Back to Work
UX Design · FinTech · 2026

Bank Loan Redesign

A high-abandonment loan application rebuilt from the ground up: reducing cognitive load, restoring confidence, and making financial commitment feel safe rather than stressful.

Role UX Lead
Year 2026
Tools Figma · UX Pilot · ChatGPT
Platform Web
✓
2
Income
3
4
5
Annual Income
$72,000
💡 Include salary, freelance income, and regular bonuses
Employment Type
Select type...
Progress auto-saved
/ The Problem

Users weren't confused.
They were scared.

The bank's loan application had a dirty secret: most people who started it never finished. Not because they didn't qualify — because the experience made them feel like they might be making a mistake.

High Mid Low Visit Basic Info Financials ↓ Documents ↓ Review Done Spike 1 Spike 2

User confidence levels across the original loan application flow. Two critical abandonment spikes identified at financial data entry and document submission.

When users hit the financial data section, abandonment spiked. When they reached document submission, it spiked again. Exit surveys surfaced words like "confusing", "stressful", "unclear what happens next." This was not a usability problem. It was a trust problem.

01
Uncertainty kills conversion

75% of drop-offs occurred not at difficult steps, but at ambiguous ones. Users didn't know how long the form was, what came next, or whether their answers were correct. Progress indicators alone reduced measured anxiety by answering "how much more is there?"

02
Jargon is a trust blocker

Terms like "gross annual income," "collateral type," and "origination fee" appeared without explanation. For first-time applicants, encountering undefined financial terms mid-form triggered abandonment: not because they didn't know their income, but because they weren't sure if they were answering the right question.

03
The stakes of the moment matter

Document submission wasn't just a file upload. For users, it was the point of no return: the moment they committed something real. The design needed to acknowledge that weight, not flatten it with a standard file picker.

/ Discovery

Understanding
the friction.

I used UX Pilot for AI-assisted wireframe exploration, Scite AI to validate research assumptions against academic literature on financial UX, and ChatGPT to synthesize findings into actionable design principles.

Research methodology
📊
Exit Survey Analysis
170+ responses coded
🗺
Journey Mapping
Current state + emotion overlay
📚
Academic Literature
Financial anxiety in UX via Scite AI
🤖
AI Wireframing
Rapid iteration via UX Pilot
🧠
Synthesis
Principles via ChatGPT + review

The current-state analysis mapped every screen against emotional response and cognitive load indicators. I built a persona: a first-time applicant, financially motivated but cautious, worried about making an error that cannot be undone. Every subsequent design decision was filtered through her lens.

Current-state journey map with emotional annotations
Current-state journey map annotated with emotional responses. Red zones mark the two primary abandonment spikes at financial data entry and document submission.
User persona and emotional journey map
User persona and emotional journey map. The vertical axis tracks confidence level, with sharp drops at the two critical touchpoints.
/ The Redesign

Five decisions that
changed everything.

The redesign wasn't about making the form prettier. It was about making every moment feel navigable, honest, and recoverable.

01
Progress visibility

A persistent progress indicator at the top of every step answered the question users were silently asking: "How much more is there?" Knowing the end is visible changes the emotional framing of the entire experience. Tested in isolation — it shifted the most-anxious-moment score from step 2 to step 4.

✓
✓
3
4
5
Step 3 of 540%
02
Plain-language microcopy

Every financial term gained a contextual helper in plain language. Not a legal definition: a human explanation. "What counts as income?" answered directly, in-context, before users had to wonder. The tooltip system expanded on demand and collapsed when focus moved, keeping the form clean by default.

Gross Annual Income
03
Grouped form sections

The original form was a continuous scroll of 40+ fields. The redesign chunked it into named sections (Your Income, Your Loan, Your Documents) — giving users a mental model of what belongs together and reducing perceived complexity. Users knew exactly where they were and what they'd completed.

04
Auto-save and resume

Life interrupts. A user who closes the tab and returns a day later should not start over. Auto-save with visible confirmation ("Your progress is saved") removed the fear of commitment to a single session. A simple return flow — enter email, resume exactly where you left — reduced restart abandonment by eliminating the sunk-cost panic.

05
Review before you commit

A full review screen before final submission — showing every answer, all editable — transformed submission from a leap of faith into a final check. Users who reached it went from the most anxious point to the most confident. The ability to go back removed the irreversibility that caused freezing.

Low-fidelity wireframes exploring form chunking
Low-fidelity wireframes exploring form chunking and progress architecture. Multiple variations tested before settling on the named-section model.
High-fidelity prototype screens
High-fidelity prototype screens. Persistent progress bar, plain-language field labels, and contextual tooltip system visible throughout.
/ Before vs. After

What the
transformation looks like.

The same step. The same data. Completely different emotional experience.

LOAN APPLICATION — STEP 2 OF 7
Gross Annual Income (Required)
—
⚠ This field is required. Please enter a value.
Employment Status
Select employment status
Collateral Type
—
⚠ Invalid selection.
Origination Fee Preference
Select option
Co-Applicant Gross Income (if applicable)
—
Tax Filing Status
Select one
↓ Scroll for 14 more fields
✓
2
3
4
5
💰
Your Income
Annual Income
$72,000
💡
Include salary, regular bonuses, and freelance income. Exclude one-time payments.
Employment Type
Full-time employee
Progress saved automatically
/ Outcomes

Clarity is
measurable.

The redesigned flow was validated through usability testing focused on task completion, navigation errors, and SUS score. Every friction point identified in research was directly addressed by a specific design intervention, then tested independently.

0
Critical drop-off points resolved

Each of the five identified friction moments received a specific, tested design response: not a general polish, but a targeted intervention for that exact form of anxiety.

0
Form fields chunked into named sections

40+ fields reorganized into five named sections with clear completion states. Users reported the form felt "much shorter" despite identical field count.

0
AI tools accelerating research and ideation

UX Pilot, Scite AI, and ChatGPT reduced research synthesis time and allowed faster iteration across multiple wireframe directions in the same sprint.

↑ SUS
System Usability Scale score in testing

Task completion rate and SUS score both improved in usability testing. More critically, the qualitative shift: "confusing" and "stressful" replaced by "straightforward" and "I knew what to do."

"It felt like someone actually thought about what it's like to apply for a loan for the first time."

Usability test participant, Session 3

The most important shift wasn't in a metric. It was in the language users used to describe the redesigned experience. Words like "stressful" and "confusing" were replaced with "straightforward" and "I knew what to do." That is the real outcome of a trust-centered redesign.