The shortest route between question and answer
Somebody is standing at a pump and is not sure in which order to shut it down. The instruction exists. It is in a folder in the office, or in a forty-page PDF on a portal they cannot get through on their phone, or in the e-learning they did last year.
In most organisations the distance between their question and the answer is several minutes and a few barriers. That sounds like little. It is enough to make somebody do it from memory, and that is where a category of mistakes begins that never shows up in the training statistics.
A code on the equipment brings that distance back to five seconds. No new didactics, no new platform, just a shorter route.
QR or NFC
The choice often gets made on instinct, and there are sober arguments.
QR is a printed code. It costs practically nothing, works with any phone camera without an app, and you can make it as large as you need. It needs light and line of sight, and it has to be clean enough to read.
NFC is a chip in a sticker or tag. It works in the dark, works through dirt and works through paint, and you only have to hold the phone against it. More expensive per unit, not every handset supports it equally smoothly, and you cannot tell by looking whether a tag still works.
For most situations it starts with QR, simply because the barrier is low enough to try it today. NFC earns its keep in places that are dirty, dark or wet, and on equipment that gets heavy use.
A third option that gets forgotten: both. The same destination, two ways of reaching it, because one is cheap and the other works when the first does not.
What breaks in practice
Labels wear out. A printed code in an outdoor arrangement is unreadable after two winters. Use material that can take it and fold replacement into the rounds that happen anyway.
Gloves. Operating a phone in work gloves is awkward. Scanning still works; clicking through to a third-level menu does not. So the destination has to be right immediately, not a landing page.
No signal. In a basement, a tank or a concrete hall there is no network. This is the most important requirement of all and the one most often skipped: the instruction has to be available locally, or downloaded in advance, or the code is useless in exactly the place it is needed.
Zoning. In an ATEX environment not every phone is permitted. That decides who can scan at all, and it is a question for your HSE colleague rather than for your supplier.
What does the code point at
This is the question that decides whether it still works in three years, and it is organisational rather than technical.
A code pointing straight at a file is a time bomb. As soon as that file gets replaced, moved or renumbered, the sticker on the machine points at nothing, or worse, at an outdated version still sitting there.
So have the code point at a fixed reference that itself forwards to the current version. Nothing on the sticker ever changes and you manage the content in one place. The same reasoning as everywhere else: one source, several entrances.
Write down who owns that reference. Without an owner this drifts apart within a year, and then there is a code on equipment promising something that no longer exists. That is worse than no code.
Starting with ten machines
This technology suits a large project badly and a small start excellently.
Pick ten pieces of equipment that generate the most questions, or where a mistake is most expensive. Make sure those ten have an instruction that is correct and readable on a phone. Stick the codes on. After a month, ask who used them and what for.
That conversation usually produces two things. You hear which instructions were no good, because people scan and give up anyway. And you hear which equipment should have been on the list too, from the people who stand there daily. That list is more usable than any inventory made in advance.
How this fits into a broader training offering is in the guide on setting up or renewing a corporate academy. What it takes on the app side is under learning apps.