Trades and home services

AI agents for trades: turn incoming requests into work orders.

The dispatch spreadsheet that worked at 5 trucks starts failing at 15, and the work orders, calls and texts keep growing while the office stays at two or three people. The hours go into the office, where the dispatcher retypes each work order email into the job system and the office manager follows up on missing details before anyone can schedule the job.

ServiceTitanJobberHousecall ProWorkizFACIL-ITQuickBooks

These systems hold the jobs, the price book and the invoices, and if one already takes requests in on its own it is enough, but in most offices people still carry each work order from the shared inbox into the job system by hand.

Signs you have it

Where the office hours go.

  • The dispatcher copies the client, address, work order number and scope from each email into the job system, field by field.
  • A job sits unscheduled because nobody has keyed in the service deadline or the not-to-exceed amount yet.
  • The techs finish the work in the field, and the office spends the evening retyping it.

How it works with your rules

Your rules from the first day, and your team decides what it learns.

We build an agent that follows the company's own rules from day one, such as each client's not-to-exceed limit, the fields a work order needs before it is scheduled and which tech takes which trade, and when the dispatcher changes a trade, moves a job or fills in a field, the team decides whether that change is saved as a rule, an example or an exception that the agent follows next time. The dispatcher reviews its work at first and decides anything that needs a decision, and the agent works on its own only within limits the owner sets.

At first, your team reviews its work. Review stops one type of task at a time, once its record on them is good and you agree. It learns how your company does this job. It does not watch, rate or profile your employees. You can read everything it has learned and remove anything.

How it runs: a fit test first, and if a ready-made tool is enough, we say so. Then a build in ten business days and ninety days of support, included. You own the agent, its tests and guide, and what it has learned. Your team runs it, or our engineers do, on your own servers, in your own cloud account or in our data center.

The use cases

Explore agent examples for trades and home services.

Every job is written for the work itself, so the four kinds of work on the use-cases page are the other way in.

Facilities maintenance contractors

From a request a business posted

Work orders from email into the job system

A dispatcher retypes every work order email into the job system, field by field, before anyone can schedule it.

Most work orders reach a trades office by email, and portal requests usually send a copy there too, so this re-entering is where the dispatcher's day goes before any job can be scheduled.

Read the full write-up