Industry:
finance
What are you looking for?
Every organization wants to harness the power of digital transformation to make their business more efficient and profitable, but this is often easier said than done. Even deciding on the best or most strategic place to begin this transformative work can feel daunting. That’s why Vishal Chawla, CTO at BSI Financial, focused on combining digital and operational transformation to support critical needs within the mortgage servicing company.
“When we started on our journey two years ago, we thought ‘if we create the most fantastic machines, we’ll see improvements’,” he explained. “But that isn’t the reality.”
Chawla quickly realized that BSI Financial’s back-office operated multiple enormous, manual processes, whose repetitive natures could use enhancement. However, the operational process design needed to be in lockstep with the tech improvements, or the changes wouldn’t be sustainable or scalable. The goal was to be able to have processes run in a predictable, efficient manner, while introducing transparency into the process to support accuracy and scalability.
“We decided to approach process optimization through a ‘crawl-walk-run’ approach,” Chawla said.
This approach consisted of three steps:
With this framework in mind, Vishal and his team examined hundreds of processes in BSI Financial’s environment and looked at how they could best orchestrate these processes. In some cases, they were looking to automate an activity for efficiency, while in others they hoped to orchestrate with a specific impact on ROI. Their first test case ended up being a process where orchestration helped deliver both: loan onboarding.
As a mortgage service company, BSI Financial plays in a complicated ecosystem with many stakeholders, including borrowers, asset sellers (both the original sellers and other brokerage firms), and investors. Layer in the many regulations that affect financial services and the housing market in particular, and every activity or business process becomes even more complex.
Mortgage servicers have to get things right to protect everyone’s interests, but the environment is constantly changing with new requirements, regulations, and massive amounts of data. The opportunity for errors is high—as are the ramifications for everyone involved.
Nowhere is this dichotomy more visible than in the loan onboarding process. When BSI Financial purchases home loans from sellers such as brokers and takes on the servicing of the loan, it requires the transfer of enormous amounts of data and documentation to support the transfer and define the original terms of the loan. This is the first step in the loan lifecycle at BSI Financial, and the most critical place to reduce errors. If errors aren’t caught at this stage of the process, the snowball effect of those errors results in servicing errors that are costly for BSI Financial to address downstream.
Unfortunately, there are multiple opportunities to introduce errors into the loan onboarding process. When the seller transfers the loan to BSI Financial, they are responsible for sending the pertinent data and documents. BSI has to match it all up in their own systems, and there will be mismatches in the system, especially if the loan has been serviced by the seller for a long time. Other opportunities for errors include duplicative data, missing data, additional data mapping mismatches, image-to-data extraction errors, dataset merge and service system upload errors.
“These errors are expected in the industry, but the scale and type of error will determine your fate,” Chawla confided. Prior to optimization, the process involved manual data entry and verification, which was time-consuming, error prone, and repetitive; it was also tough to quantify any meaningful error reductions.
Chawla and his team wanted to improve this process and remove the manual data handling. They decided to focus on error opportunities where the business wasn’t consistently capturing the data they needed, hoping to improve the accuracy level and overall efficiency of the process. Addressing this one process would have a meaningful impact on the business.
“This isn’t a onesie-twosie process,” explained Chawla. “We have so many instances or transfer running per day, simultaneously.”
With the key process identified, the team turned to process redesign and orchestration to first automate the most repetitive tasks to reduce errors and increase efficiency, then to find and automate all human tasks in the process that did not explicitly require human intervention or intelligence. They wanted to maximize the opportunities to reduce errors as far upstream as possible, where error reduction is less costly—both in resources and in reputation—to rectify.
Chawla and his team knew what they wanted to accomplish; to measure their success, they set the following KPIs:
As the team approached this project, they kept in mind their digital transformation ethos—operations had to be in lockstep with tech improvements, with an emphasis on transparency and efficiency. To reach those goals, the team realized they needed to focus on a process orchestration solution, and one that could support their visibility and transparency goals through BPMN.
As BSI Financial kicked off their transformation activities and focused on their loan origination process in late 2022, the team began to investigate process automation tools but weren’t impressed with the peer reviews that Chawla collected. Then he heard about Camunda.
Chawla has a strong product engineering experience and brings a cross-functional view to problem-solving. When he first discovered Camunda, he decided to experiment with the open-source version of the product and designed a small automation experiment with BSI Financial’s tax vendor to streamline interaction between the companies’ two systems, people and processes. The results, which was a fast and flexible automation, gave Chawla the confidence to choose Camunda to support the organization’s major process orchestration activities, even as there was internal pressure to instead leverage legacy applications and approaches.
The BPMN modeling support provided by Modeler was a major factor in the decision to choose Camunda, as it directly supported the project’s goals to increase transparency and efficiency into the process, and to align operational and technological improvements.
“We went in cold to BPMN, but it was easy to learn within the platform with the documentation provided,” Chawla said.
Other factors that made Camunda the right choice for BSI Financial were the ease of use and the flexibility that the system provided Chawla’s team.
The orchestrated loan origination process has now been in production at BSI Financial for around a year and is already delivering on its identified goals.
Prior to Camunda, it took an average of 4-5 days to onboard a loan from when the new loan documentation was sent to BSI Financial. While that timeframe was within the team’s SLA, they wanted to improve that metric and free up their human resources to concentrate on higher-value activities.
With the new orchestrated process leveraging Camunda, onboarding now takes two days or less. Additionally, secondary latency (reviewing the data collected to ensure it is complete) happens almost instantaneously, with very little manual validation.
The original process was a completely manual effort, tying up multiple employees’ time. Now, 60% of overall volume runs straight through, and around 50% of transfers are completed with straight-through processing. Those that do require manual intervention are due to outside factors beyond BSI Financial’s control; this circumstance now puts the onus on the loan seller to be proactive and get the team the data they need more quickly.
Aside from an increase in efficiency through onboarding time, the orchestrated process is supporting risk reduction efforts by reducing critical errors earlier in the loan lifecycle, stopping problems from compounding further downstream. Servicing errors are now down to <5% prior to implementing Camunda. Credit also goes to other improvements where BSI indexes and extracts data from provided collateral documents using advanced machine learning to improve accuracy.
Critically, this improved process is now delivering cost savings for the organization as well. There is approximately a 35% reduction in the cost required to onboard a loan.
As Chawla explained, “We always wanted to answer the question, ‘How do I reduce cost while still reducing risk?’”
The employees who onboard BSI’s new loans are also delighted with the process improvements. Previously they were constantly chasing paper through spreadsheets and emails, but now they have a purpose-built portal with exception-handling tools already built into the process (the “walk” phase of the crawl-walk-run of the transformation framework).
“Their world has completely transformed, and it’s changed the narrative at the company as well,” stated Chawla.
Thanks to their transformed loan onboarding process, the team is delivering cost savings, risk reduction, and increased operational efficiency to the business- and they aren’t done yet.
The success implementation has opened the floodgates for more process orchestration opportunities as the team turns to broader IT initiatives.
The team is investigating several potential processes improvements they can orchestrate with Camunda, some tactical, some operational, and some strategic. BSI puts these initiatives on a continuum when it comes to complexity, and is currently investigating additional automation opportunities, particularly for back-office activities and introduction of AI agents in reasoning and decision-making.
“The act of orchestration forces the team to re-evaluate how the process is actually organized—it’s making us break down organizational silos and remove duplicative work across departments,” Chawla stated. “We’re continually looking for opportunities to ‘BPMN-enable’ our processes. We generate Level 1 and 2 BPMN diagrams from our operations manuals using AI – this is amazing from a change management perspective as operational changes are easily cascaded to BPMN and implementation.”
The team is also using discovery tools in their environment to hook into processes to trigger downstream actions and orchestrate multiple processes triggered by those events; for example, a customer requesting documentation or support in the system can now trigger multiple processes and workflows kicked off automatically, making support faster and easier for the customer.
More business processes are also being orchestrated with Camunda, such as the payoff request quote process, where consumers who are selling their homes need to pay off the existing mortgage as part of the sale process.
There is also now a separate roadmap for continuous improvements to be made to the loan onboarding process, with a goal of 100% pipeline for service onboarding. The team is also looking into implementing decision automation into multiple processes as well to further scale their process orchestration adoption and further reduce unnecessary human intervention.
Throughout this initial transformation journey at BSI Financial, Chawla has been impressed with his team’s ability to learn and adapt, and credits Camunda in part for supporting those efforts.
“We had no background in Camunda and BPMN,” he explained. “But now, my group has enough expertise to tackle anything Camunda-related.”
He stresses to others who are looking at their own process orchestration or digital transformation efforts to remember that it’s an adoption process, and to be successful you must learn as you go. For BSI Financial, they were clear on the metrics they wanted to influence but needed to home in on the right levels of complexity and areas of their process ecosystem to focus on to drive those results and provide ROI to the business.
“When the team can watch a live dashboard and see how their efforts are literally moving the needle for the company, they feel encouraged that their process orchestration efforts are working—supporting not just one team or department, but the entire business.”
Camunda © 2024
