Automate Maintenance Dispatch with a Custom AI System

Most property management companies rely on the maintenance module within their Property Management System (PMS) like AppFolio or Buildium. These tools are effective for logging requests but function as simple ticketing systems. They create a ticket but lack the intelligence to triage the issue or dispatch a technician automatically. A property manager must still manually read every request, decide its priority, and then begin the coordination work.

In practice, this means a property manager sees a new ticket for a 'broken AC unit' in Texas in July. They then open a separate spreadsheet or Google Calendar to find an HVAC-certified technician. They have to check who is available and not already on a job 40 miles away. This process involves multiple phone calls, text messages, and cross-referencing schedules, consuming 15-20 minutes for a single urgent request. The PMS records the problem but does not help solve the operational puzzle of getting the right person to the right place quickly.

The structural problem is that PMS platforms are designed as systems of record, not systems of operational intelligence. Their architecture is optimized for storing data about leases and payments, not for making real-time, logic-based decisions. The maintenance 'feature' is just a form connected to a database. These platforms cannot parse the nuance in a tenant's message, calculate travel times, or dynamically match job requirements to a technician's specific skills. This limitation is architectural, not just a missing feature.

Syntora would start with a discovery and data audit of your current maintenance process. We would analyze 12 months of maintenance history from your PMS to understand request patterns, common issues, and resolution times. This audit identifies the key data points needed to train an effective triage model, such as keywords that signal high urgency (e.g., 'water', 'leak', 'no heat') and the skills required for different job types.

The core of the solution would be a Python service running on AWS Lambda, chosen for its efficiency in handling event-driven workloads like new maintenance requests. The service uses the Claude API to perform natural language understanding, accurately extracting the issue, location, and urgency from free-text tenant submissions. A FastAPI endpoint then takes this structured data and queries a Supabase database containing technician skills, certifications, and current locations to find the optimal match in under 500ms.

The delivered system integrates directly with your existing PMS via its API. When a tenant submits a request, the AI processes it and updates the ticket with the assigned technician and their ETA, all within seconds. The property manager shifts from being a dispatcher to a supervisor, managing exceptions rather than every single ticket. You receive the full source code, a runbook, and a system that works inside the software your team already uses.