AI that works inside your maintenance records

It means putting AI inside the system that already holds your maintenance records, instead of bolting a separate tool alongside it. Day to day that looks like three things: asking questions of your own data in plain language, handing routine review work to a schedule, and getting drafts back (descriptions, work orders, action items) that a person approves before they count.

What separates it from a general chatbot is the source. A chatbot knows how maintenance usually works. This reads your work orders, your assets, and your parts.

Inside UpKeep, yes. Nova is what this capability is called in the app. When you open UpKeep to ask a question, set up a scheduled check, or install an app, you are in Nova.

The one caveat is that AI is not only Nova. Edge sensors do the condition monitoring on your equipment and Intelligence handles the reporting. Nova is the part you give instructions to, and it reads what those two produce.

There is no AI edition and no separate SKU to buy. Nova Apps run on credits; the rest comes with your plan.

No. Output arrives as a draft or a recommendation, and a person approves it before anything is created or changed. That applies to work orders, to safety records, and to action items pulled out of a recording.

It also respects the same role permissions as the rest of UpKeep, so it cannot do something on a technician’s behalf that the technician could not do themselves.

No. Your records are used to answer your questions and run your scheduled tasks. Nova builds up context about how your team works, covering naming conventions, approval steps, and who handles what. That context stays with your organization.

Only where you have the sensors for it. Predicting a failure needs condition data off the equipment, which is what UpKeep Edge sensors provide: a reading crosses a threshold and work is raised automatically.

On an asset with no sensors, you get faster review of what already happened and better scheduling of what you know is due. Useful, but a different claim, and we would rather keep the two separate.

It helps, and improving it is one of the things the AI is good at. A scheduled task that audits asset naming, finds duplicate parts, or flags work orders missing a cost center will do more for your data in a month than a cleanup project usually does. Start there if your records are messy.

No. It removes the administrative work around the job: the retyping, the report requests, the daily read-through of the same queues. Deciding what to work on, diagnosing what is wrong, and doing the repair are still the job. Teams generally use the time it frees on backlog they were never getting to.

Nova is already in your UpKeep account. Open it, describe the check you want, and it writes the prompt for you. Most teams start with a daily review of on-hold work orders or a weekly PM audit. If you want it configured around your own workflows, book a consultation and we will do it with you.