Ah Conaill, the question I've been waiting for! I was beginning to wonder if anyone was going to get to it. There are certain things that I don't want to say because I would be violating my NDA, so I'm going to be very careful about what information I divulge, and don't divulge. Yes, my department deals with a great deal of confidential information.
I've been in my current position for about 7 months. Customer service team members become "point people" who are the primary contact people for the various brands. These people have regular meetings with both the business teams and the R&D teams, just to find out what's new and what's going to be new. I could tell you about much of the 2004 and 2005 product schedule. Of course there's no way you'll convince me to do that.
My access to R&D is good. I have access to their private email folder where I can post questions to them. These forums look not at all unlike message boards. When possible, the original designer interprets the intention of the rule for me. Often times this isn't possible because the original designer is no longer with the company. In these cases anyone in R&D who wants to take a crack at it can give the "official" answer.
I also have regular meetings with people like Ed Stark and Christopher Perkins, where I sit down with them and ask them some questions that have come up that aren't covered in the erratta. I often ask them questions about things I have answered and think I'm correct on, just in case I'm wrong. Amazingly enough, I have been wrong before, which causes a shift in the way questions are answered. Once in a great while I get multiple emails from a customer and their friends on the same subject and they get different answers. Normally the second answer given is the correct one if they contradict one another. IOW, I doublechecked the second time it came up and found out my first answer was wrong. Oops!

Let me follow this up by saying that I usually am right the first time, so if you email, please don't send several on the same subject.
My personal interraction with R&D goes a little further than most point people because I also work as a freelance designer for them (I sit here writing this rather than finishing up my work on a fairly large project that's due Monday). After a few Dragon articles and some work for Bastion Press, the decision was made to let me do some freelance work on some D&D and related books. This decision was actually made before I became part of the CS team or the D&D point person. You'll see my first contribution to D&D in October with the Book of Exalted Deeds. This was designed primarily by James Wyatt, but Christopher Perkins and I both did sections of it (For everyone whose ever read the interviews, yes James really is as nice of a guy as he appears to be). What this means is that not only am I answering rules questions, but I'm also helping to generate new rules.
People get wrong answers for two reasons. The first is that the person who answers the question specializes in something other than D&D. Amazingly enough, we also support a little game called Magic: The Gathering, as well as Pokemon, MLB Sports, and others. I personally am not a major Magic player, but two of the guys in the department are certified judges (I think I got the term right). We all have an understanding of all the games, but some people have expert level knowledge while others don't, and this comes back to the issue of the brand point person not having enough time to answer all the questions.
The second reason is simply believing I know a rule and then being wrong about it. We're talking about a rules system that consists of so many books that its impossible to be right 100% of the time, and it would be impossible to go to R&D for every single issue. Occasionally I get a ruling from R&D that
seems to contradict what's written in one of the rulebooks. In those cases I ask them about the contradiction and they usually say that either the book is wrong, or the rule was fixed in erratta or a later book. This is then backed up with some good solid game design philosophy I had previously known nothing about (and I have the official design guide).
Before I worked in rules support, I mainly did email for the ecom department. In a single day, I could reply to as many as 300 emails. In an 8 hour day, that averaged out to about 37 emails per hour, or about 1 every 2 minutes. That would have been an extraordinarily busy day, but we did have them occasionally. With D&D, each email must be researched fairly extensively. You have to check the book, the erratta, and sometimes sage advice. In a normal day I answer approximately 60 D&D emails, or about 8 per hour. That's one every 13 minutes. Some don't take as long, some take much longer. The average number of questions in a single email is 3. The most in one email I've dealt with had 22 questions. I tend to see questions from many of the same people several times per week. Also, with every new release, the number of questions we get on a daily basis tends to go upward. This should be fairly obvious, but based on the numbers I've seen, there are more emails coming in on a daily basis now than there were immediately after the release of 3E.
The type of questions range from simple, easily answered stuff that comes right out of the book to questions with feat/class ability interractions. The easy ones are usually answered with the book and page number where the answer can be found. The more difficult ones are researched and sometimes put to R&D.
Anyway, that's a pretty good peek behind the curtain. Now, back to writing.