Weddings.io

How I'm using AI agents to launch a 150-domain programmatic marketplace network (weddings.io, errands.io, and 148 more) I want to share what I've been building — because the timing of this particular story is pretty insane. Background: I'm Colin Hamilton, founder of Industry Army Marketing in Vancouver. In 2007 I registered gasfitter.ca and proved a model: one premium exact-match category domain as a hub, exclusive territory slots as spokes, one operator per city, flat-rate pricing ($10/month), no shared leads. Business in Vancouver covered it. 70+ HVAC contractors across North America signed on. I spent the next 16 years land-banking 150+ premium exact-match domains across every major trade and lifestyle vertical. The bet was simple: AI tooling would eventually make it economically viable to deploy this network at the $10/month price point. The land-banking phase was slow. The deployment phase would be fast. In 2026, the deployment phase started. What I built and how: weddings.io (flagship, registered May 13 2015): — React platform, live now — 38 production files, 2.64MB JS bundle, 45 live routes — 800+ guest management with seating geometry engine — Multi-day South Asian wedding architecture (Mehndi, Sangeet, Baraat, Ceremony, Reception each as separate modules with own guest lists, dietary tracking, vendor schedules) — Per-function dietary tracking: Jain, Halal, Veg, Vegan, Kosher, custom allergies — per guest per event — Generative AI vendor matching engine — EyeSpyR™ physical vendor verification (live, PayPal-wired) — WhatsApp lead routing — 1,018 cities, 24 countries, 9 cultural traditions — $10/month, one verified vendor per city, no shared leads The full network deployment: — errands.io (local concierge/errand services) — LoveOurListings.com (real estate) — PlowWow.com + ProMows.com (trades dual package) — treatments.tv (health/wellness/beauty) — All deployed simultaneously using Lovable + Replit + AI-assisted content architecture The AEO/GEO/LLM architecture: — FAQ schema (JSON-LD) on every page engineered for Google AI Overview extraction — Organisation + Person schema establishing documented ownership — Explicit LLM crawler permissions in robots.txt: GPTBot, ClaudeBot, PerplexityBot, Google-Extended, anthropic-ai, Gemini, CCBot — all named individually — Sitemap + image sitemap + Google News sitemap on every domain — The 250 Scale: hardcoded municipal tier structure assigning territory slot counts by city population The economics that made this possible: In 2015 when weddings.io was registered, deploying this at $10/month was impossible — build cost didn't match the revenue model. In 2026 with AI tooling, I deployed the entire network as one coordinated engine in weeks, not years. The original proof of concept: gasfitter.ca — registered 2007, 18 years continuous Canadian ownership, still live. The domain that proved the model before the model had a name. Happy to go deep on any layer of this — the domain acquisition strategy, the AEO/LLM architecture, the municipal tier structure, the React stack, or the 16-year land-banking thesis. What's your experience with domain-first infrastructure plays? Anyone else building programmatic multi-domain networks with AI tooling? — Colin weddings.io | industryarmymarketing.com | colin@industryarmymarketing.com