Agents in the Trades
Hiring

Green techs routed toward the hours their next license actually needs

It is 7:05 a.m. at the dispatch board and the service manager is staring at two screens that do not talk to each other: today’s board, already filling, and a spreadsheet showing a second-year apprentice still short on supervised service hours for his next license — after three weeks on installs.

Source: ServiceTitan, "Pantheon 2026: Live coverage from ServiceTitan" — https://www.servicetitan.com/blog/pantheon-2026-live-coverage — the coverage describes a multi-brand operator standardizing skills, job types and business units before its dispatch optimization improved, and dispatch tooling that works from those skills and job types. This post turns that discipline into a buildable workflow for apprentice development; it is not a shipped product claim, and no vendor figures are used.

The problem

A license is not one pile of hours. The next level asks for hours in categories — service, install, work under a licensed supervisor — and an apprentice collects whatever the board hands him. Nobody notices the mix is wrong until he is near eligibility and one category is still nearly empty.

Dispatch cannot fix that in the moment, because dispatch is solving today: who is closest, who is free, who can bill the call. Development is a quarter-long problem decided by a tool built for a single morning.

Get one workflow like this, every day.

The free Agents in the Trades newsletter — each workflow broken down: what it does, what it costs, and where it breaks.

How the agent workflow works

  1. The license map gets built once, per tech. The manager enters each apprentice’s next level and the hours it requires, by category — and verifies the map before anything runs on it.
  2. Completed work tallies itself. As jobs close, the agent sorts each apprentice’s hours by category from the job record — job type and supervising tech — and keeps a running gap: hours banked, hours still owed.
  3. The board gets a nudge, not an order. When a suitable job lands — right category, a licensed tech assigned who can supervise — the agent flags it with its reason: this call counts toward the category this apprentice is shortest on.
  4. The dispatcher still dispatches. Suggestions can be accepted or overridden in one tap; overrides are logged, not argued with. The agent steers the quarter, never the morning.
  5. The quarter gets reviewed, not the scramble. Monthly, the agent shows each trajectory — on pace, drifting, or stalled in one category — while work still exists to fix it, so a schedule change, not a delay, is the answer.

What you need to set it up

Where it breaks

Hours tallied are not automatically hours that count. Licensing bodies accept documented, supervised hours, filed their way. The tally is a planning aid; the official record, signatures and filing still belong to a human who checks them against the rules.

Steering has a price today. Sending the green tech to the job he needs can mean a slower call, a senior tech pulled into supervising, or a longer drive past a closer tech. On a slammed board the dispatcher will override — rightly. If that happens daily, the shop has no slack to develop anyone, and no agent creates slack.

The good jobs are not neutral. The assignments that build hours can also carry overtime, spiffs or the customers everyone likes. Route them by license gap alone and you have quietly set people’s pay. The manager still has to watch the distribution and explain it.

The bottom line

Run this if your apprentices discover their hour gaps at the finish line. Keep the map honest, tally work as it closes, and let the agent whisper to dispatch all quarter — so the next license is a date on the calendar, not a surprise delay.

Related use cases

← All use cases

One working agent workflow, every day.

Real AI agent use cases from solar, construction, and field operations — what they do, what they cost, and where they break. No hype, no theory.