If you have employees and a fire extinguisher on the wall, you are already on a maintenance schedule. OSHA requires that portable extinguishers provided for employee use be visually inspected monthly, that they get an annual maintenance check, and that the date of that check be recorded and kept on file (29 CFR 1910.157(e)).
If you run vehicles that meet the federal definition of a commercial motor vehicle, you are on a second one. Each vehicle has to have passed an inspection of the components listed in Appendix A at least once during the preceding 12 months, with documentation of that inspection carried on the vehicle (49 CFR 396.17). Separately, your drivers have to prepare a written report at the completion of each day’s work on each vehicle, covering brakes, steering, lights, tires and the rest of a fixed list, whenever they find a defect (49 CFR 396.11).
Nobody in your business decided any of those intervals. They arrived with the equipment. And most owners, asked on the spot when the extinguisher by the back door was last checked, would have to walk over and read the tag.
That is worth noticing before you sit down to build a preventive maintenance schedule, because it changes what kind of job this is. A meaningful part of your schedule has already been written by people with more authority on the subject than you or any chatbot, and it binds you whether or not it appears on a calendar. The first task is retrieval, not design.
Where the rest of your intervals are already written
Once you start looking for intervals somebody else has already set, they turn up in ordinary places, and none of them require a subscription to read.
The equipment manual is the obvious one, and it is usually specific in a way general advice never is: a duty cycle, a lubrication interval, a filter part number. The warranty document is the one with money attached, because warranty coverage is a contract and its terms typically condition coverage on the maintenance the manufacturer specifies. Whether a future claim survives depends on the wording of your particular agreement, so the useful move is to read your own warranty for what it asks of you and write those requirements down as tasks. A service contract on an HVAC unit or a copier does the same thing from the other direction: it tells you what somebody else is coming to do and when, which means those dates belong on your calendar rather than in your inbox.
Your commercial lease and your insurance policy are worth a pass too. Both routinely assign responsibility for specific systems and specific upkeep, and both are documents you can open this afternoon.
There is one more obligation that is easy to miss because it attaches to the maintenance itself rather than to a machine. If you have employees, OSHA’s lockout/tagout standard covers the servicing and maintenance of machines and equipment where unexpected energization or start up, or the release of stored energy, could cause injury, and it requires employers to establish a program and use procedures for isolating that energy before the work begins (29 CFR 1910.147). The standard defines its own scope and exceptions, so read it against your actual equipment. The point for scheduling is simple: the moment you move from calling a technician to doing the work in house, a new set of requirements arrives with it.
The one part nobody else can decide for you
Harvesting gets you a pile of intervals that cover safety, compliance and the manufacturer’s interests. It does not tell you which of your assets actually matters, because no regulator knows what your business sells.
That is the genuine judgment call, and it has a sharp test: which pieces of equipment stop revenue when they stop? Not which are expensive, and not which are complicated. To take an illustration, in a cafe an espresso machine may well outrank the delivery van, because the van has a workaround and the machine is the product. Rank on that question and the list tends to get short quickly. Three to six items is a workable size to start with.
Everything below that line still gets maintained. It just does not get the calendar reminder, the named owner and the spare part on the shelf.
The job the model is actually good at
Here is where the usual pitch for AI maintenance planning goes wrong, and it is worth being precise about, because the error is subtle.
Ask a chatbot how often to service a specific machine and you will get a confident, plausible interval. It will read like expertise. But you cannot audit it, and the authoritative number for that machine is usually printed in a document sitting in your filing cabinet. Trading a checkable source for an unverifiable guess is a bad trade even when the guess happens to be right, and it is a particularly bad trade here, because being wrong shows up as a failed inspection, a denied warranty claim or somebody getting hurt.
The useful rule is to point the model at work whose output you can check, and away from work you can only find plausible. Intervals fail that test. Your calendar passes it easily.
Because the real mess is not knowing that the compressor needs a quarterly service. It is that you have harvested forty tasks at seven different frequencies and have no idea what that adds up to in a given week. Feed the model the list, the frequencies, the start dates and the number of people who can actually do the work, and ask it to lay the year out and find the collisions. That is arithmetic over data you supplied, so you can verify the answer by looking at it. What comes back is the kind of thing a list hides and a calendar makes obvious: that the second week of every quarter, say, has more tasks stacked in it than there are people free to do them. If you want a structured way to run that conversation, our walkthrough on how to start using AI in your small business in 30 days covers picking one task and keeping a human review step on the output.
Four moves that turn intervals into a schedule
1. Write the short list first. Name the three to six assets that stop revenue when they stop. Do this before any research, because it decides how much effort the rest is worth.
2. Harvest, do not generate. For each asset on the list, and for the building itself, collect the intervals already set for you: the regulations that apply to your operation, the manual, the warranty terms, the service contract, the lease. Put each one in a single document with a note saying where it came from. That source note is what lets a future you or a new hire check the interval instead of trusting it.
3. Give every task a date and a name. A task assigned to the team is a task assigned to nobody. Put the harvested intervals on one shared calendar with one person’s name on each recurring event, and write down what the task actually involves. If the task lives in one technician’s head, that is exactly the case for capturing it properly, which is the ground covered in our playbook on turning a task into standard work instructions.
4. Run the collision check, then set a review date. Lay the whole year out, find the weeks that are overloaded, and move what can move. Then put one date in the calendar to revisit the schedule against what actually broke. A preventive schedule that is never compared against real failures slowly becomes a ritual.
Note what this schedule does not cover. It handles the failures you can see coming on a clock. The ones that arrive without an interval attached, a key person out for two weeks or a supplier going quiet, need a different instrument, which is the subject of our 90-day build for a business contingency plan. Preventive maintenance and contingency planning get confused constantly, and they are opposite jobs: one is for the things you can put on a calendar, the other is for the things you cannot.
What the software costs, and when it earns it
For a list of three to six assets, a shared calendar you already pay for is genuinely enough, and buying maintenance software before the habit exists tends to produce an expensive record of tasks nobody did. Check what your existing tools do first.
When the list outgrows a calendar, the category is a CMMS, and three vendors publish their prices openly. MaintainX has a free Basic tier at $0, with Essential at $20 per user per month billed annually or $25 billed monthly, and Premium at $65 annually or $75 monthly. Fiix also has a free tier, limited in users and capped at 25 active preventive maintenance tasks, then Basic at $45 per user per month and Professional at $75, each available billed monthly or annually. UpKeep lists Essential at $24 per user per month and Premium at $55, though its pricing page does not state whether those are the annual or the monthly rates, so confirm that before you commit.
The free tiers are the honest starting point for most small operations, and the 25-task cap on one of them is a reasonable proxy for when you have outgrown a calendar anyway.
The part that is judgment, not regulation
Everything above this heading is either a rule you can look up or a price on a vendor’s live pricing page. What follows is opinion.
Maintenance software is often sold on the promise of reducing your dependence on the one person who knows the machine. That gets the value backwards. That person is the source of the schedule, not its casualty. A good deal of what belongs on your list is not in any manual, it is in their head, and it tends to sound like the noise a machine starts making before it fails or the fitting that always needs a second turn. A schedule that captures that makes them more valuable and makes the business survivable when they take a vacation, which is a different outcome from making them redundant and a considerably better one.
The other thing worth saying plainly is who this benefits. The gain from preventive maintenance does not land on a spreadsheet first. It lands on whoever currently gets the 6:30 a.m. call when something has failed overnight, and on the crew who lose a Saturday to a repair that a Tuesday inspection would have caught. That is the argument for doing it, more than the repair bill is.
Walk over to the fire extinguisher nearest your desk and look at the tag on it. Whatever date is written there tells you something true about how your business currently handles the maintenance somebody already decided for you, and it takes about fifteen seconds to find out.
If you want a structured set of questions to work through while you build the list, the Equipment Maintenance Schedule prompt at BusinessPrompter.com is filed under Operations and Systems and is described as creating preventive maintenance schedules to minimize downtime. It is a Pro prompt, so it sits behind a paid upgrade rather than in the free set. Use it for the sequencing and the questions you have not thought to ask, and keep harvesting your intervals from the documents that actually govern them.
Frequently asked questions
How many assets should go on the schedule when I am starting out?
Three to six is usually right for a small operation. The test is which equipment stops revenue when it stops, not which is most expensive. A short list that gets followed beats a complete list that gets filed, and you can add to it after the first review date.
Do I need maintenance software for this?
Not at the start. A shared calendar with one named person per recurring task does the job for a short list, and it is worth checking what your existing tools already do before adding another subscription. The vendors compared above make sense once the list outgrows a calendar and you need parts, history and work orders in one place.
What if the manufacturer’s recommended interval seems more frequent than I need?
Write it down as the documented interval before you decide anything, because that is the number your warranty terms and any future dispute will be measured against. If you then choose to run a longer interval, make it a deliberate decision with a note explaining why, rather than a gap nobody can account for later.
Who should own each task?
One named person, never a team or a role. The point of the name is that somebody notices when the task did not happen, and shared ownership reliably produces the opposite. Write down what the task involves at the same time, so the schedule survives that person being unavailable.
