There is no shortage of rules, laws, acts, and regulations surrounding mortgage servicing. This wide range of compliance mandates – from the Real Estate Settlement Procedures Act (RESPA)/Regulation X and Truth in Lending Act (TILA)/Regulation Z to the Electronic Funds Transfer Act (EFTA)/ Regulation E and Fair Debt Collection Practices Act (FDCPA), to name a few – place a heavy burden on servicers when it comes to internal processes and operational efficiency.
Already inherently large and complex, mortgage servicing operations can be slowed by compliance requirements. This heavily regulated industry must continuously monitor, manage, and evaluate compliance across numerous requirements that are constantly evolving. One glance at the Fannie Mae website illustrates just how frequently servicer policy updates are issued, which is often several times each month. Tracking and adapting to new regulatory requirements is arduous, but non-compliance puts servicers at risk to face millions of dollars in fines and fees.
On top of that, servicers are balancing a higher volume of inquiries from borrowers in the wake of a higher interest rate environment and an economic downturn spurred by COVID-19. This article explores how artificial intelligence (AI) technology can simplify and hasten internal compliance processes while improving operational efficiency.
Manual Processes Are Major Pain Points
For mortgage servicers, for example, compliance management protocols rely heavily on manual processes, which drain internal resources. A survey of 300 loan servicing executives and managers asking about workflow pain points revealed that nearly a quarter (24%) said regulatory compliance management is heavily manual, and nine in 10 said they have implemented or plan to implement loan servicing workflow technology.
Manual processes drain internal resources. The shift toward automated workflow technology makes sense as manual processes are expensive.
The shift toward automated workflow technology makes sense as manual processes are expensive. According to the MBA’s Servicing Operations Study and Forum (SOSF), fully-loaded servicing costs tripled between 2008 and 2021, holding above $200 per loan since 2010 and the passage of the Dodd-Frank Act. In 2021, costs hit $240 per loan – increasing from $226 per loan the year prior. Corporate administration costs, which includes compliance, increased to $48 per loan in 2021 — accounting for 20% of total costs.
Agency guidelines that live as complex, long, tedious documents must be scanned for changes, and internal systems and communications must be adjusted to ensure the servicer remains compliant. Amidst increased regulatory scrutiny and internal compliance challenges, finding a way to streamline compliance and enhance operational efficiency can seem impossible.
The Complicated World of Compliance & Risk Management
Servicers face increasing operational costs in ensuring compliance with all servicing regulations. This burden can be made worse by outdated technology that inhibits fast and effective communication with internal stakeholders.
Servicers must also be mindful of risk management, which entails proper internal audits, processes, and operating procedures to address risks. Risk management teams must work hand-in-hand with quality control (QC) teams to effectively mitigate risks, and without updated technology, this can become overwhelming.
The burden of increasing operational costs in ensuring compliance with all servicing regulations is made worse by outdated technology.
As regulations and policies continue to evolve, internal compliance and risk management processes often become a tangled web, which creates more risk. Most recently, the Federal Housing Finance Agency (FHFA) announced enhancements to COVID-19-era deferral policies from Fannie Mae and Freddie Mac. These policies allowed homeowners experiencing hardship to defer up to six months of mortgage payments, which would be settled at the end of the loan as a non-interest-bearing balance. More than 1 million payment deferrals have been completed by lenders since the beginning of the pandemic. About one-third of homeowners exited forbearance via these payment deferrals since 2020.
Servicers are now dealing with these borrowers exiting forbearance programs, placing increased regulatory oversight on the industry, especially as some of these homeowners will be unable to bring mortgage payments current. This will trigger loss mitigation protocol and require mortgage servicers to offer options to these borrowers, with some that will inevitably enter the foreclosure process. All of this adds up to increased compliance requirements, processes, time, and costs.
Solving Compliance and Risk Management With AI
Natural Language Processing Simplifies Servicing Compliance from Start to Finish
Natural language processing (NLP) is a type of AI that enables computers to understand and translate human text similarly to how humans can. The benefit is a certainty that nothing is getting lost in translation – an essential certainty when it comes to compliance.
Our newest patent-pending solution, VAIDA™, is leveraging NLP to consume raw data in the form of servicing guidelines – whether in PDF, Word document, or image format – and translate it into a simplified process document. VAIDA uses NLP to digest the data, remove unimportant context, highlight keywords, and transform the information into noise-free condition- and action-based process flow charts. For servicers, the benefits are substantial.
- Reduced strain on resources – Remove reliance on manual processes and the resources required to execute with an automated system that enables resources to focus on high-value operational tasks.
- Automated compliance operations – Automate manual processes, removing the need for extra resources and reducing chokepoints and room for human error.
- Unparalleled insights – Gain the ability to drill down into servicing guidelines (Fannie Mae, Freddie Mac, CFPB, FHA, HUD, etc.) updates by section, topic, scope, business rules, and criteria – and view the number of loans in your portfolio that are affected by a rule and require a related action, with the ability to drill down to individual loan details.
- Seamless integration of updated guidelines– Run new rules across your portfolio with the ability to export results into a .xls or .csv file – or generate snippets of code to be added directly into your existing system.
How VAIDA Works
Harnessing the power of VAIDA is not rocket science for the user, though the behind-the-scenes lifting by AI is no simple feat. Users simply choose a file (like updated Fannie Mae or Freddie Mac guidelines) from their computer and run the system. The system consumes the data, conducting deep processing and creating data frames. It then extracts pertinent data, infers sentence structure, and creates simplified process documents for every single section within the chosen guidelines.
In addition to processing documents, the system generates accompanying flow charts, offering an easy-to-understand, visual element that includes rules, conditions, and actions available at a glance. The guidelines are broken down into these process documents for each section and topic, allowing the user to easily navigate the output.
Perhaps the most exciting part of VAIDA is the ability to generate a snippet of code for any section of the guidelines, which also encapsulates the rules, conditions, and actions. This code can be generated in Python or another language of choice, allowing servicers to easily plug it into their existing system and automate compliance according to updated guidelines.
VAIDA makes it easy to apply compliance updates across a portfolio of loans. Servicers gain visibility into how many and which loans meet conditions. Upon running the system on new guidelines, it will generate views into the guidelines that are easy to navigate. Each section of the guidelines can be drilled into topics within the selection. Each topic breaks down into scopes, and each scope drills down into conditions. This is where servicers can view which loans in the portfolio qualify under specific conditions.
Harnessing the power of VAIDA is not rocket science for the user, though the behind-the-scenes lifting by AI is no simple feat.
For example, all loans that meet the conditions to require sending a breach or acceleration letter would be marked as “true”, providing insights into the loans that require action (and which have actions that are either fulfilled or unfulfilled). The technology provides high-level dashboards showing the percentages of fulfilled versus unfulfilled actions for the overall portfolio, but users can also drill down into individual loan details.
With VAIDA, users can access multiple dashboard views that present a breakdown of the different actions required by each section of the guidelines, offering insights into the top ten actions. Filters enable users to drill down into more specific actions, like viewing the number of loans that require a breach or acceleration letter. Users can drill down to individual loan details, including the criteria that qualify the loan for a specific action. Rule descriptions dictate the appropriate action called for if all conditions are met. Servicers can “run” a specific rule on the portfolio, which will generate the corresponding loan numbers. This information can be exported and sent to the operations team to be added to its workflow.
Another key challenge VAIDA addresses is the issue of time. Updates to servicer guidelines are time-sensitive, adding pressure to not only remain compliant but to do so as quickly as possible. The rush to update compliance processes to meet evolving guidelines creates additional risk — something could be missed, mistakes might be made — that are only amplified by reliance on human effort and manual processes.
VAIDA reduces this risk, quickly processing data and providing fast implementation solutions. These real-time insights allow servicers to both track and speed up responses to regulatory changes via VAIDA’s relevant, meaningful, and potentially penalty-saving insights. It’s ability to extract the most important information streamlines change management and mitigates the chances of a rule violation — protecting not only the servicer, but all stakeholders, including investors and borrowers. VAIDA’s ability to track and flag issues in real-time with minimal human intervention helps servicers avoid violations and the associated fines, fees, and penalties.
It’s uncharted territory for mortgage servicers who lean heavily on manual processes that are prone to human error and that consume time and resources at lofty levels.
It’s Time for a New Path Forward
Complex and changing rules and regulations are a given in mortgage servicing. For compliance teams, this has been an ongoing pain point and, in many cases, generated unnecessary costs. Without the proper tools, many of these processes remain manual, requiring heavy manpower to make loss mitigation, default resolution, and other critical decisions. It’s inefficient, slow, and prone to human error.
Allowing tech to do the heavy lifting not only simplifies processes and reduces costs, but also allows human talent to focus on adding value to the customer experience.
Harnessing the power of AI can help servicers untangle the extraordinary complexity of industry requirements. As servicers continue to trudge the road toward digital transformation, some of these tools can integrate with existing systems, reducing the costs and headaches associated with “rip and replace” technology initiatives. Yes…it’s time.
Under increased regulatory scrutiny and facing resource constraints, servicers are well advised to seek out how AI-powered compliance technology can simplify complex compliance requirements. Allowing tech to do the heavy lifting not only simplifies processes and reduces costs, but also allows human talent to focus on adding value to the customer experience – a critical call to action for the industry as a whole.