There is still, I think, an interesting and OP-relevant aspect to this everyday phenomenon of 'prior intent' (the thing that another poster oddly said is not a feature of his daily life). Suppose you get in your car with the prior intention of going somewhere, but instead of 'automatically' putting the key in the ignition straight away, you let yourself wait quietly for a while (or perhaps something distracts you or you deliberately ponder something or allow yourself an 'om' moment of pause, or even at a pinch because you have fitted a hypothetical new gadget to yourself which can measure your non-conscious RPs and you are curious) and then insert the key, when you choose to, what effect has the previous prior intent had?
I would not think that many people other than convinced afreewillists would say that because of the prior intent, the subsequent insertion of the key was not, of itself, consciously directed when it happened, if it felt like that (eg when you snapped out of your reverie and consciously realised you had to put the key in to start the car). And yet, in a way, to get back to the OP, you knew in advance what you were going to do, or at least you had entertained (ie experienced consciously) a mental/virtual prediction about it. Iow, in that sense, thought did precede action. The question of whether non-conscious activity nonetheless preceded any particular conscious thought, at any point in time, is another matter of course. If it did, then the answer to the OP might be, yes, predictions about what they might do aside, humans really don't know what they actually are doing (at any given moment). But in principle, thought did still precede action. Which I think takes us beyond what the various lab experiments we have been discussing have looked at, as far as I know, and into deliberated decisions extending into the future, which some say are a different ballpark.
To temporarily use the model suggested earlier that consciousness, and by extension conscious intent, is possibly a matter of 'necessary or useful attention' and is related to encoding a robust/reliable memory, we might venture to suggest that the prior conscious intent to go somewhere in your car encodes a memory (briefly setting aside the question on the other thread about downward causation and provisionally assuming it to be possible) which is then recalled to facilitate (or at least prompt) the key insertion, when it happens.
This might also help to explain or at least be consistent with reasons why you might alternatively insert the key 'absent-mindedly' (ie without any apparent conscious intent) especially if starting your car is a routine action for you, because in that case (which would be different to you 'open-endedly' waiting for the conscious urge) the memory prompt is operating automatically and non-consciously, conscious attention not being needed in that case. We might think of the former case as you getting into a hire car in a foreign country (perhaps with the steering wheel on the opposite side to what you are used to) or a car that you are not familiar with, so that you do have to look for where the key goes.
Amateur ramblings, as ever.