AI & AUTOMATION · FEATURED PROJECT

AI-Powered Service Booking SaaS Platform

Designed, built, and deployed a multi-tenant SaaS platform from zero, enabling service businesses to automate bookings via AI SMS conversations, manage clients, and collect payments — all without a single manual intervention.

RoleFounder / Full-Stack Builder

StatusLive in Production

StackPython · FastAPI · PostgreSQL · Twilio

TypeSaaS · AI Automation

📌 The Problem

Small service businesses — HVAC, plumbing, landscaping — lose revenue every week because after-hours booking calls go unanswered. Hiring a receptionist is expensive. Existing software is bloated, complex, and built for enterprise.

I saw an opportunity: deploy an AI agent that handles the entire booking conversation via SMS, so business owners never miss a lead, even at 2am.

🏗️ What I Built

An AI SMS agent that conducts natural booking conversations, collects job details, and confirms appointments without human involvement
Multi-tenant architecture supporting multiple clients with isolated data, per-client branding (colors, logo, service menus), and custom domains
Stripe integration for subscription billing and in-booking payment collection
FastAPI backend deployed to Railway with a PostgreSQL database for persistent client, booking, and conversation data
A client-facing marketing site with a tiered pricing page and live booking widget

📊 Outcomes

100%

Automated booking flow — zero human touchpoints

Live

Deployed to production with paying clients

Multi

Tenant architecture supporting N clients from day 1

🔧 Sample Architecture

# Simplified booking agent flow
async def handle_sms(incoming: SMSPayload):
    session = get_or_create_session(incoming.from_number)
    
    if session.state == "awaiting_service":
        service = extract_service_intent(incoming.body)  # AI NLP
        session.service = service
        return ask_for_datetime(session.client_config)
    
    elif session.state == "awaiting_datetime":
        dt = parse_natural_datetime(incoming.body)
        booking = create_booking(session, dt)
        notify_business_owner(booking)
        return confirm_booking_sms(booking)
        

🛠️ Technologies Used

PythonFastAPIPostgreSQL Twilio SMSStripe APIRailway AI Prompt EngineeringMulti-tenant SaaS RESTful API DesignHTML/CSS/JS

💡 Key Learnings

Building for non-technical clients means obsessing over the onboarding experience — the AI only works if the business data is set up correctly
SMS compliance (A2P 10DLC / TCPA) is a critical business constraint that requires deep understanding of carrier regulations
Multi-tenancy requires architectural decisions made at day 1 — retrofitting it later is painful
The best product insight came from talking directly with service business owners, not building in isolation