TL;DR
- Production-ready AI-first platform for trade and contractor businesses, trunk-based development at v1.8.0
- AI voice receptionist handling inbound calls, smart scheduling, crew dispatch, quoting and AI growth coaching
- Validated that the Solo-Squad Execution Model transfers across verticals with different operational patterns
The Problem
Trade businesses run on phone calls, text messages and spreadsheets. Missed calls mean missed jobs. Quoting is manual. Scheduling is a whiteboard. The existing software options are either too complex (enterprise field service management) or too basic (simple booking tools that ignore the reality of crew-based, multi-site work).
The Approach
After building OpenChair, I wanted to validate whether the same AI-augmented solo operator approach could work for a completely different vertical. The cross-vertical story covers both builds together. Beauty and wellness is appointment-based, individual-client focused. Trades are job-based, crew-dispatched, multi-site. Different operational patterns, different AI requirements.
Key Decisions
-
AI voice receptionist as the entry point. For tradies, the phone is the business. An AI that answers calls, qualifies leads, books estimates and routes urgent requests changes the economics of a one-person or small-crew operation. I cover the technical and design details in the voice agent playbook.
-
Smart scheduling with crew dispatch. Unlike single-practitioner scheduling, trades require multi-person crew assignment, travel time calculation, job duration estimation and equipment tracking. The AI scheduling layer accounts for all of these.
-
AI growth coach. Analyses business data to surface actionable insights: pricing recommendations, seasonal patterns, customer retention opportunities. Not a dashboard, a proactive advisor.
-
Shared architecture with OpenChair where sensible. The tRPC + Drizzle + Supabase stack transferred directly. The AI orchestration layer was reusable. The domain-specific features (quoting, crew dispatch, job costing) were net new.
Results
- 50+ AI features in production
- 26 tRPC routers, 41+ database migrations
- AI voice receptionist, smart scheduling, quoting, crew dispatch
- AI growth coach and website builder
- Production-ready, trunk-based development (v1.8.0)
- Solo-built replacement for 8+ specialised roles
What I'd Do Differently
- Deeper customer discovery with actual trade business operators before building the quoting module
- More aggressive feature flagging to support incremental rollout
Tech Stack
Frontend: Next.js 16, React 19, Tailwind CSS 4
Mobile: React Native (Expo SDK 55)
Backend: tRPC 11, Drizzle ORM, PostgreSQL (Supabase)
AI: OpenRouter, Vercel AI SDK, Retell AI
Infrastructure: Supabase, Stripe (Billing, Connect), Vercel, GitHub Actions