It's possible that the Z9 overlay just isn't keeping up - they'd had that issue in the past. However, I was thinking the same thing and looked over about a hundred different shots couple weeks ago. I looked at where the camera said the AF area was (looking at the image on the back of the camera) and compared that to the sharpest part of the photo in Lightroom. I went through one by one and, overall, if the camera showed the AF point on a given spot, that was the sharpest point. I may re-test this with another type of bird (these were mostly long-necked bird and the Z9 really struggles with those in flight).
Also, keep in mind that often focusing on the body will still yield a very sharp eye - especially if the bird is flying parallel to you. There's also DoF to consider. I tend to fill the frame most of the time so any focus errors are very obvious. Below is a quick screen grab from Lightroom to give you an idea of what I try for (lol, don't always get, but try for!). None of those are cropped (yes, I do get some clipped feathers sometimes).
View attachment 34595
From my experience with other shooters, the bird is usually less than half this size (usually about a quarter TBH) in the frame and this gives them a
lot of extra DOF to play with. Not saying that applies to you, just what I've seen.
I also bring this up because I have seen a few "reviews" where the reviewer was amazed at the accuracy of the camera, but the birds were so heavily cropped that as long as focus was
anywhere on the bird, it would be fine thanks to the extra DoF at that distance. These are all tack-sharp at the eye, but only beaucse I had subject detection off. Subject detection had a terrible time finding the face and eye and would constantly focus on the body - putting the keeper rate to zero. The a1 can struggle with these kinds of shots too, but it gets it more often than not. One reason is that the a1 only looks for faces and eye, whereas the Z9 looks for faces, bodies, and eyes. This means that even if I have something like Small Wide on the face of the bird, if it doesn't see the face or eye but does see the body, it'll move AF to the body. (With subject detection, as long as a wide AF area is anywhere on the subject, subject detection has priority and can place the AF area where it wants on that target). With the a1, if the camera doesn't see the face or the eye, it just tracks normally and doesn't prioritize the body.
I think a HUGE improvement for the Z9 would be an option that told subject detection to ONLY look for faces and eyes and not bodies. That way, you can try to keep the AF area where it belongs and if the camera sees the face/eye, it'll refine the focus area - if not, at least you still have a shot at a sharp image. The way it works now, if the camera doesn't see the face/eye and goes for the body, it can subvert any chance you have at a sharp series.
Sorry, probably more explanation than you needed!