A gifting store that knows what they'll love
Send a gift to India or Nepal from anywhere in the world, and let AI find one they'll actually love. We designed and built the store, its recommendation engine and its global checkout.
- Client
- Gifting e-commerce brand
- Year
- 2024
- Duration
- Ongoing
- Services
- Product design, AI engineering, E-commerce platform
Overview
An e-commerce gifting website for people who want to send gifts to India and Nepal from anywhere in the world. AI-powered recommendations tailor the store to each visitor, turning a crowded catalog into a short list of gifts that feel personal. Our team handled it end to end: discovery, architecture, design, the AI engine, development, testing, launch and ongoing optimization.
- Platform
- Responsive web · Desktop and mobile
- Our role
- Discovery and product strategy, Solution architecture, UX and UI design, AI and machine learning, Full-stack development, Payments integration, QA and launch
- We delivered
- Gifting website, AI recommendation engine, International checkout
- Headline
- 2 countries delivered to, from anywhere in the world
The challenge
Sending a gift home from abroad is harder than it should be. Most stores don't deliver to India or Nepal, international cards fail at checkout, and endless catalogs leave people guessing. The gift arrives late, or it's the wrong one.
AI integration
The heart of the platform is a TensorFlow engine that learns what each person likes, so finding the perfect gift takes seconds, not an evening of scrolling.
The engine looks at how people browse, what they say they like and what they have bought before, then ranks the catalog for that visitor and surfaces the best matches on the homepage.
It keeps learning. Every click, search and order feeds back into the model, so suggestions improve over time and each return visit feels more personal than the last.
Our approach
Every decision came back to the sender's two worries. Is it the right gift, and will it arrive?
- 01
Start with the sender
We began with people living abroad who want to send something meaningful home, and designed every step around their two worries: choosing the right gift, and trusting it will arrive.
- 02
Let AI do the guessing
We built a recommendation engine that learns from behavior, preferences and past orders, so the homepage shows each visitor the gifts they're most likely to love.
- 03
Make global feel local
We paired international payments with reliable delivery in India and Nepal, so checkout works from any country and the gift lands on time.
- 04
Ship, then keep learning
After launch we kept measuring and tuning, so the recommendations and the catalog improve with every order.
The store
Recommendations sit right on the homepage, and the journey from browsing to checkout takes only a few steps.
Features
Three moments carry a sender from "who is this for?" to a paid order: narrowing the store down, choosing with confidence, and paying the way they already pay.
Smart filtersGender, age and budget turn a big store into a short list. 01 / 04Smart filters
Shop for the person, not the catalog
Senders start with who the gift is for. Filters for gender, age and budget narrow hundreds of products to the handful that fit, and the same signals feed the recommendation engine.
- Filter by recipient gender and age range
- Sort by price, or set a budget band
- Occasion and category chips one tap away
- Every choice sharpens future recommendations
Product pagePrice, size and delivery choices before the sender commits. 02 / 04Product page
Everything needed to decide, on one screen
The product page answers the questions that stop a sender: what's inside, what it costs today and when it will arrive. Delivery is a clear choice, not a surprise at checkout.
- Price, saving and size options up front
- Standard, express or scheduled delivery
- Delivery cities listed before you pay
- Recommended for you, right below the fold
CheckoutA sender's own country code, the recipient's currency, local wallets. 03 / 04Checkout
Pay from anywhere, deliver at home
Most senders live abroad. Checkout starts with a mobile number from their own country, shows the order in the recipient's currency, and offers the payment methods people in Nepal already trust.
- Mobile sign-in with country codes from Australia to Qatar
- Order summary in NPR, with promo codes
- Local wallets: eSewa, Khalti and Fonepay
- One continue button from number to payment
Mobile webBrowse, filter in a sheet, see what fits: the same store on a phone browser. 04 / 04Mobile web
The whole store, one thumb away
Many senders shop from a phone browser. On mobile, filters open in a bottom sheet, applied filters stay visible as chips, and a dot on the filter icon shows the list has been narrowed.
- Filters in a bottom sheet, within thumb reach
- Applied filters shown as removable chips
- A dot on the filter icon when results are narrowed
- Clear all to start again in one tap
Gift packaging
We designed the unboxing as carefully as the store: a gift should feel exciting from the moment it arrives.
Senders can add a moment to the box itself. A gift can play music as it's unwrapped, burst balloons when it's opened or release a shower of confetti.
For anything more personal, a Talk to us button puts senders in touch with the team, so packaging can be designed around one person and one occasion.
How we delivered it
We ran the project from first conversation to post-launch care. Each phase ended with a clear outcome the next one could build on.
- 01
Discovery
We set out to understand the need for a global, personalized gifting service, and defined three requirements: an AI recommendation engine, international payments and reliable delivery to India and Nepal. Outcome: a focused roadmap built on personalization, global access and streamlined delivery.
- 02
Planning
We chose Python and TensorFlow for the recommendation engine, Node.js for the backend, React for the front end and Stripe for secure international payments. Outcome: a defined architecture, stack and development milestones.
- 03
Design
We designed for simplicity and personalization, with recommendations on the homepage and a UI that guides people smoothly from browsing to checkout. Outcome: high-fidelity prototypes and user flows focused on effortless gift selection.
- 04
AI integration
We built the TensorFlow engine that analyses behavior, preferences and past interactions, and keeps learning from every input. Outcome: a recommendation system that helps people find the perfect gift, effortlessly.
- 05
Development
We built Node.js services for fast, reliable order processing and a responsive React front end that works across every device. Outcome: a fully functional platform with AI recommendations, ready for testing.
- 06
Testing
We tested recommendation accuracy, usability and payment security, and ran cross-border tests so delivery and payments worked for senders everywhere. Outcome: a platform proven for global users.
- 07
Launch
We launched the platform, letting people around the world send gifts to India and Nepal with AI suggestions guiding the way. Outcome: a live, user-ready platform with personalized gifting and global reach.
- 08
Post-launch
We monitor performance, gather feedback and tune the recommendations, with more gift categories and a sharper algorithm on the way. Outcome: continuous improvement built into the product.
Capabilities
Everything a sender needs to find the right gift and trust it will arrive.
- 01
Personalized homepage
Every visitor sees gifts ranked for them, based on their behavior, preferences and past orders.
- 02
A model that keeps learning
Clicks, searches and purchases feed back into the engine, so recommendations sharpen over time.
- 03
Pay from anywhere
Stripe handles secure international payments, so senders can check out in any country.
- 04
Delivery to India and Nepal
Gifts ordered anywhere in the world arrive at the recipient's door in India or Nepal.
- 05
Guided checkout
A short, clear path from browsing to paying, designed to remove hesitation at every step.
- 06
Works on every device
A responsive React front end that feels just as fast on a phone as on a laptop.
- 07
Reliable order processing
Node.js services keep orders moving quickly from payment to fulfilment.
- 08
Built to grow
New gift categories and model improvements slot in without rebuilding the store.
- Picked for you
- Pay from anywhere
- Delivered in India
- Delivered in Nepal
- Learns every visit
Architecture
Every visit teaches the engine something. Behavior flows in, the model scores the catalog, and the homepage reorders itself for that visitor.
01 · Signal
- Browsing behavior
- Stated preferences
- Past orders
02 · Model
- TensorFlow
- Scores the catalog
03 · Rank
- Top gifts per visitor
04 · Show
- Personalized homepage
- Picked for you
05 · Learn
- Clicks and purchases
- Feed the next visit
- Front end
- React
- Responsive design
- Cross-device
- Backend
- Node.js
- Order processing
- REST APIs
- AI
- Python
- TensorFlow
- Recommendation engine
- Payments
- Stripe
- International cards
- Fulfilment
- Delivery in India and Nepal
Build decisions
Four choices that kept the store fast, trustworthy and genuinely personal.
- 01
Recommendations on the homepage
We put the AI where people land, not buried in a tab, so the first thing every visitor sees is a gift that fits.
- 02
Python for the model, Node for the store
TensorFlow in Python gave us the best tools for machine learning, while Node.js kept order processing fast and simple to scale.
- 03
Stripe for global checkout
Senders live everywhere. Stripe let us accept international payments securely without building a payments stack from scratch.
- 04
Test across borders, not just browsers
We tested delivery and payments from different countries, because a gift that fails at the border is worse than no gift at all.
The results
0
countries delivered to, India and Nepal, from senders anywhere in the world
- phases, from discovery to post-launch, run by one team
- 0
- recommendation engine that learns from every visit and every order
- 0
Selling online?