Duplicates are one of the most important features of the entire Q&A system. They are the consequence of the decision to create a repository of knowledge and to try and increase the signal-to-noise ratio compared to forums. In the old times the most dreaded response to a forum post was "that has been asked already, use the search function", without a link to the actual answer and search results that mostly lead to other poor souls getting the same unhelpful response. Duplicates on Stack Exchange sites, if used correctly, are as good if not better as real answers.
They are also one of the most hated aspects of SE that people complain about a lot on platforms outside the network. You cannot read any thread where SE or SO are the topic that doesn't have at least one person complaining about the bastard moderators that just close everything as duplicates even if it isn't a real duplicate.
And while I do think that duplicates are sometimes applied too broadly to questions that are only superficially similar, I suspect that a large part of the people complaining were not the target of a bad closure, but simply did not understand how the duplicate helps them and how it actually answers their question.
Duplicates are answers, but we try our best to make them appear as a bad thing, as shutting down the question. You get a banner with small text that does explain that the answer is somewhere else, if you read it carefully. That banner contains 4 links and 2 buttons, only one of these links leads to the answer the user is looking for (the duplicate). I can imagine that people unfamiliar with the system get confused here, especially if you account for some kinds of banner blindness.
Duplicates often are more abstract and generic. Especially on technical sites the problem a user has is often a specific instance of a much more general problem. The general problem often has been answered with a really thorough canonical answer, but especially less experienced users might get confused when their question gets closed as a duplicate in these cases. They're not getting one answer specific to their problem, but a complex set of answers meant to cover several different cases and explaining the underlying systems and ideas.
To solve the first issue I would show the answers from the linked duplicate in the answering area of the original question. That's the area where you look for answers, and putting the real answers there is more useful than having them behind a single link above the question that people might miss. There are several details here that might not be that easy to handle well, we don't want to confuse people here about the fact that the answers are from a different question. So that should be marked somehow and distinct from a real answer list in some way.
And if you think this idea is flawed and causes more problems than it solves, just put a big banner in the answering area that links in a more obvious way to the duplicate.
For the second part of the problem I think it would be useful to allow close voters to specific in a bit more detail which part of the duplicate answers this particular question. I'm not sure how to achieve this best, to some degree this is already possible by manually posting a comment when closing. But I think it could be useful to attach this more directly to the duplicate closure itself and encourage people to explain this a bit if it's not completely obvious why it's a duplicate. It would also be nice to be able to highlight specific parts of the duplicate answers for this, but that's kind of a big feature that is probably only worth implementing if it can be used somewhere else as well.
I think it is also useful to nudge the close voters towards thinking about the question "how specifically does this duplicate answer the question". If you don't have an answer to this question, maybe the duplicate isn't as appropriate as you initially assumed.
One reason I'm posting this is because I think duplicates deserve more attention than they currently get. Some parts are much harder to fix like better incentives for correct duplicate closures and better tools to handle the large amounts of duplicate questions on sites like SO more efficiently. There is a limit to simply adding more features, queues and stuff on top of the Q&A system, we also have to improve the core primitives like duplicates.