Now assisting mortgage lenders with cloud automation is Vaultedge
“People would have laughed at you if you had told them a decade ago that Vaultedge is helping mortgage lenders automate in the cloud. The internet has made it easier than ever to learn just about anything you want. So with all this content and information available, how did you know which ones to learn from and which ones actually work?”
Luckily, you are in Real Estate Heaven and for today’s blog we will be talking about Vaultedge. Did you know that they are now assisting mortgage lenders with cloud automation? Remember, that while the first step is completing any of this amazing information, the second and possibly more important step is taking action even if it’s imperfect action.
The goal of Vaultedge is to fully automate mortgage underwriting and due diligence. The splitting, indexing, data extraction, and validation of mortgage documents are automated by Vaultedge’s AI-powered mortgage document processing software, saving up to 80% of time and money in loan processing, post close, and loan boarding. Leading mortgage companies use Vaultedge’s software to automatically process thousands of loans each day.
Through the cloud marketplace Microsoft AppSource, Vaultedge Software is making its AI-powered document recognition and data extraction software available to mortgage lenders and servicers.
Vaultedge Mortgage Automation (VMA), which was developed on the Microsoft Azure platform, can be integrated with loan origination systems like ICE Mortgage Technology’s Encompass and can process up to 10,000 loans per day, the company said on Monday.
Using optical character recognition and artificial intelligence (AI), VMA can automatically identify documents, extract data from them, and determine income. According to the business, loan processing, post close, and loan boarding can all be done with time and cost savings of up to 80%.
According to co-founder and CEO Murali Tirupati, Vaultedge is “razor-focused on helping lenders and servicers reduce costs and improve response time through automation, while at the same time making life easy for end-users” like loan officers, loan processors, and underwriters.
Vaultedge, a 2015 startup, announced a partnership with ICE Mortgage Technology and Encompass integration in July. After an abrupt increase in mortgage rates abruptly put an end to the refinancing boom, mortgage lenders are looking for ways to stay competitive and reduce costs. One of those ways is to automate the processing of mortgage applications, underwriting, approvals, and closings.
Solutions are frequently cloud-based, and companies like Salesforce, Google, Amazon, and Microsoft all market services designed to assist digital mortgage service providers in digitizing and processing mortgage applications and documents.
Rocket Mortgage, the largest mortgage lender in the country, is providing banks and credit unions with an end-to-end “mortgage-as-a-service” through Salesforce Financial Services Cloud.
Microsoft Cloud for Financial Services was introduced last year with partners like Mortgage365 and Finastra, the creator of the Fusion Mortgagebot platform, and integrates cloud services across Microsoft’s public cloud offerings.
Lending DocAI, a specifically designed mortgage solution from Google Cloud that was also introduced last year, assists partners like Roostify and Mr. Cooper in automating data entry and designing and customizing document processing workflows.
Underwriter Assist, an AI-powered mortgage solution from Black Knight, uses Amazon Textract and Black Knight-developed mortgage-specific algorithms to extract data from documents like W2s and pay stubs. Underwriter Assist runs on Amazon Web Services (AWS).
In order to automate loan processing for mortgage lenders, especially using Encompass® by ICE Mortgage Technology, Vaultedge, a leading OCR platform, today announced that its integration is now built on the most recent Encompass Partner ConnectTM API Platform available through ICE Mortgage TechnologyTM, part of Intercontinental Exchange, Inc. (NYSE: ICE), a leading global provider of data, technology, and market infrastructure.
Through its AI-powered document recognition and data extraction software, Vaultedge seamlessly integrates with Encompass to streamline document processing. To provide OCR capabilities to Encompass users, Vaultedge announces integration with Encompass® by ICE Mortgage TechnologyTM.
With the help of Vaultedge, Encompass users can quickly extract data from loan documents, cross-validate it against other documents, and automatically index loan documents, which streamlines the loan processing process. Without spending hours manually sorting through countless documents, the processing teams can obtain underwrite-ready loan files within Encompass.
“The industry pioneer in guiding mortgage lenders’ digital transformation efforts is ICE Mortgage Technology. Vaultedge is a specially developed OCR program that offers users of Encompass a seamless loan processing experience. In line with our shared commitment to maximizing efficiency and value for our end customers, the partnership between Vaultedge and ICE Mortgage Technology is a logical next step “said Vaultedge’s CEO and co-founder, Murali Tirupati.
With the aid of the AI-based Automated Document Recognition (ADR), Automated Data Extraction (ADE), and Income Analyzer platform from Vaultedge Mortgage Automation (VMA), mortgage lenders, servicers, and investors can automate the processing of mortgage documents with an accuracy rate of 99+ percent. With VMA, manual processing costs are cut by 80% while loan closing times are shortened by 2 weeks.
Some of its main advantages are:
- built with high scalability (10K+ loans processed/day) on the Microsoft Azure platform
- indexes documents, extracts data, and calculates income through integration with loan origination software like Encompass. Official ICE Mortgage Technology OCR partner Vaultedge.
- identifies significant exceptions (data discrepancies, etc.) quickly and performs direct cross-validation with the source.
- saves you manual work by automatically writing verified documents and data to your loan origination and servicing systems.
- Standard/structured and non-standard/unstructured documents: Data Extraction
- Rating of confidence for exception handling simplicity
- on-demand instruction for new documents
- can be put into place in as little as 4 weeks and requires no user training.
For the most part, loan origination and boarding have been manual, paper-intensive processes. Additionally, it has been noted over the last ten years that headcount productivity has decreased by more than 50% while loan production costs have increased at a CAGR of 7%
As a result, the loan origination and boarding process is not operating as effectively as it could.
A typical loan file has between 250 and 400 pages worth of documents, with thousands of data elements that need to be processed and verified. Such documents and data handling by hand is laborious and prone to mistakes. Important mistakes may appear as these documents progress, which could make the loan file unusable. Such mistakes seriously impair loan quality and put lenders and servicers’ businesses at risk.
The bottlenecks caused by a manual and paper-intensive process that result in loan defects will be examined in this article. Based on this knowledge, we will research the main business risks associated with these loan defects and how automation can help to mitigate these risks.
Process Overview for Loan Origination & Boarding
The origination and onboarding of loans is a complicated process with many stakeholders and choke points. What the entire process appears like when viewed from above is as follows:
Numerous channels, including brokers, third party originators, correspondent lenders, co-issue, retail lenders, etc. are used to originate mortgages. The origination process begins with the applicants submitting their employment details, financial statements and credit history for prequalification. Applicants who have been prequalified move on to the loan processing stage, where they must submit a variety of paperwork that typically totals hundreds of pages. These typically include things like their bank account statements, tax returns, pay stubs or W-2s, gift letters, and their history of rental payments.
Following the completion of the loan documentation process, the loan proceeds to the underwriting, credit decisioning, quality control, loan funding, and closure stages. Separate loan assets and MSR assets are created following loan closure. When these assets are given to a mortgage servicer for the sale of MSR, a post-close audit is performed on them. As an alternative, they go through a due diligence procedure as part of managing investor portfolios.
Lenders, servicers, and post-close teams are required to quickly and in bulk ingest data in non-standard formats (paper, scanned images, PDFs, data feeds) and analyze it at each stage of the process. This is a significant bottleneck that could have a negative impact on loan quality and create significant business risks. Let’s start by comprehending these bottlenecks.
Knowing the bottlenecks that cause loan defects
- People intensive processes – leading to lower output
Mortgage origination and servicing operations have always been highly labor-intensive. If we examine the breakdown of production costs, we can see that 66% of the total cost is attributable to personnel costs. Personnel costs alone for production and fulfillment make up 21% of the total.
Ironically, a cost structure with a high human component does not always translate into higher productivity. The average monthly productivity, on the other hand, has decreased over time.
According to the average monthly productivity for retail production between 2003 and 2018, we can see that from 2003 to 2018, the productivity of non-producing retail FTEs and monthly loan officers steadily decreased by almost 60%.
As a result, loan production, which is largely a human-driven process, is seeing a significant decline in output per person.
- Document indexing takes a lot of time and is difficult.
One of the loan origination & boarding process’s most time-consuming steps is document classification, sorting, and validation. When carried out manually, it could typically take 10 to 14 days to finish these tasks.
This is due to the fact that loan packages typically include a variety of documents that total hundreds of pages and differ in format. In a manual process, it is very challenging to quickly and repeatedly determine the accuracy of 500 complex pages of information!
Therefore, manual indexing could soon lead to declining productivity in an environment with a high loan volume.
- Data updating and validation are error-prone processes
The fact that a substantial amount of loan documents are printed out during the loan origination, closing, and boarding stages, even though they are digital, presents a significant challenge. Maintaining an accurate version control is challenging when there is manual data entry or updates in paper documents. Data fields on paper documents and digital files in the loan origination system may not match as a result.
Data validation errors present another problem. A typical loan package contains 250–400 pages in formats ranging from paper documents like photo ID copies to digital documents like scanned PDFs and excels. Thousands of data fields in each of these packages need to be verified. However, there are two main difficulties with manual data validation:
-First, updating customer data consistently across all pertinent documents and sub-systems is very challenging when done manually.
-Second, the origination and boarding processes include embedded federal & GSE regulatory compliance processes. New information must be gathered and added to the core lending systems at each step of the mandatory compliance process. Manual handling makes updating data points in accordance with compliance requirements more prone to error.
Therefore, handling documents and data manually is time-consuming, ineffective, and prone to error. This turns into one of the primary causes of subpar loan quality.
Risks to Business from Loan Defects
- The issue with bad loans and business risks
When we talk about loan quality, we’re talking about loan files that have accurate and sufficient documentation and that adhere to the pre-established loan policies of an originating lender, loan guarantor, loan investor, and/or regulator. Loan defects are any features of the loan asset that do not meet these pre-established requirements.
As a result, the quantity and severity of these flaws serve as a benchmark for loan quality. The quality of loans is lower the more defects there are and the more serious they are. For all the parties involved in the loan production and boarding value chain, it generates two significant business risks:
Lenders’ risk of buybacks
Loans with errors might be returned to lenders so they can be fixed. The worst-case scenario might involve investor buyback requests.
- Lender and servicer financial risk
increases overall production costs, prolongs the loan processing and boarding cycle, and decreases per-employee productivity of overworked staff. As a result, we note that poor loan quality caused by flaws is a significant source of business risks. Let’s take a closer look at these risks in order to reduce them.
- Buyback risks resulting from serious flaws:
Three levels of risk severity are prescribed by the FHA and Fannie Mae loan defect taxonomy.
The critical defects, which prevent a loan package from being sold, are the ones that pose the greatest risk for buy-back. As a result, a lender must account for losses related to loans that cannot be sold. From the viewpoint of a mortgage servicer, critical flaws result in non-compliance problems and fines imposed by regulators and external auditors. Let’s examine some intriguing trends to determine the seriousness of this risk.
Over the past year, the critical defect rate has increased.
It is important to note that over the past year, critical flaws have generally increased. According to the most recent survey by ACES quality management, the overall critical defect increased by 25% from Q2’s 1.88 percent to Q3 2020’s 2.34 percent.
The second-largest source of critical flaws is loan documentation.
In Q3 2020, loan documentation accounted for about 19 percent of all critical defects, placing it as the second-largest contributor when we examine the major causes of critical defects.
- Critical errors in loan documentation increased by 80%.
If we compare the growth in critical defects within each category, we can see that between Q2 2020 and Q3 2020, the number of critical defects related to loan documentation increased by 80%! As a result, one of the main causes of critical flaws in FY 2020 was loan documentation.
In conclusion, critical defects have grown during FY 2020–21, and mistakes in the loan documentation process have emerged as the second most significant contributor to such defects.
Financial risk brought on by operational overstretch
Lenders and servicers face significant financial risks as a result of critical defects.
Price modifications based on loan quality.
Capital allocation for flawed loans
penalties for breaking the rules
Probability of early payment default and foreclosing
Another important, yet overlooked aspect is the cost of re-working a defective loan.
Demand and loan volumes are anticipated to remain high during FY 21–22 because 30-year fixed interest rates are still at historically low levels of 3% and there is a lack of housing inventory.
It’s interesting to note that ACES has found a connection between rising manufacturing-related issues and performance problems. This implies that the likelihood of defective loans is higher the more loans that are produced. As a result of the massive stockpile of loan volumes, the loan production and quality assurance teams are likely to continue being overworked. Reworking flawed loan files will further tax the staff beyond what they can already handle.
This will only lead to a further decline in productivity and an increase in the cost of loan production. According to an analysis of historical loan production costs based on data from MBA.org, loan production costs increased steadily from $ 3600+ in FY 09 to $ 7500+ in FY 20.
If critical defects continue to clog up lenders’ and servicers’ available resource bandwidth, this trend is predicted to worsen.
A Case for Document Automation in Loan Defect Mitigation
Let’s examine how document automation can minimize the aforementioned process risks.
Three workflows are stacked on top of one another. The main tool used by loan processors and underwriters in the core data processing workflow is LOS. It receives data from various sources, including banks, credit agencies, and points of sale. Concurrently, document processing workflow sits in parallel to LOS and ingests data directly from the point of sale. However, it provides features that LOS does not.
Document capture, indexing, classification, and storage are all done automatically by document automation software. The next step is data extraction from both structured and unstructured documents. Additionally, it compares these data points with those from the LOS as part of the validation step.
Additionally, it gathers and manages updated documents to guarantee data consistency with LOS. The ability for the lender or servicer to identify exceptions and automate their resolution is most important. This is crucial because manual exception processing is time-consuming, costly, and prone to mistakes.
Finally, by integrating a document automation software with LOS, it is possible to create and deliver error-free digital loan packages to investors, document custodians, mortgage servicers, and other parties. Basically, document processing automation reduces the risks of loan defects by:
capturing and accurately indexing different document types and formats automatically.
data extraction from these documents and LOS data validation. Correct version control and automated exception handling are both possible. This reduces the likelihood of incomplete data, incorrect data entry, or manual data entry.
Critical flaws were consequently adequately mitigated, along with associated buyback risks and financial risks.
Automating Document Processing: Core Capabilities
Since we realized how important it was to incorporate LOS and document automation software to enhance loan quality and lower business risks. Let’s examine the main features of such software to help with proper evaluation.
Recognizing and Classifying Documents
This entails the automatic identification of various documents and forms, including the W-2, deed of trust, pay stubs, and form 1003. The pages that have been assigned to the correct document types are then indexed. Following that, a dynamic checklist is used to compare these documents and find any missing ones.
Development of Documents
In addition to separating and labeling individual forms and documents, the system should offer a straightforward user interface (UI) to view pages sorted by document type.
Verification of data and handling of exceptions
Structured and unstructured data can be extracted from the mortgage file using the data extraction capabilities. Software platforms can extract more than 2000 data points from the ingested data set and can label (or color-code) the data with a confidence score to show whether or not it requires user review. By simply clicking on the extracted value, users can review these highlighted data points and double-check the error.
Our analysis led us to the conclusion that the primary cause of serious loan defects is the manual and paper-intensive nature of the loan production process. Lenders and servicers face buyback and financial risks as a result of these defects. However, such risks can be reduced when a document processing automation system is integrated with the loan origination system.
They provide servicers and investors with error-free loan packages by automating document indexing, data extraction and validation, faster exception resolution, and delivery of loan packages. In certain circumstances, manual interventions can be cut by 80% to 90%, releasing locked human capital and reducing buyback and financial risks. To put it another way, automating document processing is a low risk, high reward strategy that can significantly reduce risk and increase productivity throughout the entire loan production & boarding process.
We will eventually see a rollback of the forbearance extension and foreclosure moratorium as the COVID pandemic’s effects wane and the economy gradually improves. The Mortgage Bankers Association (MBA) estimates that 1.6 million homeowners nationwide are still enrolled in forbearance plans based on its most recent data. The percentage of loans in forbearance in the portfolios of lenders and servicers is still high as a result of this slowdown in forbearance exit. This presents a significant operational challenge for servicers when transferring MSRs in bulk because it calls for careful handling of borrower data and applications for loss mitigation.
This is because the loan boarding team is overburdened with paperwork when a mortgage servicer purchases bulk loans or MSRs. To ensure that documents and data fields are in order, it takes a tremendous amount of manual work to go through hundreds of documents in non-standardized formats and perform a “stare and compare” analysis. The story doesn’t end there; for servicers, time begins to run out as soon as the MSR transaction is completed. In order for servicing to begin on schedule for newly originated loans or to continue uninterrupted for previously originated loans, these loan files typically need to be onboarded within a few weeks. The loan boarding teams are under a lot of pressure because they have to process documents quickly and accurately in a limited amount of time. However, when we take into account the Consumer Finance Protection Bureau’s regulatory requirements and the slow forbearance exit, this operational challenge multiplies (CFPB).
The CFPB maintains strict oversight over MSR transfers and whole loan portfolio transfers because millions of homeowners are still in forbearance, preventing avoidable foreclosures. In CFPB Bulletin 2021-02, the Bureau advises mortgage servicers to take all necessary precautions now to stop “a wave of avoidable foreclosures in the fall”—when borrowers leaving COVID-19 moratoriums and forbearance plans start to ask for loss mitigation options.
Discrepancies that creep into loan files during bulk transfers may cause such avoidable foreclosure situations to occur, which could snowball into improper handling of borrowers’ loss mitigation requests. This is due to the fact that loan boarding is, as was previously mentioned, primarily a manual and paper-intensive process. Unintentional mistakes made during a manual onboarding procedure might lead to incorrect mortgage payment application, inaccurate borrower contact information, delayed interest rate adjustments, etc. Any of these occurrences could put the borrowers’ loss mitigation options in jeopardy, which could lead to financial hardship or, worse, foreclosure.
The CFPB has also established specific procedures to properly handle the transfer of borrower information and specifics of loss mitigation applications during bulk MSR transactions in order to reduce such systemic risks.
Mortgage servicers are under a lot of pressure to comply with the CFPB’s regulatory requirements as well as to onboard loan files more quickly. Once on board, servicers are required to handle borrower requests for loss mitigation in accordance with Regulation X, state laws, and GSEs-established guidelines. Therefore, any mistake in the upstream manual onboarding process has the potential to quickly escalate into numerous compliance issues and subpar service.
Automate First Strategy to speed up purchasing large quantities of MSR
As was mentioned above, a flood of incoming paperwork and data during bulk MSR transfers can stifle loan boarding teams’ ability to process information. Mortgage servicers could use a “automate first” strategy in three crucial areas to ensure frictionless onboarding:
Standardizing document and data transfer at scale: Downstream servicing issues may result from non-standardized loan file delivery from MSR purchases from various lenders. Therefore, one of the top priorities for mortgage servicers should be to standardize the delivery and receipt of loan file data and documents at scale during such transactions. This might entail integrating loan servicing software that can automatically ingest documents in a standardized format and is compatible with the lender’s LOS.
Version control and document indexing: During the loan boarding process, a substantial amount of the documents are printed out for making small corrections and omissions. However, it is tedious and time-consuming to sort through and rearrange hundreds of pages of loan files. Additionally, maintaining an accurate version control is challenging when updating paper documents manually. Mortgage servicers are consequently exposed to compliance risks brought on by breaking general transfer-related rules and procedures established by the CFPB. Service providers should automate their document processing workflow so that documents can be indexed and version control can be performed with a single click in order to reduce these risks.
Data extraction and validation: Data verification and discrepancy identification are both labor-intensive processes, just like document classification. Being 100% accurate while manually confirming the integrity of hundreds of data fields across multiple documents in bulk loan files is practically impossible. As a result, mortgage servicers should use automated data extraction software that can identify inconsistencies in values for fields with similar names across documents. The user will then have more time to concentrate solely on fixing the exceptions. This can change exception handling from being entirely manual to requiring only 10%–15% of documents to be reviewed manually.
In other words, deploying a powerful loan servicing software along with a document processing automation system would be necessary to automate the loan boarding process. Such a system is capable of automatically ingesting non-standard documents, classifying them into the appropriate document types, and extracting pertinent data to verify data integrity.
That’s it for this blog Real Estate Heaven Fan, remember to subscribe to our channel and if you feel like we’ve delivered value please share this blog with ONE person. That’s right just one person as a token of your appreciation for the hard work we put into making content that educates and helps you on your mission of building your own fortune. Remember you can read blog after blog, but it isn’t until you actually take action that you’ll start to see results. See you soon!