Full case study coming soon

perks

A white-label cashback benefit for employers.

Shop the brands you already use, cashback shows up. Zomato is the skin. The same product re-skins to any employer.

Role

Lead Product Designer

Scope

0 to 1 · Marketing site and member portal

Timeline

15 active build days

screens shipped
across two surfaces

brand detail pages
with a live rate

1,500+

brands in the
catalogue

Summary

What it is, and what I did

The product

An employee cashback benefit

Shop the brands you already use, cashback shows up. The employer funds it, so it costs the employee nothing.

My role

Sole designer, end to end

Brand, design system, every screen on both surfaces, and the quality bar the build was held to.

Scope

Two surfaces, 71 screens

A marketing site and a member portal: six shopping categories, a wallet, a profile and a help centre.

RoleLead Product Designer, CashKaro. Sole designer on this project

Timeline1 to 22 July 2026 · 15 active build days

TeamMe on design and quality, one engineer on the member portal

I ownedPositioning and voice · brand identity · design system · every screen · motion · copy · quality review

DeliveredMarketing site · Member portal · 13 chapter brand book · Invite email

Screens

The product

Eleven screens, one scroll. Keep going and the pages walk past, or jump with the chips. See it live opens the real page in a new tab.

See it live

Mobile

On the phone

Not a reflow. The phone gets its own layouts, drawn for one thumb and one column.

Marketing site on a phone: the salary hero with fanned category cards — **Marketing**Fanned static cards instead of draggable ones. Same argument, one thumb.

**Marketing**Fanned static cards instead of draggable ones. Same argument, one thumb.

Portal home on a phone: welcome Angad, a search pill and a category grid — **Home**Opens with a question, not a dashboard. Search first, categories second.

**Home**Opens with a question, not a dashboard. Search first, categories second.

**Shopping**Every gutter equals every gap: 24px, one rule across the portal.

Credit cards on a phone: a watercolour hero and card comparison list — **Cards**Filters move into a sheet, so the list keeps the full width.

**Cards**Filters move into a sheet, so the list keeps the full width.

Flights and hotels on a phone: tabs and partner tiles — **Travel**Flat rupee amounts, not percentages. Travel pays differently.

**Travel**Flat rupee amounts, not percentages. Travel pays differently.

Luxury on a phone: a jewellery hall hero and occasion cards — **Luxury**Shop by occasion. Nobody browses luxury by product type.

**Luxury**Shop by occasion. Nobody browses luxury by product type.

Loans on a phone: a loan type toggle, an amount slider and lender cards — **Loans**One thumb sets the amount, the lenders reorder underneath.

**Loans**One thumb sets the amount, the lenders reorder underneath.

Insurance on a phone: cover chips and insurer plan cards — **Insurance**Cover chips instead of a quote form. No jargon.

**Insurance**Cover chips instead of a quote form. No jargon.

Wallet on a phone: all-time earnings on a red passbook card — **Wallet**Where is my money, answered in one card.

**Wallet**Where is my money, answered in one card.

**Profile**The member’s corner. Deliberately not a tab, so the nav never lies.

Help centre on a phone: an ask pill and topic tiles — **Help**Ask first. The assistant reads live data, so it cannot invent a rate.

**Help**Ask first. The assistant reads live data, so it cannot invent a rate.

The system

A system built to wear other brands

The product is white label, so the design system had to hold its shape while taking on someone else’s language. Thirteen chapters covering colour, type, motion, icons, numbers and voice, each with the reasoning attached.

Open See it as in the page below and switch partner. The layouts, spacing and behaviour stay exactly where they are; only the brand changes. Every colour is derived from one hue, so adopting a new one is a token change rather than a redesign.

/branding

Process

Building it with Claude

Seventy-one screens in fifteen days is an unusual pace for a single designer. I set the direction and Claude handled the implementation.

Every screen started as a Figma frame plus a short written brief
The brief covered what the screen had to achieve and what it should avoid
Build, review, revise. Some screens landed first time, the login took eight
The judgement stayed with me: which objection a section answers, whether a figure is honest, when an interaction costs more than it returns

Documentation

A written decision log

Every decision recorded with the reason behind it
What was tried first, and why it was replaced
About 3,300 lines by the end
No question reopened twice

Quality assurance

Testing what a screenshot cannot show

Most states only appear while scrolling
Scripts scroll through, trigger each state and capture it
Four rounds of performance work came out of it
Emulation reported no problems, a real handset disagreed

Editing

Knowing what to remove

Signup bonus, star rating, activity ticker: built, then cut
A placeholder success screen: cut
Every cashback rate shown is a real one
Smaller figures, credible product

15active build days

167commits

8login versions before the final one

97automated test scripts

Where it stands

Not live yet

Perks has not launched. Everything shown here is designed, built and running. The commercial side is still in progress.

Brand partnerships and employer agreements are still being negotiated
Screens, copy and rates continue to change as those conclude
Read this as the current state, not the finished product
Zomato is a demonstration brand here, not a signed partner

Full case study coming soon. angaduiux@gmail.com