Why Digital VOD Is Only Half the Battle
In the quest to build a truly digital mortgage, asset verification was supposed to be one of the “easy” wins. The industry successfully moved from paper-based Verification of Deposit (VOD) forms to direct digital connections via providers like Finicity, Plaid, and Blend. The promise was clear: no more PDFs, no more “stare and compare,” and instant data validation.
However, many operations managers have discovered a frustrating reality: while the collection of asset data has been digitized, the underwriting of that data remains stubbornly manual.
Even with a digital data stream, underwriters are still performing a rhythmic, tiring dance. They are looking at transaction logs to source large deposits, cross-referencing recurring withdrawals against the credit report to find undisclosed debts, and manually calculating months of reserves.
The industry has traded a paper bottleneck for a data bottleneck. To break free, lenders need to move beyond simple data connections and toward AI-driven asset verification—a system that doesn’t just “show” the data, but understands the financial story it tells.
The “Long Tail” of Asset Friction
The reason asset verification remains so manual isn’t because we lack the data; it’s because mortgage guidelines—specifically those from Fannie Mae and Freddie Mac—require a level of “forensic accounting” that traditional systems aren’t built to handle.
Consider the “Large Deposit” rule. According to the Fannie Mae Selling Guide, underwriters must source any deposit that exceeds 50% of the total monthly qualifying income for purchase transactions on one-unit properties. In a legacy workflow, this means an underwriter must:
- Calculate the 50% threshold based on the qualifying income.
- Scan two months of bank statements (often 10-20 pages) to identify every deposit.
- Compare each deposit against the threshold.
- If a deposit exceeds the threshold, they must manually add a condition for a Letter of Explanation (LOE) and supporting documentation (like a bill of sale or gift letter).
This is the definition of “clerical” work that drains underwriting capacity. As we discussed in our look at 10 manual tasks killing underwriting productivity, these small inefficiencies pile up fast. When you multiply this process by the number of accounts a typical borrower has, you realize why asset clearing is often one of the last hurdles before a loan is “Clear to Close” (CTC).
AI-Native Analysis vs. Simple Data Connections
A modern, AI-native Loan Origination System (LOS) treats asset data differently than a legacy system with a “digital VOD” plug-in.
In a legacy system, the digital VOD data is usually flattened into a PDF or a static table. It’s a “digital filing cabinet” approach. An AI-native system, however, ingests the transaction data into a structured model that is constantly monitored by machine learning algorithms.
1. Automated Large Deposit Sourcing
Instead of an underwriter scanning lines of code, an AI-native system automatically identifies every deposit that meets the GSE threshold. It doesn’t just flag them; it attempts to classify them. Is it a payroll deposit? The system cross-references it with the AI-driven income calculation results. Is it a transfer between accounts? The system identifies the “out” transaction on the corresponding account and automatically reconciles the two, clearing the need for an underwriter to even see the flag.
2. Identifying Undisclosed Debts
One of the leading causes of post-close quality issues is the discovery of undisclosed liabilities. AI-native systems analyze transaction patterns to find recurring payments that don’t appear on the credit report—alimony, child support, or “shadow” debt like Buy Now, Pay Later (BNPL) loans. By surfacing these risks during the initial processing phase, lenders can avoid the dreaded “denial at the finish line” that occurs when a final QC check finds a new monthly obligation.
3. Real-Time Reserve Calculations
Reserves aren’t static. In a volatile market where closing costs can shift based on interest rate locks and prorated taxes, the required reserves for a loan can change daily. An AI-native LOS performs real-time reserve calculations, instantly notifying the team if a borrower’s available funds fall below the required threshold due to a change in loan terms.
Closing the LOE Gap
The biggest time-waster in asset underwriting isn’t the analysis itself—it’s the “chase.” When a large deposit is found, the file often grinds to a halt while the processor or underwriter requests a Letter of Explanation from the borrower.
In a progressive automation strategy, this “chase” is eliminated. As soon as the AI identifies an unsourced large deposit, the system can automatically:
- Generate a targeted task in the borrower portal.
- Explain the requirement in plain English (e.g., “We noticed a deposit of $5,000 on Jan 12th. Please tell us the source of these funds and provide documentation, such as a gift letter or bill of sale.”).
- Mark the condition as “Pending Borrower” without an underwriter ever touching the file.
By the time the underwriter opens the loan for the first review, the LOE and the supporting documents are often already in the file. This is the shift from “stare and compare” to exception-based underwriting, where the machine handles the clerical gathering and the human focuses on the final decision.
Day 1 Certainty and the Path to Repurchase Relief
The shift toward AI-driven asset verification isn’t just about speed; it’s about certainty. The Government-Sponsored Enterprises (GSEs) have provided a roadmap for this through programs like Fannie Mae’s Day 1 Certainty and Freddie Mac’s Asset and Income Modeler (AIM).
These programs offer lenders relief from representation and warranty (rep and warrant) claims on specific components like assets, provided the data is validated through an approved vendor and meets specific criteria.
However, simply “using an approved vendor” isn’t enough to get the full benefit. Lenders often fail to achieve Day 1 Certainty because their LOS doesn’t correctly map the data to the GSE’s requirements, or they allow manual overrides that void the relief.
An AI-native LOS acts as a “guardrail.” It ensures that the data moving from the bank connection to the GSE’s portal remains “clean” and consistent. By automating the extraction and validation process, the system helps lenders achieve higher rates of rep and warrant relief, which is a critical component of reducing repurchase risk.
The ROI of Moving Beyond the PDF
For many lenders, the cost of asset verification is buried in the general overhead of processing and underwriting. But when you look at the ROI of AI-native asset analysis, the numbers are compelling:
- Reduced Turn Times: Automating the identification and sourcing of deposits can shave 24-48 hours off the “conditional approval” phase.
- Higher “Loans per Month” (LPM): By removing the clerical “forensic accounting” work, underwriters can handle more files without increasing their stress levels or error rates.
- Lower Post-Close QC Costs: Catching undisclosed debts and unsourced deposits early reduces the need for expensive post-close corrections and buyback defenses.
As we noted in The Story of Loancrate, we didn’t build a new LOS just to make it “digital.” We built it to make it intelligent. Assets aren’t just a balance on a screen; they are a dynamic part of the borrower’s financial story.
From Static Balance to Dynamic Data Model
The mortgage industry is moving away from a model where the LOS is a passive recipient of documents and toward a model where the LOS is an active participant in the credit decision.
Lenders who continue to rely on manual “stare and compare” for bank statements—even digital ones—will find themselves at a significant disadvantage in a market where speed and certainty are the primary differentiators. The future belongs to the AI-native lender who can verify assets in minutes, not days, and provide a seamless, automated experience for the modern borrower.
At Loancrate, we’re building that future today. By turning raw transaction data into actionable underwriting insights, we’re helping mortgage teams move beyond the bank statement and toward a new standard of operational excellence.