A contractor in Texas has been on ServiceTitan for four years. $42,000 a year in software spend. A "ServiceTitan coordinator" on the org chart - a dedicated person whose job is to maintain the platform, build reports, configure automations, and train new staff on the system. He's also still dispatching manually because the routing logic in ServiceTitan doesn't account for certification requirements - his HVAC techs have different EPA classes and he can't build that constraint into the dispatch board without custom development work he was quoted $15,000 to do.

He's not angry at ServiceTitan. It does what it does. He's frustrated that he's paying $42,000 a year plus a coordinator salary for software that still requires him to manage the most important parts of his operation manually.

What ServiceTitan Actually Is

ServiceTitan is an FSM platform - field service management software. Good job management, solid CRM, decent reporting. Built for the middle of the field service market: companies large enough to need structure, small enough to manage from one platform. For a 5-20 tech operation that does standard residential service work, it provides real value.

The structural limitation is that it's software a business configures and manages. Every automated workflow was built by someone inside the company using ServiceTitan's tools. The dispatch board is a visual tool that a dispatcher uses to manually assign jobs - it doesn't make decisions, it helps a human make them faster. The reporting is powerful but requires someone who knows how to build and maintain the reports. When the ServiceTitan coordinator leaves, a significant portion of institutional knowledge about how the platform is configured walks out with them.

That's not a criticism. That's what a configurable platform does. The question is whether that model fits where the company is trying to go.

What Custom AI Actually Is

Custom AI is not software the company manages. It's a system built around how the company operates, installed running, and operated alongside them. The distinction is foundational. ServiceTitan makes existing workflows faster. Custom AI replaces the workflows.

A company running custom dispatch doesn't have a dispatcher scheduling the same way but with better software. It doesn't have a dispatcher scheduling standard calls at all - the system handles those, and the dispatcher handles exceptions. A company running billing automation doesn't have an admin building invoices faster from better tools. The invoice builds itself from the work order at job close. A company running revenue leakage detection doesn't have a billing manager who's better at catching unbilled work. The system flags every gap automatically before the billing cycle closes.

The result is an operation that runs more of itself. Not because software is doing tasks faster, but because the infrastructure is built to make the decisions that previously required a person in the loop.

Who Should Use ServiceTitan vs Custom AI

ServiceTitan is the right call for companies that have the internal resources to configure and maintain a platform, whose workflows are standard enough that the platform fits without heavy customization, and who want a structured environment to manage job history and customer records. It serves a real market well.

Custom AI is the right call for companies that have outgrown manual operations and don't want to hire more admin staff to scale, whose workflows have specific requirements - certification matching, custom routing logic, multi-trade complexity - that platforms can't accommodate without expensive custom development, and who want a system that expands with them rather than requiring more management as volume grows.

The companies that move from ServiceTitan to custom AI are typically running 20-100 techs, have already spent significantly on platform implementation and coordination staff, and have hit a ceiling where the next increment of growth requires either more admin overhead or a different approach to how the operation runs.

Most companies start by fixing one workflow - dispatch or billing - and the ROI funds the next build. It's not a platform decision. It's a systems decision.
Three questions that tell you which camp you're in: First - when you dispatch, is a human assigning each job or is a system routing them? If a person is making every assignment, you're in manual operations regardless of what software you're using. Second - when a job closes, does the invoice go out automatically or does someone have to generate it? If an admin is still building invoices, that's a workflow, not an automation. Third - do you know today which completed jobs haven't been billed? If the answer is no, you have a revenue leakage problem that software configuration can't solve - you need a detection system.

TMI doesn't compete with ServiceTitan at the platform level. TMI builds the AI layer that sits on top of or alongside whatever job management system a company already uses, and then extends into the autonomous systems that platforms can't provide. The starting point is The Audit - 30 minutes identifying which workflows are costing the most and what the system would recover. That's where the decision between more platform and more infrastructure gets made with actual numbers attached.

TMI Field Notes · Operations
← All stories