While I agree that the OODA loop is most often presented as a (fairly obvious) decision model for individuals, the model itself is not the big idea. Rather, the primary value comes from the realization that in a competition (like war, or business, or sports, etc) between two or more individuals/groups the ones who can "cycle through" the OODA loop will be able to adapt faster and often gain the upper hand through superior decision making.
In the context of the military, there are ways of reorganizing your command structure to enable faster OODA loop cycling. For example, a major driver of the "slowness" of traditional armies is their centralization of command. Propagating new intel up the chain and orders down to the troops takes a lot of time, especially when intermediate nodes keep dropping out. If you can delegate your decision making to the lowest possible level, this will make the average decision slightly worse, but because you can make each decision much faster you can still come out ahead overall. This is one of the ways an organization can "practice" the loop. (And coincidentally, one that growing startups often struggle with since it is very difficult to transition from direct command to delegation based command)
I also don't agree with your claim that you can't "practice" the loop on an individual level. Anyone who suffers from indecision in the face of uncertainty and overwhelming options ("analysis paralysis") should know that it is something you get better at over time, especially when you need to be doing it under time pressure.
Source: Was a Navy officer for 14 years, we had tons of discussions of "how to get into the opponents loop" during briefings and trainings. Note that in the military it is sometimes possible to actively slow down the opponents OODA looping, something that is probably illegal for most civilian companies. (Though see https://www.joelonsoftware.com/2002/01/06/fire-and-motion/ for a legal example)