOEs across the 1003, AUS and loan file) was eliminated entirely; those checks now happen automatically before a processor opens the file.
• The setup team increased loan throughput without adding headcount, because classification is no longer a human task.
• Ocrolus’s document mapping feature enabled self-directed implementation, letting the team correct classifications in real time without routing requests through support.
• Income analysis is the next planned integration, extending automation from document sorting into underwriting calculations within the same Encompass environment.

FAQs

What is automated mortgage document classification? 

Automated mortgage document classification is the use of AI to sort incoming loan documents into the correct folders within a loan origination system, such as Encompass, without human intervention. Instead of a processor manually placing each document, the system reads and routes it automatically. Key Mortgage Services reduced per-file indexing time by up to 67% using Ocrolus Classify.

How does Ocrolus Classify integrate with Encompass? 

Ocrolus Classify integrates directly within Encompass. When a loan is submitted to fulfillment, Ocrolus automatically classifies each document and places it in the correct location before the setup team begins review. The integration also includes a document mapping feature that lets administrators correct misclassifications in real time without requiring support tickets.

How much time can automated document classification save mortgage lenders?

Results vary by loan type and file complexity, but Key Mortgage Services saw indexing time drop from 20 to 30 minutes per file to 5 to 10 minutes per file, a reduction of up to 67%. For a team processing 50 loans a month, that equates to recovering 17 to 25 staff hours previously spent on manual document sorting.

Why do mortgage setup teams still rely on manual document indexing?

Most lenders rely on manual indexing because it is a known, controllable process. The risks feel manageable even if the time cost is high. The shift to automation requires trust in the classification accuracy of the underlying AI model. As Luba Mainz of Key Mortgage noted, the key is vetting the vendor thoroughly, then committing to trusting the technology rather than second-guessing every output.

What is stare-and-compare in mortgage processing?

Stare-and-compare refers to the manual practice of verifying that a document such as a verification of employment (VOE) appears consistently across multiple places in a loan file, including the 1003, AUS and loan file. It is time-intensive and prone to human error. Automated document processing tools like Ocrolus perform these cross-checks automatically before a processor opens the file.