On the face of it, option B would seem to be clearly better, but I'm just trying to understand how an approval system can work if it automatically just approves things, that sounds more like slight delay system than an approval system. Maybe I'm misinterpreting, the way I was reading it sounded something like "the process of approving updates would be cumbersome and time consuming for humans to do, that's why the process of calling things approved is automated" but perhaps what you were saying is the "the process of evaluating whether approval should be granted is automated and done by software that can figure out if the update will or won't cause problems and then either does or doesn't approve depending on the evaluation" which sounds great, but I just didn't think that was actually a thing that could be done by software. Is that actually how it works? There's software that can determine if OS updates to phones does or doesn't cause unexpected problems with an entity's existing systems? I just thought for sure you'd need a human to do that given how hard it is to define a 'problem' and how specific the needs of an enterprise would be.
If my initial understanding was correct, that the software just does the job of ticking 'approved' for you, so you don't have to tick it yourself, then I am completely at a loss in understanding how that is any better than simply having no approval process and just allowing updates without oversight since it's functionally the same, except a little bit slower (albeit only a little slower because it's automated).