what_is_conversational_ux

Have you been navigating the booming chatbot industry for a little while?

If yes, chances are you have, at some point, run into the term 'conversational UX'. If you haven't, you most certainly will.

The often misunderstood term is at the centre of many debates and tribulations within the community. In this article, I want to define exactly what conversational UX is and expand on why it matters to chatbot builders.

 

What is conversational UX?

Conversational UX is the way a user experiences an interaction with a business or a service through an automated conversational interface that learns along the way.

Crystal clear? Great, see you next week!

 

Really, though, what does it mean?

Admittedly, the definition wasn't super understandable so let's expand on it.

Conversational UX is, essentially, how you (the user) experience chatting with a machine that learns from what you are saying. As the back and forth of messages between you and the machine goes on, the bot learns about what you want and adapts until it finally delivers what you were looking for.

Let's say you were looking for blue shoes and decide to engage a shopper chatbot (easier to talk to than to say out loud!). The conversation could go like this:

 

         conversational-ux-chatbot-example-inputconversational-ux-chatbot-example-output_min

I have made this bot up. To experience something close to this, talk to the eBay bot!

 

As you can see, the chatbot did not just interact with you on a superficial level, it learned from each of your messages, retained the information, and adapted until it was able to finally deliver the thing you were looking for.

The conversational UX, here, refers to the way you as the user interacted with the machine.

 

What defines good or bad conversational UX?

There are many different ways the conversation above could have gone, some worse than others. For example, the back and forth above looks like pretty good UX compared to this one:

 

You: I need blue shoes

Chatbot: Great. Please use the command ITEM - COLOUR - SIZE - BRAND - TYPE

Y: Ok I want blue Nike shoes size 8 for sports.

C: Sorry, I did not get that. Please use the command ITEM - COLOUR - SIZE - BRAND - TYPE

Y: Shoes blue 8 Nike sports

C: Sorry, I did not get that. Please use the command ITEM - COLOUR - SIZE - BRAND - TYPE

Y: SHOES - BLUE - 8 - NIKE - SPORT

C: Here is a selection for you to pick from.

 

Although the end result is the same (you received your selection of shoes to pick from), the experience in the second example is significantly worse for the user. He/she had to fail three times in a row before learning the exact and unique way the chatbot would finally understand what it needed to display.

Good conversational UX is seamless.

The user should, ideally, not even realise what is going on. The user should not learn anything new to make the interaction successful (commands, for instance). The user should not use anything other than the conversation (text or voice inputs) to make the interaction successful.

As you can imagine, this is no easy feat!

good-conversational-ux-seamless.jpg

 

Why conversational UX matters to chatbots

Conversational UX defines the way we build chatbots.

There are infinite ways to reach the end goal of a user interacting with a chatbot, but which one makes the most sense? Which one is going to make the user feel good about this interaction?

Answering this question requires two types experience.

The first is field experience. Building chatbots and seeing what works versus what doesn't and in so doing, gaining knowledge you cannot get elsewhere.

The second is experience in developing artificially intelligent chatbots. As we have seen above, conversational UX has a component of learning into it. The chatbot learns from the interaction with the user and adapts to it. That requires AI.

As you can see, achieving great conversational UX demands knowledge and experience - nothing a DIY chatbot builder can help you with.

 

Are we really there yet?

I want to end on a short note about the current state of the chatbot industry and reassure the more chatbot-savvy readers.

I am not here saying perfect conversational UX is happening right now. I am also not here saying it is even attainable yet. Even further than that, I don't believe we should strive to attain perfect conversational UX right now (more on that in a future article, although it relates to why I believe faking it is a bad idea in the chatbot field).

We are at a fantastic point in the chatbot industry where we get to learn the best practices of building a robot that interacts with humans. We have enough technological knowledge to finally start this process.

Exciting times!