How to determine what to automate in mortgage

How to Determine What to Automate in Mortgage


As mortgage businesses face a volatile market environment in 2022, it is time to reassess automation priorities. According to research, interest on 30-year fixed-rate loans could reach 4% by the end of 2022 (lower than historic levels but higher than the last 18 months), which could dent demand for refinancing. On the other hand, the housing market is booming, and mortgage providers can look forward to a 9% rise in purchase mortgage originations. This will take place in the backdrop of a complex regulatory climate that seeks to mediate housing supply to all groups, without excessively burdening lenders. Inevitably, mortgage businesses will find themselves spending more on human effort and manual processes as they grapple with this new reality. Mortgage process automation will be a vital enabler – a business essential and not a nice-to-have, that drives efficiency while managing overheads.

Gauging Your Mortgage Automation Readiness

Mortgage process automation entails the use of technologies like robotic process automation (RPA) scripts, software bots, and artificial intelligence (AI) processes to replace or reduce the human interventions needed to perform key tasks.

According to Fannie Mae’s July 2021 survey, streamlining business processes is now the no.1 priority for mortgage providers, ahead of compliance, cost-cutting, marketing, and introducing new products and services. Automation is among the best ways to remove bottlenecks in mortgage, unlock efficiencies, and make processes more scalable in order to nimbly adapt to more (or less) than predicted volumes without incurring costs. However, the businesses’ readiness for automation will vary, and much of its success relies on how accurately one can gauge the baseline.

Mortgage automation readiness refers to a state of the backend, middle office, and frontend operations that is conducive to the use of automation. For instance, if a lender is still operating on legacy data formats and different data types for different functions, it will be extremely difficult to migrate the process to an automated workflow. A lender’s readiness also depends on the work culture and employee skill levels, which will determine whether they are able to use the automation effectively.

Without adequate readiness, real adoption will remain low. For instance, research indicates that 43% of lenders are equipped with automated loan e-closing technology, but only 12% of this group use it in a meaningful way.

The Starting Line: Deciding What to Automate in Mortgage

Companies with some degree of digital maturity and cultural readiness can progress to the next step – determining which mortgage process to automate. This decision will depend on the following factors:

  1. Volume and frequency – Mortgage process automation is ideal for high-volume, high-frequency tasks. It is by automating processes like data entry, common customer interactions, document reviews, etc., that lenders can accrue the most ROI. If one automates a task that does not occur too often, the automation solution will be left sitting idle for the majority of the time, taking away from your investments.
  2. Existing digital maturity – Mortgage processes with some degree of digital maturity are the prime candidates for automation. For example, if a lender has recently upgraded from paper-based document submissions to PDF and digital formats, then automating the process is the next logical step. An automated solution could use AI techniques like optical character recognition (OCR) to quickly extract data from PDFs and store it in a centralized, standardized repository – triggered by origination events to eliminate the need for human intervention.
  3. Process variability – There are typically two types of processes in a financial service organization – high-variability processes that change significantly from one instance to another and low-variability processes that are iterative and happen with very rare exceptions. Handling problematic customers is a high-variability process, which is unpredictable and is best performed by a human professional. Low-variability processes like reviewing multiple documents for the same set of data fields and compliance checks can be easily automated.
  4. The presence of exceptions – Process exceptions can either be predictable or unpredictable. For instance, loan documents submitted at the time of origination typically follow the same template. But once in a while, the lender may have to accommodate a new template as per the scenario or regulatory change. This is an exception that automation tools can easily handle through machine learning (ML). A human in the loop may register the exception and enter the new template for the first time, and the tool will remember the decision for future instances. Unpredictable process exceptions, on the other hand, are unsuitable for automation.
  5. Processes related to data – Mortgage processes and tasks related to data usually fulfill all the parameters needed for automation readiness. Whether it is data analysis to calculate borrower risk or generating reports for compliance and audits, these processes are frequent, high in volume, low-variability in nature, and have predictable exceptions. Automation also helps to consolidate data spread across different mortgage systems in a single environment for 360-degree intelligence.

Assessing Your Mortgage Automation Priorities

While it is tempting to choose a standalone solution that automates a point process (i.e., a chatbot that answers customer FAQs), this approach is not sustainable and will hold back adoption. It is more advisable to conduct a holistic assessment, gauge one’s automation readiness, set up robust digital systems, and integrate the automated workflows for maximum value.

For example, conversational AI that can fetch customer data and respond with dynamically updated information is far more useful than an FAQ bot. The journey must begin from low-hanging fruits evaluated on the five parameters mentioned and gradually progress towards a wholly digital and automaton-ready landscape.

At Nexval, our team of 1000+ subject matter experts combines mortgage industry knowledge with cutting-edge digital innovation. We provide lenders with 40+ automation tools and function-specific bots that are powered by AI, ML, and RPA to unlock maximum value from your automation investments.

Ready to commence your automation journey? Talk to our mortgage Tech Gurus to determine your best-fit starting point.

Anwesha Basu

About the author

{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}