When was the last time anyone counted what’s in your maintenance shop?

Not a rough idea. An actual count. With dollar signs attached.

There’s cash sitting on those shelves. Parts that don’t match anything installed on the property. And a system held together by one person’s memory.

The maintenance shop at an apartment community is a black hole with a budget. Parts go in, some come out, the rest sits on shelves nobody inventories.

And this isn’t a messy closet problem. It’s an information problem. The part your tech needs might actually be in the shop. Buried behind something, left over from a job two months ago, never logged. But nobody knows it’s there, so they order a new one. The absence of information is more expensive than the absence of the part.

Your maintenance supervisor isn’t bad at inventory either. The incentive structure makes hoarding the rational move. Resident calls at 2 AM with a blown water heater. If the part is on the shelf, the problem disappears by sunrise. If it’s not, the super is driving to Home Depot at 6 AM and the property manager hears about it by 7. Overstocking is quiet. Stockouts are loud. So supervisors overbuy, keep emergency appliances tucked in corners, and nobody questions the spend because the alternative is worse.

Then there’s the fixture problem.

A 300 unit property that’s been through at least two renovation cycles over 15 years. How many faucet models are installed across those units right now? How many supply lines, toilet kits, garbage disposals? The shop has parts for fixtures pulled out in 2019 and is missing parts for what replaced them. Every unit turn becomes a guessing game. What’s in the unit, what’s on the shelf, do those two things match. The default answer is “just order new” because confirming what you already have takes longer than replacing it.

The technology to change this already exists. It’s just never been pointed at an apartment maintenance shop.

A company called AMPAworks builds small AI cameras that mount on shelves in hospital surgery centers. They photograph inventory when they detect movement and auto-count in real time. When stock drops below a threshold, the system triggers a reorder. Healthcare, where a missing supply can delay a surgery, solved this with a camera and an algorithm.

The interesting part for multifamily isn’t the camera. It’s what the camera could connect to. The rent roll. The move-in schedule. The work order queue. Unit-level asset records showing which fixtures are in which apartments.

Move-in scheduled for Unit 204 next Tuesday. Work order shows a faucet replacement. The system checks whether the correct model for that unit’s fixture package is in stock. If not, it flags the gap days before the tech discovers it the hard way.

MarginPoint already tracks parts across multifamily properties with a mobile app and automated reordering, and it works well for teams that use it consistently. Where the next generation picks up is removing the manual input entirely. Instead of a tech logging what they pull, the system sees the shelf itself and cross-references the operational data.

Getting there isn’t simple. Someone has to barcode and map every item in the shop before vision AI can read it. And cameras in a maintenance shop are going to feel like surveillance before they feel like inventory management. The trust conversation with the team matters as much as the technology.

But the core question stands. Every other system in property management got digitized. Leasing, accounting, work orders, resident communications. The maintenance shop is still managed by memory and eyeball.

How long does that last?