UX case study · Fintech, digital lending
FinVista
An assisted lending app: sales executives run the whole journey, and the customer meets it only at the moments that need their own hand.
Outcome
What the work moved.
Project overview
One lending app for six products, run by executives, handed to the customer at the moments that need them
FinVista is the assisted lending app of one of India's largest NBFCs (client anonymized under NDA): 25 million+ customers, a loan book above Rs 30,000 crore, loans originated in 500+ cities. I led UX across all six products, with the two-wheeler flow as the proving ground.
TL;DR
Paper-heavy origination became an 11-step native Android journey run by sales, call centre and store staff, with designed handoff moments where the customer signs, consents and verifies in person. Six products on one 100+ component system. Processing ~60% faster, agent training down ~40%.
Senior UX Lead
5 months
2 designers, 1 PM, 8 devs
Native Android app
Figma, FigJam, JIRA
End-to-end LOS
How it ran
Four phases across five months.
Discovery
3 weeks
Research
3 weeks
Design
10 weeks
Validation
4 weeks
Product landscape
Six products, one system. The highest-volume one proved it.
Two-Wheeler Loan
New and used bikes, scooters, electric. This case study is this flow.
Consumer Durable · Personal · Used Car · Tractor · Three-Wheeler
Five more products on the same 100+ component system.
Who runs it
Three roles drive this app. None of them is the borrower.
Sales Centre Executives
Run loans over phone and app, a dozen applications at once, and own the compliance. Speed is the job.
Store Managers
Own the dealer portfolio, the pipeline and the escalations. They read the app more than they type into it.
The customer
Not a user. A participant, for the eight or so moments only they can complete: live photo, consents, OTP, uploads, e-signature.
The problem
One app, two personas, six products, and a gap from flagship phones to entry-level Android on patchy coverage
One scalable Android app, reusable across six products, holding up on every device and network in India.
The old flow was paper, agents and fragments. The new one had to be fast enough for twenty applications a day, instantly readable the second the phone turns around to face the customer, solid where the coverage is patchy, and RBI-compliant end to end.
Three dimensions
What made it hard.
Multi-Persona Complexity
Executives drive; the customer takes over for signatures and consents. Same screens, two completely different readers.
Information Overload
One application is 40+ fields. The IA does the heavy lifting.
Device and Network Diversity
Flagship phones on good coverage to entry-level Android where the signal drops. The app holds up across the gap.
Scope
What I took on
- End-to-end origination for 6 products
- Native Android design system (100+ components)
- Sales, call-centre and store-manager pathways
- KYC, identity verification, consent flows
Out of scope
- Backend credit scoring
- Payment gateway UI
- Admin and back-office
- Post-disbursal servicing
Teardown
What the best Indian lending apps already do, and the gap they all leave
I pulled apart the leading Indian lending apps and ran a heuristic sweep with stakeholders.
| Pattern | Navi | KreditBee | Bajaj Finserv | Home Credit | Muthoot |
|---|---|---|---|---|---|
| Contextual navigation | ✓ | ✓ | ✓ | - | - |
| One-step-at-a-time | ✓ | - | ✓ | ✓ | - |
| Micro-assistive text | ✓ | ✓ | - | ✓ | - |
| Progress tracking | - | ✓ | ✓ | - | ✓ |
| Help always available | ✓ | ✓ | ✓ | ✓ | - |
| Multi-product support | - | - | ✓ | - | - |
What the teardown settled
Every competitor nails a single-product flow. None solve multi-product, multi-persona. That gap is the opportunity.
Material Foundation
Material and card layouts. Android users already know them.
Clean and Minimal
One hierarchy. Matters most when a customer is holding a phone they have never seen.
Progress Always Visible
Everyone knows where they are and what comes next.
Two readers
Two readers, one journey
The executive and the customer read the same screens, so both had to be designed for at once.
Rahul Sharma
“Move fast, keep the customer calm, never miss a compliance step.”
Time
Many customers at once. Every extra tap costs him.
Continuity
A call picks up exactly where the app left off.
Ankita Agarwal
“Give me a loan I understand, on this phone, with no hidden charges.”
Trust
Cautious about financial data. Needs visible safeguards first.
Transparency
Clear rate, EMI, fees, total. No fine print surprises.
Non-negotiables
The three rules the design had to keep.
- Progressive Disclosure. One category of fields at a time. The load never spikes.
- Trust at Every Step. Visible security cues and a plain reason for every field.
- Two Readers, One Screen. Dense and fast for the executive; stripped back and plain at handoff.
The eleven steps
The Two-Wheeler loan journey in 11 steps
One screen, one job, progress always visible.
Personal Details
Identity plus live photo
PAN Verification
PAN or Form 60
Address Details
Pin code and landmark
Employment Details
Type and employer
Income Details
Applicant and household
Product Details
Vehicle, make, model, scheme
Schemes & VAS
Add-ons and insurance
Bank Account
UPI, account, mandate
Photo & Documents
Capture and upload
Notifications & Consent
E-consent preferences
Loan Summary
Review and submit
Options on the table
Four decisions that carried the design
Low-fi locked the IA, mid-fi set direction, hi-fi sharpened against real feedback.
The hardest screens in grayscale, options side by side.
Explorations
Flat vs 3D icons, input styles, layout density.
Four decisions
What the design actually rests on.
- No-Scroll Screen. One screen, one job. It mattered most for agents running 15 to 20 applications a day.
- Dual Progress. Journey-level percentage plus step-level indicator. Orientation without overwhelm.
- Assistive Onboarding. Every step opens with plain-language help, so the executive never translates the screen out loud.
- Pre-verified Data. Existing customers get pre-filled fields with checkmarks and one-tap confirm.
The library
A 100+ component library, one visual language across six products
Figma auto-layout, variants and tokens, so engineering handoff stayed clean.
What the library bought
New product journeys assembled from existing components in days, not weeks.
In the store
Three rounds that changed the handoff
No formal lab. I shared Figma prototypes with real users, watched, and refined. These are the three rounds that changed the design.
Internal stakeholders
Scrollable multi-action screens overwhelmed everyone on wireframes. I pivoted to one screen, one action.
Sales executives
Six executives on mid-fi. Nobody could find the way back to edit a completed step, so edit icons became persistent.
Customers at handoff
Customers signed without reading. I rewrote every handoff screen to one plain question plus what happens next.
Where it landed
A journey an executive can run at speed and hand over cleanly at every point that needs the customer.
What changed
From paper to digital
vs. paper flow
one design system
reusable library
for agents
Honestly
What worked, and what I would change.
What went well
- One screen, one action killed form fatigue for executives running back-to-back applications
- Progressive disclosure made the 11-step journey "not feel long"
- The early design system let later product journeys assemble in days
What I would do differently
- Bring sales executives in on round one. Their deepest insight only surfaced in round two
- Test on low-end Android devices sooner
- Design error states from the start, not at the end
The shipped app
Ten of the 46 screens
A grab-bag from across the two-wheeler journey, from login to disbursement.
Structure, flow and content are the product as it shipped. The visual layer is a Material 3 refresh in the current design language.
Thank you
Thanks for reading.
Happy to walk through any part of this in more detail, including the decisions that did not make it.
