Built with AI: How Enso Labs Shipped a Production AI Site in 24 Hours

The Stack

Built with AI is a case study documenting how Enso Labs shipped ensolabs.ai — a production Next.js 14 website with 71 JSON-LD schemas, MCP endpoint, and GA4 event tracking — in 24 hours using Claude Chat, Claude Design, Claude Code, and Vercel. The same AI-native delivery process we use for client engagements built our own studio presence.

• Claude Chat (Opus 4): Strategy and research. Competitive analysis, positioning, content architecture, JSON-LD schema planning, and SEO strategy — all reasoned through in conversation before a single line of code was written.
• Claude Design: Visual prototyping. Layout exploration, component design, color system definition (OKLCH), typography pairing (Inter Tight + JetBrains Mono), and responsive breakpoint planning.
• Claude Code: Production engineering. Every line of Next.js 14, TypeScript, and CSS was pair-programmed with Claude Code — from the App Router page structure to the dynamic OG image generation endpoints.
• Vercel: Auto-deployment from GitHub. Push to master, site is live in under 60 seconds.

What We Shipped

The site isn’t a landing page. It’s a complete studio presence with production infrastructure:

• 71 JSON-LD schema blocks across 13 pages — Organization, Person, ProfessionalService, Product, FAQ, LocalBusiness, Article, Breadcrumb, WebSite, ContactPoint. Every schema validated with zero errors.
• MCP endpoint at /.well-known/mcp.json — making the site discoverable by AI agents and Claude integrations.
• signal2noise intelligence: The studio’s news-intelligence engine, published natively on the Insights page (ensolabs.ai/insights) — daily AI market signals, scored and curated.
• Dynamic OG images: Edge-runtime generated Open Graph images for every case study and insight article — unique social previews without a static image pipeline.
• GA4 event tracking: Form submissions, share button clicks, signal2noise feed interactions, and page engagement — all tracked with custom events, not just pageviews.
• Mobile responsive: Every page tested across breakpoints. The navigation transforms to a full-screen overlay on mobile. Typography scales. Cards reflow. Nothing breaks.
• AEO-optimized: Every page leads with a definition-style sentence designed for AI answer engine extraction. robots.ts explicitly allows GPTBot, ClaudeBot, PerplexityBot, Applebot, and Google-Extended.
• RSS feed + sitemap: Auto-generated from the content data layer. Search engines and feed readers pick up new content automatically.

The Process

The 24-hour timeline wasn’t a sprint — it was a demonstration of what AI-native delivery looks like when you remove handoff friction:

• Hours 1–4: Strategy. Claude Chat mapped the competitive landscape, defined the three-pillar positioning (AI Transformation, Agentic Systems, Financial AI), planned the page structure, and drafted all JSON-LD schemas.
• Hours 5–8: Design. Claude Design iterated on layout concepts, established the dark navy + teal accent color system, and produced component-level designs for hero sections, cards, timelines, and the FAQ accordion.
• Hours 9–20: Build. Claude Code generated the entire Next.js 14 codebase — App Router pages, client components, server components, API routes, metadata configs, and the complete CSS file. Every component was reviewed and refined in real-time.
• Hours 21–24: Polish and deploy. Mobile fixes, performance optimization, OG image validation, schema testing, and the final push to Vercel.

Why This Matters

This site is the proof that AI-native delivery is not theoretical. A single senior practitioner with the right AI toolchain can ship production infrastructure that would take a traditional agency team weeks — with better SEO, more structured data, and tighter code.

That’s the thesis behind Enso Labs: AI doesn’t replace senior judgment. It amplifies it. The human decides what to build and why. The AI handles the volume, velocity, and consistency.