Past the hype
There was a period when every organisation had to be doing something with VR. That period is over, and that is a gain. What remains is a tool with a narrow but real field of application, and the question is no longer whether it impresses, but whether it pays for itself.
We build VR trainings and we regularly advise against them. That is not modesty but arithmetic: for the majority of the training questions we get, a well-made e-learning is faster, cheaper and equally effective.
The three conditions
VR earns its keep when all three of the following hold. At two out of three it is usually already not worth it.
You cannot rehearse it in reality. A fire in a machine hall, a gas leak at a compressor, a thermal incident in a battery container. Situations too dangerous to recreate, or that would stop production. This is the most important condition and immediately the most underestimated: if you can rehearse it on the installation, do that.
It is about acting, not knowing. The question has to be what somebody does when it goes wrong, not whether they know the procedure. For knowledge transfer, VR is an expensive way to achieve what a module also achieves.
It comes round often enough to repeat. A scenario run through once rarely justifies the build cost. The value arises in repetition: every shift, every new employee, every year again. Then the cost per participant shifts towards something defensible.
Where it does not pay off
The other side explicitly, because that is where most of the misspent money sits.
For pure knowledge transfer, regulation, procedures, product knowledge, VR is overkill. For operations that call for physical feel, torque settings or tool handling for instance, the simulation does not come close to the real thing. And for subjects that change quickly it is outright unfortunate: adapting a VR scenario is considerably more expensive than changing a text screen, so whatever will be different next year is better left out of VR.
There is also a practical limit that rarely gets discussed in advance: somebody has to manage the headsets. Charging, updating, cleaning, handing out. Without an owner they are in a cupboard within six months.
The honest calculation
The question that makes the decision is not what VR costs but what the alternative costs.
Set the build cost against what you do now. An organisation that has four shifts rehearse for a day each year with external supervision and a shut-down installation pays more for that than people assume, and those costs come back every year. Against that background, a scenario that lasts five years and can be repeated per shift is quickly defensible.
Skip that comparison and VR stays an expense with nothing on the other side of the ledger, and then it never wins against an e-learning at a tenth of the price.
Our rule of thumb: start with the most expensive incident you nearly had last year. If that scenario cannot be rehearsed in reality, you have a candidate. If it can, there is probably something cheaper that delivers more.
If VR drops out, the question is which form it becomes instead. That overview is in the guide on setting up or renewing a corporate academy. What it takes technically is under VR and XR simulations. The thermal incident in the first condition belongs to batteries and storage.