We replaced a manual dispatch process with an AI that books itself

A field services company was running their repair and maintenance ops through HouseCall Pro. Every time a customer called in, a dispatcher would manually check which techs were available, eyeball the map for nearby jobs, and pick a slot and assign someone.

It worked. But it didn't scale.

We built an autonomous scheduling agent that handles all of this in seconds — intake, job creation, smart slot booking, and tech assignment, without a human in the loop. The customer picks their tech. The system handles everything else.

Instead of random assignment, the AI clusters available slots by geography. It finds technicians who already have nearby jobs and suggests slots that minimize drive time. Less travel means more jobs per day.

We embedded a retrieval-augmented knowledge base covering common HVAC issues — furnace noise, AC failures, refrigerant leaks, no-heat situations. When a customer describes their problem, the system instantly retrieves relevant context to help the tech arrive prepared.

Not all problems are equal. “Furnace making noise” is a P3. “No heat in January” or “suspected gas leak”? That is auto-escalated to P1 — with slot urgency adjusted accordingly.

Still early days, but this is what AI-native operations looks like in practice — not a chatbot bolted on top, but intelligence woven into the workflow.