A somewhat less cynical, but hopefully still realistic take.
First, there is a certain "activation energy" for the professor (and even more so, the department/faculty) to pay attention. The reality is that there is a certain baseline of largely-unwarranted complaints about all TAs, professors, etc., since there is a certain baseline of students who complain about everything, or fundamentally don't get (yet?) that at University, they are responsible for their own learning, and won't get spoon-fed. Of course not every situation/complaint fits that mold, but it has to break out of that noise to get acted upon. Complaints that come from articulate students, with generally good performance in the course and in their academic careers to date, that provide concrete evidence, and avoid name-calling, are less likely to be ignored. And multiple such complaints, not necessarily coordinated, help.
Second, immediate remedial action in the current course depends on the professor. And professors vary widely in their interest in teaching and in their willingness to expend additional time. When I used to teach courses, I would try to attend my TAs' tutorial sessions once in a while, but there were certainly trimesters I never got around to it. However, even then, I think if potential issues had bubbled up to me, I would have made the time. I have subbed in and been "acting TA" for my own course when a TA fell ill during the term, and I'd like to think I would have done the same thing if a TA were hopelessly bad.
Professors generally have office hours. If TAs are good, very often the professor's office hours are poorly attended. A student or group of students could exert pressure, deliberately or not, on a middle-of-distribution-in-terms-of-caring professor by showing up regularly at office hours with questions, and being factual (rather than just complaining) about useful help not being received from the TA. If a professor is regularly being asked "elementary" questions, and hearing the TA could not or would not answer them, it doesn't take a rara avis teaching-prioritizing professor to realize they need to do Something.
In all the above, the unfortunate reality is that it is very difficult for a department to replace a TA during a course. So immediate remedial action is essentially limited to the professor or some other teaching coordinator coaching the TA, and/or the professor doing more of the work themselves.
As to consequences for future courses, again it depends. I wish it were better, but the number of universities/departments that have meaningful performance management and professional development for TAs, in the way I have gotten used to in the industry part of my life, is pretty darn small. The reality therefore is just hope that some coaching helps, and/or there is enough of a clear situation that the TA is not assigned to that course next time.
As to removing bad TAs from the future pool altogether, it depends on the supply/demand balance of TAs in that department. I've taught in applied science-type departments where TA-ship is a building block of financial support for all grad students and it would take someone really bad to be taken off teaching--there's a perpetual hope they'll do better next time around. I've been in a math department where teaching was core, but there were more students than teaching spots and poor quality did have consequences. And I've seen language departments where TA/lector spots were highly competitive and a poorly-student-rated lector was just not going to get hired again.
Finally, decades ago it was pretty awful, but Universities are getting better at policing misconduct. So all the above applies to poor quality teaching by TAs. If the issue is blatantly unfair grading, unresolved material grading errors, or--even worse--misconduct towards students (or others), there are usually now independent processes to take care of that. There may well be challenges whether those processes move swiftly enough and are impartial enough, but it's a whole different ball game.
This answer has a North American, science/math/STEM bias.