OK Car Hire — Technical & Commercial Case Study

Summary: In under two months we delivered an enterprise car-rental platform that combines clean architecture, engineering excellence, high performance and AI acceleration, designed for large development houses and consultancies that demand quality, speed and scalability.

🏗️ Stack & key components 🔧

• Frontend: React 18 + TypeScript (strict)
• Build / Dev: Vite (HMR, tree-shaking, minification)
• Styling: Tailwind CSS + PostCSS
• Backend ERP: Odoo 18 via JSON-RPC 2.0
• Design: Figma with hand-off and plugins (VSCode ↔ Figma sync)
• Typed services & contracts: vehiclesService, categoriesService, helpdeskService and DTOs in odoo-contracts.ts
• Image optimization: WebP / OptimizedImage (lazy, srcset, decoding="async")
• E2E tests: Playwright with Page Object Model

🔌 Enterprise integration & architecture

Resilient JSON-RPC client handling authentication, errors and CRUD operations over Odoo models (fleet.vehicle, vehicle.rental.agreement, vehicle.rental.category, etc.). For a deeper technical write-up of the backend and delivery process, see: From zero to production in 8 weeks — Building a vehicle rental system with AI co-pilots.

• Togglable mocks with VITE_USE_MOCK for offline development and fast iteration.
• Decoupled services and strict contracts for maintainability and scalability.

We also automated PR generation and documentation to keep the workflow reproducible.

🤖 AI as a strategic lever (new & highlighted)

AI agents and automations were used to:

• Generate base components and E2E test templates.
• Create realistic test data and mocks.
• Suggest refactors and code optimizations, and generate PRs/documentation automatically.

Result: significant reduction in implementation and review time without compromising technical quality.

⚡ Performance & code optimization (new & highlighted)

• Lazy loading of images and components to reduce initial bundle and improve LCP.
• Code-splitting & dynamic imports to load only what’s necessary per route.
• Tree-shaking and minification with Vite to remove dead code.
• Image optimization (WebP, srcset, conversions) and attributes loading="lazy" + decoding="async".
• Caching & compression: Nginx configured with Gzip, Cache-Control and expirations for static assets.
• Additional practices: selective prefetching, dependency analysis and TTFB/TTI optimizations.

Impact: measurable improvements in load times, interactive experience and reduced hosting/bandwidth costs.

📱 Fully responsive and SEO-friendly (new & highlighted)

• Mobile-first: design and testing across breakpoints for consistent mobile, tablet and desktop experiences.
• SEO-friendly: semantic HTML, optimized meta tags (title/description), Open Graph and clean URLs; use of preconnect/prefetch to improve crawlability and perceived speed.

Combined with performance and accessibility, this improves indexing, CTR and conversion rates.

🧪 Quality & deployment

• Playwright E2E automates critical flows (forms, bookings, multi-step navigation).
• CI/CD friendly: Docker multi-stage builds (Node → Nginx) with healthchecks and security.
• Best practices: code reviews, integrated documentation and linters/TS to reduce technical debt.

💼 Business value for large dev shops and consultancies

• Rapid, repeatable delivery: complex MVPs on aggressive timelines thanks to AI + automation.
• Reduced risk: strict TypeScript and testing reduce regressions and operational costs.
• Enterprise scalability: ERP integrations and modular architecture ready for growth.
• Differentiation in proposals: pixel-perfect design with Figma + team sync, combined with high-performance engineering.

✅ Conclusion

OK Car Hire is an operational blueprint: robust architecture, high-quality standards, optimized performance, SEO and strategic use of AI agents that speed delivery and improve quality. It’s a replicable, scalable solution for enterprise projects seeking speed, maintainability and high business impact.

Interested in applying this approach? See /en/servicios/ai-driven-development or contact us.

Acknowledgements

Special thanks to Omar Díaz and Jancel Moreno. This project would not have been possible without them.

---