How field research into loan-officer workflows shaped a faster way to compare, price, and share mortgage options with borrowers.
Role Product Designer, partnered with lead Experience Designer, Reported to VP of Product
Skills User research & interviews, value proposition design, service blueprinting, user profiles, personas, user flows, user journeys, sketching, wireframes, prototypes
Skills User research & interviews, value proposition design, service blueprinting, user profiles, personas, user flows, user journeys, sketching, wireframes, prototypes
As the Senior Product Designer for this project, I partnered with Product and Engineering leads and a Mortgage Advisory Council to research loan-officer workflows, synthesize our findings, and translate them into the Product Selector experience. The resulting feature gave loan officers a faster way to move from borrower information to product comparison and sharing, helping reduce the average loan-closing timeline from several weeks to 7–10 days and contributing to 30% to 100%+ increases in loan conversions for customers within 1 quarter.
The problem: fragmented workflows
Our on-site research showed that loan officers, processors, and underwriters were relying on multiple tools to gather borrower information, qualify applicants, look up loan products, and calculate rates and pricing. We also learned that product selection was not just a data problem: conversations with borrowers about their goals, timing, and financial situation played an important role in determining which mortgage options made sense.
Researching and designing with mortgage experts
Our research combined on-site interviews with loan officers, managers, and assistants with ongoing feedback from a Mortgage Advisory Council of subject-matter experts. The council included client administrators who configured Roostify for their loan-officer teams as well as Roostify employees with previous loan-officer experience. I helped synthesize the research into customer profiles, value maps, conversation scenarios, and service blueprints, then used the council’s domain expertise to continuously evaluate and refine the prototype as the workflow evolved.
(Above) Early research notes: Pains, gains, and jobs to be done for loan officers, processors, & borrowers.
Shopping user flow service blueprint
Journey map for various users
Finding the minimum information needed
One of the most important design questions was determining how much information a loan officer needed at each stage of the conversation. I created scenarios and touchpoint diagrams to identify where borrowers could provide just enough information for loan officers to calculate several likely options, then organized the required loan criteria into logical groups for the product-search workflow.
(Above) Wireframes showing required loan criteria inputs.
Turning research into a faster workflow
Research led to several decisions that reduced unnecessary work: borrower information already available in Roostify could be pre-filled, forms could be generated dynamically based on the scenario, and loan officers could begin a product search directly from a borrower’s pipeline record. I also separated inputs from results so loan officers could focus first on the criteria needed for the search, then compare selectable products and expand rows for more detailed fee information.
Supporting the loan officer–borrower conversation
The experience was designed to support the loan officer’s role as a trusted advisor, not simply return a list of rates. Loan officers could compare products, review monthly payments and closing costs, generate fee worksheets, and share personalized options with borrowers so they had clearer information to support an evolving financial decision.
A mockup of the loan officer leads pipeline showing how a search can be initiated for a prospective customer.
The product search input form generated from a leads list would prepopulate with the customer's information.
What I learned
This project reinforced that improving a complex B2B workflow often starts by understanding where information already exists, when people actually need it, and how it moves between roles and systems. Rather than simply digitizing the existing process, we used research and service design to identify opportunities to remove repeated work, simplify decision points, and better support the human conversations that were already central to the mortgage process.
