Mostly I’m in the same boat at you: I have some visualization ability but I can’t just superimpose such mental images over my vision, without vision completely dominating. The mental object is so fidgety, dim and fleeting that the input from the eye completely washes it out.
But that seems to me to be a matter of the degree to which my mental images are usually pronounced.
In dreams and those weird hypnagogic intervals, I’ve had very vivid mental imagery, that makes me understand the mind is at least in principle capable of presenting mental objects of an intensity that matches sense impressions.
Now let’s say someone’s faculty of visualization is developed to the degree that their intentional mental imagery has an intensity and consistency comparable to that of regular vision.
I don’t see why they couldn’t, with the appropriate effort, “mix” the two in such a way that they are superimposed without any side dominating the other.
The best way to describe it is that I'm copying what I'm seeing through my eyes into my mind's eye, and superimposing the image there. This happens in real-time, effortlessly. I then have two streams of visual information that I can chose to pay attention to, one with the superimposed object and one without. If I chose to focus on the scene in my mind's eye, it's as if I have superimposed the object into my actual visual field. I can switch back and forth between the two parallel visual streams at will.
I'm not sure which option this would place me in.
But this contrasts sharply to music for example, most of the time if I think about a song I can "hear" it; not like I'm actually hearing it, I hear actual reality around me. But the fidelity of the song can be quite sharp, I can hear all the little details as if I actually had it playing.
But then again people talk about having trouble manipulating images in their mind to solve spatial puzzles and the like, and that's no issue for me; I'm quite good at those. And on the other hand, dreaming for me is not like watching a movie. Sometimes there is rough imagery, like what I get when I try to picture an apple, but mostly they're purely abstract: thoughts, words, ideas, not video.
I suspect it all comes down to quirks of how consciousness works for different people, and how we communicate about them, rather than being a hard binary thing you have or don't have.
I can’t at will persist an image in my mind (to a decent extend), but my mind can slap together related images in quick succession, which can serve well enough to rotate things in my mind.
Once I had a couple of lucid dreams though, it became clear that my mind can absolutely show me very vivid scenes.
The fidelity is sadly not there when I try to recall it.
I also think I’ve maybe gotten better at visualizing over time, but the improvements are rather gradual.
I have no mind's eye, and I definitely consider it an advantage. I genuinely thought it was a euphemism until I was about 20, drunk, and surrounded by friends at college, playing a game in the student bar and the "mind's eye" thing came up. They couldn't believe I was serious. I couldn't believe they were serious... For a while at least.
My mind works on rules, not imagery. If I am asked to "not think of an elephant in a room", I (of course) immediately think of an elephant in a room, but it's not a visual picture - it's relationships between room and elephant (does it touch the walls, the space around it, does it press the light-switch on, can the door open if it opens inwards, ...) It's the concept of an elephant in a room. There's no visual.
Similarly, I don't know my right from my left - instead I have a rule in my head that I run through virtually instantaneously "I write with my right". That then distinguishes for me which is which. If someone gives me directions "first right, second left, right by the pub and next right" I run through that rule for the first instance, and then I have the concept of "not-right" for the "second left" bit. It gets "cached" for a while, and then drops out.
So where's the advantage ? I can consciously build these rules up into complicated (well, more complicated than people expect) structures of relationships and "work them". It's not like running an orrery backwards and forwards, but it's the best analogy I can give. I can see boundary conditions and faults well before others do - and often several complex states away from the starting conditions. I'm often called into meetings just to "run this by you" because I can see issues further down the line than most. I'm still subject to garbage-in-garbage-out, but it's still something of a super-power.
I'm told I sort of gaze into the middle distance, and then I blink, come back, and say something like "the fromble will interact with the gizmo if the grabbet conflicts with the womble during second-stage init when the moon is waning". Someone goes off and writes a test and almost all the time (hey, I'm human) I'm correct. Mental modelling is what I gain from a lack of visualisation. I think of it as literally building castles in the sky, except the sky isn't spatial, it's relational.
I think the bigger takeaway is not to think that having a particular sort of richness of any kind of mental imagery (regardless of the sense modality) is somehow automatically better or worse. The mind is malleable and an intelligent mind can find clever ways to exploit what it has most richly, and there are surely disadvantages to imagery of some sensory modalities being too rich or real (e.g. schizophrenia).