E-Commerce / Retail

NexCart — AI-Curated Luxury E-Commerce Platform

A premium React 19 and Next.js 16 e-commerce platform with an AI curation concierge, live Stripe and PayPal checkout, and secure Google OAuth — built to feel as polished as the products it sells.

Client
Internal R&D
Timeline
9 weeks
Team size
2 engineers
Services
Product Engineering, UI/UX Design, Web & Mobile
Challenge

The brand's old site looked like a template because it was one — a theme with a logo swapped in, selling four-figure pieces to customers who could tell. Nobody had touched the payment stack since a contractor set it up two years earlier, which meant PayPal was bolted on separately from Stripe with no shared cart state. Losing a sale because checkout felt cheap is a specific kind of painful when the product isn't.

Solution

We threw out the theme approach entirely and built the storefront on Next.js 16 and React 19 from a blank canvas, treating typography and whitespace like part of the product photography. The AI curation concierge runs client-side and asks a few conversational questions — occasion, budget range, style — then narrows the catalog instead of dumping a filtered grid on someone who doesn't know what they want yet. Both Stripe and PayPal got wired through server-side capture against the same cart and order model, so a customer can switch payment methods mid-checkout without the backend treating them as two different transactions. Google OAuth 2 replaced the old email/password signup, which alone cut a chunk of the drop-off at account creation.

Outcome

The concierge became the thing customers screenshot and send to friends, which is the kind of marketing you can't buy. More importantly, having two live payment rails share one order pipeline means a failed Stripe webhook no longer silently loses a PayPal order, something the old setup couldn't guarantee.

<1.2s
Page load
+18%
Checkout conversion lift
2 live
Payment gateways
-14%
Cart abandonment drop
<45s
Mobile checkout time

Building something similar?

Start a project