Prototyping for voice

Thanks for sharing your thoughts on what makes good conversational design. One aspect of good design is prototyping – creating simulations or sample versions of the product you are designing so that you can test out ideas and make improvements before you launch the final version.

To help explore the prototyping process, we asked Fiona Morgan, Chief Operating Officer from labworks.io to talk us through her personal journey of developing an interactive voice prototype for the BBC.

Everyone has heard of Alexa and the Google Assistant; they’re being used to turn lights on, play music and set timers in households all around the world, but there’s much more you can do with these devices. As usage grows, so does the expectation from users on what they can do – and the people designing these experiences right now, are the ones who are defining the platforms.

My experience in voice has been within interactive entertainment. I am the chief operating officer at labworks.io, one of the leading voice game studios worldwide, and our mission is to become synonymous with playing games with your voice. Previously I worked as a producer at the BBC and created the world-first play-along radio on smart speakers with the Greg James breakfast show on BBC Radio 1.

Deciding on a smart speaker experience

I want to take you through some of the decisions Paul Jackson, the Senior UX designer and I made on a project and what influenced them. Starting with the initial idea…

We wanted to focus on what the BBC could offer that made us different to everyone else. It’s a sensible place for any project to start. For us it was talent, access to lots of engaged potential users and high production quality – for you I’d imagine it will be very different. It was very early on that we decided that we would record the voice rather than using text-to-speech, and we knew we wanted to add production to it – music, voiceovers and sound effects. And if we wanted the audience, we needed to go to them, so we settled on making something for BBC Radio 1.

Given that, for many, this would be the first time they’d use their smart speaker to do anything other than for simple transactions (turning lights on, playing music, etc.). It was decided that we wouldn’t make it more complicated by introducing a new game, with new rules that our audience had to learn. This limited us to moulding one of the already successful radio features into something that worked on smart speakers. We needed to keep it familiar so it didn’t require much investment from the user, but also make sure it would be a great smart speaker experience in its own right.

Making prototypes

There are great options like VoiceFlow which allows you to build experiences with no coding knowledge. Trying something out is my best piece of advice. Some experiences you can get a good feel for with a bit of role-play. I prefer to knock up a quick prototype because, for me, the interaction with the device is key, even at an early stage. You don’t know how it sounds and feels until you get your experience onto a device and have someone play it. I must have made at least 20 different games during this time, some more advanced than others, to test out ideas and learn from users. One clear example of something I learned is that if a user says “pass” in a quiz you might naturally send them down the ‘incorrect’ path, but actually that sounds strange to them. Have a go at role-playing your idea and you’ll see what I mean.

This took a good few months; building prototypes, reworking the script, collaborating with audio producers, changing the content… until we finally narrowed it down to something we thought was great.

Refining the product

We were lucky that our prototypes and workshops lead us to one feature, What’s My Age Again?. As with most radio features, it does what it says on the tin, which means the rules were simple to follow even if you’d not played it before. The fact that you only needed to interact using numbers meant that handling the interaction model was much more straightforward and it was relatively short. Bingo!

It wasn’t plain sailing from there, though. We seriously underestimated the amount of recording time needed with Greg James. There were over 50 clips per edition (and we’d decided to do it daily to fit with the breakfast show). The ‘loudness’ level of an Alexa speaker is also different from the levels of radio, which meant a whole lot of learning for our audio producer, Liam Hadley.

Everything came together and it was considered a huge success. Would I do it again in exactly the same way? No. However, I would do it again taking into account all the learnings we discovered throughout the process.

Where to start?

In the next step, we’ll get some top tips on prototyping from Fiona Morgan. Before we do, think about a project you have been involved in before. Have you used any prototyping techniques? What did you find useful?

Share your thoughts in the comments section.