Oh. Oops. I thought that was google talking. So it's one of those programs that looks at what kind of browser the visitor has got, and builds the page accordingly? If so, I wonder what the plainclothes robots get. Saying outright "bots get the naked version" sounds unnervingly like "this site serves cloaked content" doesn't it?
Where does the redirect happen? Within the page code itself? What does your config file or htaccess say about the original URL?