People often say to me, "you ask good questions." I've certainly had lots of practice. And I enjoy it. The key to starting any search is posing the right question. The basic lesson for any beginning librarian who wants to work on the reference desk is how to find out what the user really wants. People often start off with questions that sidestep the key issues.
Example: Someone asks for a directory of lawyers. You tell them that there are different directories depending on the type of law practice and the jurisdiction they practice in. After a few more questions, you find out that what they are really looking to do is divorce their spouse. Not an easy issue to discuss. So the questions then come around much more to the laws in NJ and the different ways that one can file for divorce, many of which can be done without lawyers. And in fact, if both parties are amicable, it can be done in a no-fault method. So the question about finding a lawyer wasn't really the question. The question was really "tell me about the steps to take and the possible approaches to filing for divorce in NJ.".
So whether it is due to awkwardness or uncertainty, it is not always easy to come up with the question that directly hits on the issues you need answered. It helps to write down your topic - create a mind map or list the key issues you must get answered. Oftentimes, the real question(s) will emerge.