A Toolkit note
Choosing and assessing people
A team-roles inventory tells you how someone likes to work. It does not tell you whether to hire or promote them, and it was never meant to.
I hired a good many people over twenty years, and the ones I got wrong I got wrong in the first ten minutes, because I liked them. Liking someone is a good thing, but of course it does not tell you whether they are going to be the right person for the job.
So this is what I would do now, and it is not clever. Decide what the job actually needs before you meet anyone, in a dozen lines: the three or four things this person has to do well, and the one thing that would make them wrong for it however good they were. Then ask every candidate the same questions, in the same order, about things they have done rather than things they would do. "Tell me about a time you had to tell a client something they did not want to hear" gets you a story you can check. "How would you handle a difficult client" gets you a speech.
Score each answer against your dozen lines while it is fresh, before you talk to anyone else about them. If you want an assessment on top, use one that has been validated for selection and says so in writing; a personality questionnaire built for team development will tell you how they like to work, and that is a different question. The Belbin entry on this shelf is honest about that, and so was Belbin.
Then take up the references yourself, by telephone, and ask the one question that matters: would you have them back?