Patofy
Global AI Growth Platform for WhatsApp-First Service Businesses
Overview
Patofy is an autonomous AI-driven growth engine purpose-built for appointment-based service businesses—dental clinics, aesthetic centers, wellness spas, and consultancies—across WhatsApp-dominant markets in Europe, Latin America, the Middle East, and East Africa. Designed around the 'Four R's Framework' (Reception, Reputation, Retention, and Referrals), Patofy unifies front-desk scheduling, Google review collection, loyalty ledgers, and lapsed-customer re-engagement into a single, closed-loop automation platform.
The Challenge
Service businesses lose upwards of 30% of high-intent bookings because over 60% of customer inquiries arrive after hours (between 8 PM and 7 AM) when staff are unavailable. When replies are delayed until morning, prospective clients have already booked with competitors. Furthermore, existing CRM solutions are fragmented and require manual data entry. The core engineering hurdle was architecting a multi-tenant system that connects Meta's WhatsApp Cloud API with stateful AI runtimes, handles burst-message debouncing, guarantees sub-60-second human handoff recovery, and enforces strict tenant data isolation at scale.
The Solution
I architected and engineered the end-to-end platform from data model to production deployment. The backend utilizes Payload CMS with MongoDB, orchestrating 35+ collections spanning appointments, practitioner rosters, automated marketing broadcasts, and audit logs with strict row-level security. The conversational engine pairs WhatsApp Cloud API webhooks with Voiceflow runtimes, leveraging a distributed Redis debounce buffer to assemble burst messages into coherent turns, managing multi-turn memory within Meta's 24-hour service window, and activating an automated 60-second fail-safe handoff when clients request human staff. The entire production stack is self-hosted on a Hetzner Cloud VPS orchestrated via Dokploy behind Cloudflare edge security, eliminating serverless cold starts and slashing infrastructure costs by over 90%.
Key Results
Transformed WhatsApp from a chaotic chat inbox into a predictable 24/7 booking engine and retention machine. The platform autonomous front desk resolves inquiries instantly, automates appointment bookings, triggers post-service review workflows, and re-engages dormant clients with personalized offers. Currently in pilot deployment across selected service businesses, with a validated commercial roadmap targeting 20+ global accounts and $1,900+ MRR.
- Multi-Tenant SaaS Architecture with Strict Row-Level Security (Payload CMS + MongoDB)
- Conversational AI Receptionist with Distributed Debouncing (WhatsApp Cloud API + Voiceflow)
- Fail-Safe 60-Second Human Handoff with Automatic Timeout Recovery
- Automated Growth Engine (Post-Care Google Reviews, Loyalty Points, Reactivation Broadcasts)
- High-Performance Self-Hosted Cloud Infrastructure (Hetzner VPS + Dokploy + Cloudflare Edge)
Engineering Deep Dives
In-depth architectural postmortems, system design trade-offs, and production blueprints written for this project.
Architecting a WhatsApp AI Receptionist
A deep dive into bridging WhatsApp Cloud API with conversational AI runtimes and 60-second handoff recovery.
Ditching Vercel for a $15 Hetzner Server with Dokploy
A practical breakdown of multi-tenant data isolation, row-level security, and self-hosting production infrastructure behind Cloudflare.
How to Revive Dormant Clients on WhatsApp
A practical win-back playbook for service businesses: when to message, what to say, and how to track recovered bookings.
How We Built an Automated WhatsApp Reputation Engine
How we collect reviews, catch negative feedback privately, and turn chat compliments into social proof on WhatsApp.
Tech Stack
Deep Dives
Architectural essays exploring the code and infrastructure: