A regional HVAC contractor
We replaced the enterprise platform they were renting with software they own. It was a fixed-fee project with clear math: $200K off the books in year one.
Before
A regional HVAC and mechanical contractor ran several operating regions on ServiceNow. The platform had been configured for them rather than built for them, and they kept paying the per-seat bill because there was no better option on the market. Dispatchers worked around the software. Techs in the field opened it as little as they could get away with. The cost and friction kept climbing, and nobody liked the answer when they asked what else the company could run on.
The first bet
A field-service system built around how they actually work
We replaced ServiceNow with a platform built specifically for this contractor, on top of a well-respected open-source CRM as the foundation so they own the backbone and can keep customizing on top of it. Multi-tenant from day one, so the same platform serves every region they operate in. Fixed-fee project on our side, and around $200K in license fees came off the books in the first twelve months.
What we noticed
The first time the dispatchers used it, they stopped working around the software. The screen showed what they needed to see, in the order they needed to see it, and the techs in the field opened the app on purpose instead of avoiding it. The rollout started landing region by region, on the same platform, with no per-region license renegotiation behind it. For the first time the software felt like theirs instead of the vendor's.
What we built next
A mobile app the techs actually use
We built them a custom mobile app for the field, designed around what a tech is actually doing between calls. Work orders, parts, customer history, and sign-off all in one place, on a screen that was built for a person standing in a mechanical room.
AI workflows pulled into the day-to-day
Custom AI workflows now sit inside the platform: drafting quotes from a tech's notes, pulling the right part numbers off equipment records, and suggesting how to slot a new call into a busy day. The AI isn't a separate product anyone has to remember to open. It runs where the work already lives.
One platform across every region
Multi-tenant from the start, so each region operates on its own data while sharing the same backbone. As they expand the rollout, the next region inherits everything the previous one already shaped, without standing up new software.
What was hard
Own the foundation without rebuilding commodities
We built on a respected open-source CRM instead of recreating authentication, records, and other standard parts from scratch. The contractor owns the backbone while the project budget stays focused on dispatch, field work, and billing.
Roll out one region at a time
The platform was multi-tenant from the start, but adoption happened region by region. That limited cutover risk and let each rollout improve the shared product before the next region moved over.
Design for the mechanical room
The field app was shaped around a technician standing between calls, not a desktop user completing an enterprise form. Work orders, parts, customer history, and sign-off had to fit that moment.
How they use it now
We work with them on a rolling basis: ship the next region, ship the next workflow, then look at what the new data unlocks. The platform belongs to them, and we're the team they call when they want the next thing built on top of it.
What's now possible
Around $200K in license fees came off the books in the first twelve months. The contractor now owns one platform across its operating regions, including the mobile app and AI workflows used in daily field work. Growth no longer adds another per-seat tax or requires a vendor's permission to change the workflow.
What's next
Agent-assisted dispatch and receivables automation are the next candidates. They are not part of the completed proof above; they are opportunities made cheaper by the platform already in place.
“This is the closest I've come to the conversation I was anticipating having… nobody else is leveraging AI like this.”
Client stakeholder
Real quotes from real clients. We anonymize on the site until each one says yes to being named. Most are in the queue. References available now.
What we built