The Statement Builder is a core tool used by credit analysts to create standardized financial statements and forecasts. Despite strong customer demand, only 16% of sessions resulted in a completed statement, largely due to an interface that combined multiple workflows into a single experience. I led a three-week redesign to restructure the experience around goal-based workflows, increasing success rates by over 3x and halving task completion time.
Lead UX Designer
Research, prototyping, UI design
Expanding and maintaining our design system
Research & Design: 2 weeks
Development: 1 week
Increased success rates over 3x
Increased user sessions 86.56%
2x faster completion
As a credit analyst, I need an easy way to transform financial data across periods and model future performance so that I can evaluate borrower health, assess credit risk, and make informed lending decisions.
Despite strong demand from customers, the statement builder had extremely low success rates.
Only 16% of sessions resulted in new statements generated.To understand why analysts struggled with the tool, I reviewed FullStory recordings of users interacting with the feature. These recordings revealed consistent confusion about where to begin and frequent abandonment before completing the workflow.
Original UI
.png)
Every user: Where do I start?
Research sessions with internal subject matter experts revealed that users approach the statement builder with specific goals in mind. Additionally, customers had new goals for the feature that weren’t currently supported, indicating the need to expand functionality to better meet these objectives.
Key Insight: The problem wasn't just complexity - it was also orientation. By combining multiple workflows into a single interface, users were often unsure where to start and had to mentally map the system to their goal before making progress.
Guided by this insight, I explored three potential approaches:
Pros: Provides a clear entry point, scalable architecture, flexible UI, and implementable within time constraints
Cons: Requires manual input from analysts
Pros: Potentially minimal engineering work
Cons: Still mixes multiple workflows, increasing cognitive load
Pros: Automates part of the process based on configuration
Cons: Highly complex, less flexible, and requires a bigger engineering lift
The redesigned experience introduces a goal-selection entry point. Selecting a goal determines the structure of the workflow, including available inputs, validation logic, and how data is transformed. This allows each flow to be optimized for a specific task rather than generalized across many. Benefits:
Clear starting point: Instead of forcing multiple workflows into one screen, users now start by choosing their goal.
Reduced cognitive load: Each flow surfaces relevant options and contextual cues so analysts can immediately recognize what to do next, eliminating the need for users to interpret how the interface maps to their task.
Scalable structure for new statement types: Working closely with engineering, we built and shipped the first two redesigned flows in one week, gathered customer feedback, refined them, and then expanded with additional flows.
New Goal Selection UI
.png)
Dedicated UI per goal
.png)
This redesign transformed the Statement Builder from an underperforming feature into a scalable tool that analysts could reliably use in production workflows
The increase in success rates shows that users were able to complete workflows that previously led to abandonment, while the growth in sessions indicates the feature became reliable enough to support repeated, production-level use.
16% -> 70%
Success rate increase
+86.56%
User session growth
2x Faster
Completion time halved
+2 New Flows
New capabilities added
Automation opportunities: Early exploration with AI shows promise for efficiency gains, but accuracy remains a top concern.
Usability testing: Success rates improved significantly, but more testing can uncover how to increase it further.
Technical Evolution: Future releases can evolve the validation system to improve usability (completed).
