Case Study

FixFlow.

Appliance Repair Dispatch Platform

A SaaS platform that gives property managers a branded intake portal. Tenants submit repair requests, the LLM triages them into a structured work order, and dispatch routes them to completion.

Next.js TypeScript Supabase OpenAI Tailwind CSS Vercel
Alpha Properties
Maintenance Request Portal
Your Name
Jane Doe
Phone Number
(206) 555-1234
Describe the Problem
The dishwasher is leaking water from the bottom...
Submit Work Order
Powered by FixFlow
Multi
Tenant Architecture
GPT-4o
AI Issue Triage
5
Work Order Statuses

The Problem

Repair companies that service property managers have a messy intake problem. Tenants call, text, or email. Details get lost. Dispatch is slow. There's no structured way to capture, triage, and track a request from first contact to completion. FixFlow is the single pipe all of that runs through.

Multi-Tenant

Branded Intake Portals

Every property management company gets their own slug-based URL — e.g., /dispatch/alpha-properties. Their tenants see the company name, not FixFlow. Clean, simple, branded.

OpenAI GPT-4o

AI-Powered Triage

GPT-4o-mini reads the issue description and predicts appliance type and the two most likely failure points. The dispatcher already knows what they're walking into before they call the tenant back.

Status Workflow

End-to-End Tracking

Work orders flow through a structured pipeline: Received → Scheduling → Dispatched → Completed → Invoiced. Every stage is stored in the database, and Twilio SMS keeps tenants updated automatically.

How It Works

Step 1
Property Manager Onboarding
A property management company is added to the database with their name, contact info, and a unique dispatch slug. That slug becomes their permanent intake URL. No setup required on their end.
Step 2
Tenant Submits a Work Order
The tenant hits the portal URL and fills out the intake form: name, phone, unit, and a description of the issue. Client-side validation catches bad data before the request hits the server.
Step 3
AI Triage Runs Automatically
A Next.js Server Action processes the submission. If an OpenAI API key is configured, GPT-4o-mini analyzes the description and writes predicted appliance type + failure points directly to the work order record in Supabase.
Step 4
Notifications Fire
Twilio sends the tenant an SMS confirmation. Resend fires an email to the property manager so they know a new request came in. Both are conditional on their respective API keys being configured.
Step 5
Dispatch & Resolution
The dispatch team moves the work order through the status pipeline until the repair is done and invoiced. The full record (including the AI triage) lives in Supabase for later review.

What It's Built With

Frontend
Next.js 16 (App Router) TypeScript Tailwind CSS 4 React 19 Server Actions
Backend & Database
Supabase (PostgreSQL) Row-Level Security Supabase SSR Service Role Auth
Integrations
OpenAI GPT-4o-mini Twilio SMS Resend Email Stripe Billing Vercel Edge
Architecture
Multi-Tenant SaaS Slug-Based Routing Conditional API Activation Server-Side Validation

See It Live

Try the demo portal or browse the project on GitHub.

Try Demo GitHub Read the build post ← All Projects