Acting without awareness is usually described as a failure of attention. It is more accurately a design feature, and one that carries a specific cost in the places where it is applied to more than routine.
Automation exists because attention is scarce
Conscious attention handles a small amount at a time and is metabolically expensive, so the system offloads anything sufficiently repeated to processes that run without it.
A practised route, a familiar sequence of movements or a well-learned task can proceed with minimal conscious involvement and often runs better that way.
Without this offloading, ordinary activity would consume the whole of available attention and leave nothing for anything unfamiliar.
The transfer works by prediction
Repeated situations build internal models of what usually happens, and the system increasingly runs on the prediction rather than checking the input in detail.
Awareness is recruited mainly when the prediction fails, which is why an unexpected obstacle on a familiar route restores attention immediately and completely.
Everything conforming to expectation therefore passes with little registration, which is why time in highly routine periods is remembered as having passed quickly.
Reactions are automated on the same principle
The mechanism does not distinguish between motor sequences and interpersonal responses, so a habitual reaction to criticism or to a particular person becomes automated too.
These run before any awareness of choosing them, which is why people describe having responded before deciding to and then having to account for it afterwards.
The automated response was appropriate to the situations that trained it, and it continues to fire in situations that only resemble those.
Prediction also filters what gets noticed
Running on a model means the actual input is checked loosely, so changes within the expected range go unregistered.
This applies to relationships and environments as much as to routes, which is how a gradual change in a familiar person or place can be missed for a long time.
The filtering is not laziness; it is the same efficiency that makes routine possible, applied where the efficiency is not what is wanted.
What deliberate attention is actually for
The aim is not to run everything consciously, which would be exhausting and would degrade performance on well-learned tasks.
It is to be able to introduce awareness selectively, particularly at the points where an automated reaction is about to determine what happens next.
That selective capacity is what attentional practice trains, and it is a narrower and more plausible claim than being present for everything.