Understanding Operational Workflow to Inform the New Design of Internal Loan Management Platform

The Business Challenge

The existing internal platform supported critical operational processes such as reviewing loan applications, verifying documentation, and managing customer accounts. Over time, the system had grown into a complex environment with multiple interconnected modules.

As the organization prepared to transition to a new platform, product and engineering teams needed to understand:

• how different operational roles navigated the existing system

• which workflows were most critical to daily operations

• where users experienced friction or inefficiencies

• what capabilities should be improved in the future platform

However, this knowledge was largely undocumented and scattered across teams.

 

 

 

 

 

 

 

 

 

 

 

 

Research Goals:

1. How do internal users navigate the current system during key operational workflows?

2. What pain points and workarounds exist in the current experience?

3. What opportunities should be prioritized in the next-generation platform?

​

Research Strategy:

Given the complexity of the system and the diversity of user roles, a mixed-methods research approach was designed to capture both qualitative insights and behavioral patterns.

​

1. User Interviews and Workflow Discovery

Goal: To understand how internal operational roles navigate system complexity to process and service loans.

I conducted in-depth context inquiries and workflow discovery interviews with cross-functional operational roles. My analysis moved beyond task-mapping to uncover the cognitive load and "hidden efficiency taxes" created by the existing platform.

Key Findings:

• Fragmented Information Architecture: Users suffer from a "multiple window penalty." Locating comprehensive case data for a single application or inquiry requires redundant navigation, forcing users to mental-model the connections between disparate modules.

• Operational Workaround Culture: To manage complexity, users have developed undocumented, ad-hoc workflows and external documentation trackers, signaling significant workflow friction and efficiency gaps.

• System/Reality Conflict: Major friction occurs around data latency and status conflicts. The system cannot accurately reflect the real-time truth of a case (especially in high-volume call scenarios), hindering decisive decision-making.

​

​

​

​

​

​

​

​

​​

​

​

​

​

​

​

​

​

​

​

​

2. Critical Workflow Mapping & Analysis

Goal: To translate interview narratives into a quantifiable, technical model of the existing operation, identifying where systemic complexity directly creates human error and operational latency.

Based on context inquiries and system walkthroughs, I created a cross-functional workflow map (Service Blueprint) that modeled the full user path across the legacy system's siloed modules. This visualization moved beyond a linear path to quantify the complexity of an ad-hoc process, exposing the "shadow workflows" that users created to manage platform limitations.

Friction Analysis:

1. Redundant Data Retrieval & High Cognitive Load: Users suffered from "Context Fragmentation," requiring navigation across up to seven disparate screen modules to access core data for a single loan (e.g., separating applicant ID, income verification, and collateral data). This creates a recursive memory tax and multiple vector paths for error.

2. Workflow Stalling & Verification Chokepoints: The map identified non-value-added steps, specifically repetitive verification locks. Users were forced into manual, multi-click confirmation cycles for automated data (e.g., verifying a clear credit report) before they could proceed.

3. Data Latency & "Status Blindness": Major friction occurred at the intersection of "Real-World State" vs. "System-of-Record State." The workflow exposed instances where the platform could not accurately reflect recent case activity, leading to "Status Conflicts"—particularly during simultaneous review scenarios.

​

​​

​

​

​

​

​

​

​

​

​

​

​

​

​​

​

​

​

​

​

​

​

​

​

​

3. Persona & Mental Model Mapping

Goal: To move beyond generic "user" categories and define the distinct behavioral archetypes, mental models, and specialized needs of the internal operational staff. This ensures the future platform supports the specific cognitive load and decision-making requirements of each core role.

What I Did:

• Behavioral Synthesis: Synthesized qualitative data from the initial 1-on-1 interviews to identify recurring patterns in task prioritization and system interaction.

• Mental Model Auditing: Analyzed how different roles "visualize" a loan case—distinguishing between the transaction-focused view of a Loan Officer and the risk-focused view of a Compliance Agent.

• Impact-Needs Matrix: Developed a matrix to cross-reference each persona's "critical daily outcomes" against the platform friction points identified in the workflow mapping phase.

Key Findings:

• The "Context-Switcher" (e.g., Loan Officer): Primarily motivated by speed and case-volume. Their mental model is linear, yet they are the most impacted by the "Multiple Window Penalty." Their primary need is a Unified Case Summary.

• The "Deep-Diver" (e.g., Compliance/Underwriter): Motivated by accuracy and risk mitigation. They don't mind complexity but are stalled by "Status Blindness." Their primary need is Real-Time Data Integrity and Audit Trails.

• The "High-Volume Responder" (e.g., Call Center/Customer Support): Motivated by immediate inquiry resolution. They suffer the most from "System-of-Record Latency." Their primary need is Instant Status Clarity.

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​​

​

​

​

 

4. Card Sorting & IA Reconstruction

Goal: To move from a siloed, module-based architecture to a user-centric information structure.

What I Did:

• Hybrid Card Sort: 15+ participants across core personas organized 40+ key data points.

• Statistical Analysis: Used cluster analysis and dendrograms to identify the strongest mental groupings.

Key Findings:

• Lifecycle Pivot: Users naturally grouped data by State (e.g., "Active Review") rather than legacy technical modules.

• Critical Persistence: High-risk indicators (Fraud, Status Conflicts) were identified as needing persistent visibility across all views.

• Taxonomy Alignment: Standardized naming conventions to match operational vocabulary, reducing "search-and-find" latency.​

​

5: Usability Testing & Iterative Validation

Goal: To evaluate the cognitive fit of early design concepts against real-world operational workflows. The focus was on validating whether the new "Case-Centric Dashboard" successfully reduced the "multiple window penalty" and improved decision-making speed for high-stakes tasks.

What I Did:

• Moderated Task-Based Testing: Conducted sessions with 12 participants across core personas using low-to-mid-fidelity interactive prototypes.

• Cognitive Walkthroughs: Observed users as they performed complex, multi-step scenarios, such as "Resolving a Status Conflict during a live call center inquiry."

• RITE Method (Rapid Iterative Testing and Evaluation): Worked closely with designers to make "just-in-time" adjustments to the prototype between testing rounds to immediately test potential solutions.

Key Insights:

• Confirmation of Unified View: Users reported a significant reduction in perceived cognitive load when data points previously stored in separate modules (e.g., Identity and Income) were presented in a single, high-density summary view.

• The "Actionable Navigation" Gap: Early concepts used a standard sidebar; however, testing revealed that users needed "State-Aware" navigation that automatically surfaced the next logical step in the loan lifecycle.

• Success of Persistence: The "Persistent Risk Banner" (for Fraud and Status Conflicts) was validated as a critical safety feature, preventing users from proceeding with "dirty data."

 

Research Impact:

This research initiative transformed a decade of undocumented "tribal knowledge" into a validated, scalable framework for the organization’s next-generation operational platform. The study achieved the following:

1. Architectural De-Risking

By pivoting the platform from a Module-Centric to a Case-Centric Information Architecture (IA), we eliminated the "multiple window penalty." This shift reduced redundant navigation by an estimated 60%, directly lowering the cognitive load for high-stakes decision-makers.

2. Operational Efficiency & Accuracy

The identification and resolution of "Status Blindness" and "Data Latency Conflicts" provided users with real-time truth for the first time. This minimized the risk of processing "dirty data," significantly reducing the vector for human error in loan servicing and call center inquiries.

3. Cross-Functional Strategic Alignment

The Workflow Maps and Persona Archetypes became the organization’s "single source of truth." These artifacts unified Product, Engineering, and Design teams, ensuring that every engineering hour spent on the new platform was anchored in a validated user need rather than technical assumptions.

4. Inclusive & Scalable Design

By prioritizing Inclusive Design and accessibility testing during early usability sessions, we ensured the future platform is equitable and resilient, capable of supporting a diverse, global workforce across multiple operational domains.