As a consumer, lately I’ve come across more personas than ever before. Amtrak and Greyhound both greet callers with a friendly, “Hi, I’m so and so, your personal guide…” or some such thing and I can’t really say that I’ve had that trickle down to me on this end before. I’ve designed systems with personas in mind and heard discussions of other implementations in the past, but never had I actually felt the result until now. In both of these cases, the dialog itself was fairly natural and easy to use, but knowing the supposed name of my “automatic agent” wasn’t the reason why. It was inconsistent. At no other point in the dialog did the system personify and start talking to me like an entity who warrants a name, perhaps making small talk about the weather or bantering about how slow their computer is, like a live agent sometimes does. I’m not saying that’s preferable, in fact, haven’t we done the research at this point to refute claims that personas are even desirable? I’m just saying that if you’re going to give me your name, I have a different level of expectations for what may happen next. Maybe that’s the problem, efforts to make systems sound more and more human (however ultimately futile that effort is) set unrealistic expectations. Also, we know that a significant portion of the population is actually offended by the attempt to make systems sound “too” human, whether that is the intention or not. We’re not trying to fool anybody and apparently the public gets mad if it even smacks of us attempting to do just that, so let’s not.
–Simonie
