In the last year, chat bots and conversational UIs have been one of the hot topics. I believe that when we’re working with conversational interfaces, knowing how conversations work is a key requisite. That’s why I want to introduce you to conversation analysis. (slide)
Conversation analysis is a tool for studying, well, conversations. So how does it work? (sld) Basically, conversation analysts record conversations and then systematically look at what we do when we talk to each other because even though conversations may look messy, they are in fact highly structured. (sld) There’s certain patterns that keep reoccurring and certain unspoken conversational rules that we all orient to—but they’re often about small, seemingly unimportant details such as how we ask a question or how long we pause so that when we try to think about them, we often don’t remember how we do things.
Let’s look at an example. Let’s say we had to write a phone call between two people. It might look like this. (sld) The problem is, an actual phone call, when we transcribe it, looks more like this. (sld).
It’s different. It looks like there’s a lot more going on. (sld) And I know, at first this looks much more messy. But, if we take a close look, we can see that in this mess, there’s lots of orderliness and that a lot of these details are quite systematic and quite important in the conversation.
I can’t talk about everything, so I just want to look at these ‘mhms’ (sld). Yes, they look unimportant. So why should we care? Well, if we look at where they’re said, we can see that they’re used very strategically. Rita always says that after Evelyn has said something that’s new. (sld) These ‘mhms’ are one of the natural mechanisms that we use for affirming content.
I want to mention one more part. (sld) The beginning. Some time ago, I came across a blog article that argued that (sld) “Hey Google” was a much better command than “Alexa!”. I couldn’t agree more. Because of this. (sld) Greetings are something that we do all the time. We never barge right in and just say what we want to say. We almost always have something that comes before. Very often, that’s a greeting. Consequently, “Hey Google” sounds more natural to us than “Alexa! Attention! Do A.” It’s these little things that make a conversation.
(sld) There’s one key thing to take away from this when working with conversational interfaces. (sld) Use real conversations. In our day-to-day interactions, we follow lots of unspoken conversational rules. When we invent conversations, we often don’t remember these fine-grained conversational strategies. And yes, at first, using real life conversations may look messy, even scary. But if we move past that (sld), we’ll see that there’s lots of systematicity and structure at work. We as information architects can be the ones to embrace these structures and use them to architect good conversations. (sld)