Skip to main content
Home Furnishings Retail HomeHaven UAE & Saudi Arabia 7 months

Enterprise E-Commerce Platform with AI Personalisation

HomeHaven, a premium home furnishings retailer operating across UAE and Saudi Arabia, was running on a heavily customised Magento 1 installation that was end-of-life, performing poorly (average 8-second page load), and costing $180,000/year in maintenance.

0.8s LCP
Page Load Time
+138%
Conversion Rate Improvement
+22%
Average Order Value Increase
−73%
Platform Maintenance Cost
+89%
Organic Search Traffic (6mo)
+AED 4.2M
Revenue Growth (Year 1)

The Challenge

HomeHaven, a premium home furnishings retailer operating across UAE and Saudi Arabia, was running on a heavily customised Magento 1 installation that was end-of-life, performing poorly (average 8-second page load), and costing $180,000/year in maintenance. Their conversion rate was 0.8% against an industry average of 1.9% for their segment. They needed a complete platform rebuild that would handle Arabic/English bilingual content, regional payment methods (Tabby BNPL, local bank cards), and AI-driven personalisation to close the conversion gap.

Our Solution

We rebuilt the entire e-commerce platform as a headless commerce architecture — Shopify Plus as the commerce backend for order management and inventory, combined with a fully custom Next.js storefront optimised for performance and Middle Eastern market requirements. We built an AI personalisation engine that analyses browsing behaviour, purchase history, and home style preferences to serve personalised product recommendations and dynamic homepage content.

Platform Screenshots

cannytechnology.com/app

Screenshot: Homepage with AI Personalisation — product recommendations grid tailored per visitor

UI Mockup Placeholder
cannytechnology.com/app

Screenshot: Product Detail Page (Arabic RTL) — bilingual product page with right-to-left layout

UI Mockup Placeholder
cannytechnology.com/app

Screenshot: Checkout Flow — BNPL options (Tabby, Tamara) alongside standard card payments

UI Mockup Placeholder

Technical Architecture

Headless architecture: Next.js 15 storefront + Shopify Plus commerce backend

Shopify Storefront API (GraphQL) for product, cart, and checkout data

Next.js with server components for sub-1s LCP on both Arabic and English pages

Bidirectional Arabic/English internationalisation with RTL layout support

AI personalisation engine: user event tracking → Pinecone vector store → real-time recommendations

Tabby, Tamara, and Stripe integration for regional payment methods

Algolia search with Arabic-language search optimisation

AWS CloudFront with Middle East edge locations for regional performance

Contentful CMS for marketing team control over landing pages and promotions

Technologies Used

Next.js 15Shopify PlusShopify Storefront API (GraphQL)ContentfulAlgoliaPineconeOpenAI EmbeddingsTabby APITamara APIStripeAWS CloudFrontVercelGoogle Analytics 4Segment

Project Timeline

1

Discovery & Architecture

3 weeks
  • Magento platform audit and data inventory
  • User research with UAE and KSA customers
  • Headless architecture design and vendor selection
  • Arabic i18n requirements mapping
  • SEO migration strategy to preserve organic rankings
2

Shopify Backend Setup

4 weeks
  • Shopify Plus store configuration
  • Product catalogue migration (4,200 SKUs)
  • Pricing rules, discounts, and collection setup
  • Regional tax and shipping configuration
  • Order management workflow setup
3

Custom Storefront Development

8 weeks
  • Next.js storefront with Shopify Storefront API
  • Bidirectional RTL/LTR layout system
  • Arabic and English content architecture
  • Performance optimisation (image pipeline, lazy loading, prefetching)
  • Contentful CMS integration for marketing pages
4

Integrations & Payments

4 weeks
  • Algolia search with Arabic language support
  • Tabby and Tamara BNPL integration
  • Google Analytics 4 and Segment event tracking
  • Wishlist and comparison features
  • Customer review system
5

AI Personalisation Engine

4 weeks
  • Behavioural event pipeline (browse, wishlist, purchase)
  • Product embedding generation with OpenAI
  • Pinecone vector store setup and indexing
  • Real-time recommendation API
  • A/B testing framework for recommendation variants
6

Testing & Launch

5 weeks
  • Performance testing across MENA network conditions
  • Cross-browser and cross-device QA
  • SEO redirect mapping and validation
  • Soft launch with 10% traffic split
  • Full cutover and Magento decommission

We went from one of the slowest sites in our category to one of the fastest. The personalisation engine genuinely surprised us — we were sceptical AI recommendations would work for furniture (such a personal choice), but the data is clear. Customers who see personalised recommendations convert at 2.7x the rate of those who don't.

Layla Al-Mansouri, Head of Digital, HomeHaven

Long-Term Impact & ROI

The new platform launched 10 weeks ahead of the peak Ramadan shopping season in 2025. Page load times dropped from 8 seconds to 0.8 seconds (10x improvement). Conversion rate rose from 0.8% to 1.9% (matching industry average) within 3 months, and reached 2.3% by month 6 as the personalisation engine accumulated more data. Revenue grew by AED 4.2 million in the first year against a platform investment of AED 950,000.

Want Results Like This?

Every case study started with a free discovery call. Let's explore what's possible for your business.

Book Free Consultation