See which machines pay their way
Utilization, cost per hour and AI drafts that people approve.
The fleet already throws off the numbers: engine hours, idle time, fuel, work orders, parts. Fleet intelligence sorts them into which machines pay their way, what each hour costs and what the shop should do next.
Hours here, costs over there
Hours live in a tracking portal, repair costs in accounting and fuel in a stack of tickets. Nobody can say what the oldest wheel loader costs per hour, or whether all six excavators on the books need to be there. So the next purchase gets decided on a hunch, and the next rental on a phone call.
Answers the owner asks for
Utilization by machine
Engine hours against available hours per machine, job and week, with idle time split out.
Cost per hour
Owning and operating cost per hour from purchase, repairs, parts and fuel, per machine.
Drafts, not decisions
The AI agent drafts work orders and parts lists. A person approves every one.
Your data, any system
REST API, webhooks, an MCP server and full CSV export, all under your permissions.
Which plan answers which question
Field plan
Utilization and idle reports per machine and job, plus full data export, on every connected machine.
Shop plan
Adds owning and operating cost per hour, built from work orders, parts and the hours each machine ran.
Command plan
Adds the AI maintenance agent, the full API, webhooks and MCP server, and single sign-on.
Data export
Download every hour, position and work order as CSV whenever you like, including after you cancel.
What the numbers cover
Utilization by job
Hours worked, idling and parked per machine, rolled up by job, machine class and week.
Owning and operating cost
Purchase or lease, repairs, parts, fuel and tires per machine, divided by the hours it ran.
Underused machine list
Machines under your utilization benchmark for 30 days, ready for a keep, move or sell decision.
AI maintenance agent
Reads fault codes, hours, inspection notes and your manuals, then drafts work orders for approval.
Questions in plain language
Ask “which loaders idled most last week?” in English or Spanish and get the table back.
API, webhooks and MCP
REST endpoints, event webhooks and an MCP server that respects each user’s permissions.
The parts of fleet intelligence
Utilization and cost analytics
Utilization, idle ratio and cost per hour for every machine, from its own engine.
AI maintenance agent
Drafts work orders and parts lists from your manuals; a person approves each one.
API and integrations
Fleet data flows into the systems you already run, through gates you control.
Example: a 200‑machine road contractor in El Paso
A road and bridge contractor runs 200 machines from yards in El Paso and Las Cruces, and its owner is weighing three more excavators. The utilization report shows nine of the current 30 averaged under ten hours a week last quarter, most of them parked at a finished job near Anthony. Cost per hour puts two older units far above the rest. The owner moves five to active jobs, sells the two and buys one. The shop lead now approves the AI agent’s draft work orders each morning, and the controller pulls hours into job cost nightly through the API.
Numbers, drafts and open data
How utilization is counted. Utilization is engine hours divided by available hours. Gytix takes hours from each machine’s engine controller and available hours from your shift calendar, so a machine that runs 40 of a possible 50 hours scores 80%. Idle time is reported separately, because an idling engine burns fuel and hours without moving dirt. Set a benchmark per machine class, and anything under it for 30 days lands on the underused list.
Cost per hour, from records. Owning cost covers purchase or lease, interest, insurance and depreciation. Operating cost covers fuel, repairs, parts, tires and wear items. Gytix pulls repairs and parts from work orders and fuel from fuel logs; you enter the owning figures once. The total is divided by the hours each machine ran, giving a cost per hour you can set beside your rental rate or your bid rate.
What the AI agent does, and does not. The AI maintenance agent reads fault codes, hours, inspection notes and the service manuals you upload. It drafts work orders and parts lists and answers questions about the fleet in plain English or Spanish. It never closes a work order or orders parts without a person approving it. A draft can be wrong, which is why a mechanic reviews each one, and every draft lists the codes, hours and notes it was built from.
Open by default. The full API, webhooks and MCP server come with Command. The API serves hours, positions, work orders and costs to accounting or job cost systems, and webhooks push events such as a new fault code as they happen. The MCP server lets AI assistants your team already uses query fleet data and create draft work orders, with the same permissions as the user who connects it.
Analytics and AI questions, answered
How do you calculate equipment utilization rate?
Divide the hours a machine actually ran by the hours it was available to run. Gytix takes engine hours from the machine and available hours from your shift calendar, splits out idle time, and reports utilization by machine, job and class. Our guide to equipment utilization rate works the formula through step by step.
Does the AI maintenance agent make repair decisions on its own?
No. The agent reads fault codes, hours, inspection notes and your service manuals, then drafts work orders and parts lists for a person to approve. It never closes a work order or orders parts without human approval, and your shop can edit or reject any draft before it becomes work.
How accurate is the AI maintenance agent?
We do not publish an accuracy figure, because each draft depends on your own records and manuals. Treat a draft as a first pass: it lists the codes, hours and notes it used, and a mechanic approves, edits or rejects it before anything happens on the machine or the parts shelf.
What is an MCP server, and why would we want one?
MCP is an open standard that lets AI assistants connect to outside data. Gytix’s MCP server lets the assistants your team already uses query fleet data and create draft work orders, with the same permissions as the user who connects it. It is included in the Command plan, priced per connected machine per month.
Can Gytix send engine hours to our accounting system?
Yes. On the Command plan, the API and webhooks send engine hours, costs and work orders to your accounting or job cost system, and we build the connector during onboarding. On every plan, you can download the same data as CSV at any time, including after you cancel.
How is fleet intelligence priced?
Per connected machine, per month. Utilization reports come with the Field plan, owning and operating cost per hour with the Shop plan, and the AI maintenance agent, full API, webhooks and MCP server with Command. Every plan is month to month with unlimited users. Your written quote, sent within two business days of a walkthrough on your own fleet list, gives the exact figure.
Ask your own fleet a question
Send the fleet list you already keep. We will say plainly whether fleet intelligence is the right place to start.