How I Built Fernwood Hills with AI

Fernwood Hills started as 40 acres of raw forest near London, Ontario. No brand, no platform, no systems. Today it's a fully operational hospitality business with a 1,000+ page booking platform, AI-powered guest communication, 28 automated workflows, and a multi-machine AI fleet running 24/7.

I built all of it. The brand identity, the photography, the website, the AI systems, the content pipeline — every piece. This post breaks down how AI became the backbone of the operation.

The Problem

Running a hospitality business as a small team means doing everything: guest communication, booking management, content creation, maintenance scheduling, marketing. The volume of repetitive work was unsustainable.

I didn't hire a team. I built one — with AI.

The AI Stack

The system runs across four machines, each with a specific role:

• Mac Mini (Hub) — orchestrator running agent frameworks, monitoring, and routing
• GPU Workstation — runs local LLMs for content generation and guest communication
• Ops Machine — handles streaming infrastructure and the Brain API (RAG system with property knowledge)
• MacBook Air — development machine for the Next.js platform

What the AI Actually Does

This isn't theoretical. These systems handle real operations every day:

• Guest Communication — AI drafts responses to guest inquiries using RAG (retrieval-augmented generation) with 267 chunks of property knowledge. It knows the property, the policies, the local area.
• Content Pipeline — automated image processing, social media content generation, and scheduling
• 28 Cron Jobs — automated workflows handling everything from booking confirmations to maintenance reminders
• Wildlife Detection — YOLO-based computer vision on live camera feeds, identifying animals and announcing sightings to YouTube streams

The Brand Side

AI handles the operations, but the brand was built the old-fashioned way — creative direction, photography, and positioning. Every photo on the site is mine. The visual identity, the tone, the content strategy — that's the creative direction side of what I do.

The combination is what makes it work. Brand vision sets the direction. AI systems execute at scale.

What I Learned

• Start with the problem, not the technology. Every AI system I built solved a specific pain point — not a hypothetical one.
• Local AI is underrated. Running models on your own hardware means zero API costs, full control, and no vendor lock-in.
• Automation compounds. Each workflow I automated freed up time to build the next one. The system gets more powerful over time.
• Brand still matters most.AI can't replace taste, vision, or creative direction. It multiplies whatever you feed it — so feed it something good.