How to Design for Voice

Posted / 25 March, 2021

Author / Enginess

How to Design for Voice

Voice technology is becoming more prevalent (not to mention better) with each new iteration. AI and cloud computing, plus the fact that there’s a massive amount of recorded video, give us the tools to make products that can reliably understand human speech.

So now, we need to design for that reality. 

And while a lot of design practices and principles carries over, there are voice-specific things to watch out for.

Here are a few rules of thumb to consider when you’re designing for voice.

1. Decide if You Need a Voice App

First off, figure out if a voice application is good for the user. Sometimes, this will be obvious. For example, if you look at the out of the box Alexa apps they were mainly focused on things like music, the weather, reminders, and small queries and questions. 

Basically, making connections between people and the simple information they wanted.

Now, contrast that with something like buying a car. That’s a much more considered decision, that people will research and review to make sure they buy the right one. Even if they could ask “what car should I buy and why” they wouldn't be able to retain that information. So in that instance, a voice application for, say, a used car lot, might not make a lot of sense.

So before you start designing, make sure a voice application is the right modality.

2. Understand Your User Expectations

How frustrating is it when you’re on the phone with an automated system and they don’t understand what you’re saying? 

That’s because (1) you’re on the phone with an automated voice system, which no one ever calls just for fun, and (2) our expectations around voice are incredibly high. Talking is so critical to human communication and as humans, we’re so good at it, we expect machines to be the same way.

So one of the first things you need to do is understand what your user’s expectations are around speech. What do they want to do with voice technology, and what are they going to say to get there? An intersectionality approach can be valuable here, paying particular attention to the specific circumstances in which they’ll be using your voice application.

3. Consider all the Permutation of Your Commands

You’ll likely be building on top of a voice system like Alexa, which should take care of a lot of the word recognition and order for you. However, you’ll still need to design a way for those words to trigger specific commands, which means you’ll need to think through all the various ways your users will be framing those questions. 

For instance, let’s say you're a headphone shop and you want an Alexa app to be able to answer questions about where to get headphones. 

People might say: where can I buy headphones, cheap headphones near me, headphone price, how much are headphones, how much are earbuds, how much are airpods… you get the idea.

If you’re designing for voice, you need to at least be aware of the various flavours your app can take.

4. Set Parameters for the Options Available

A good way to control for the problem above is to set clear parameters around what users can and can’t do with your voice application. If you do that, then you can avoid a lot of user frustration (even if the functionality is limited) because it will make it much easier for users to learn quickly what they can use your voice app for.

5. Create Other Feedback Systems 

Voice is tricky for designers because the feedback is all audio. However, that’s easily missed by users who are busy, which is the context voice is most often used in. So if you can, try and design some other feedback as well. For example, you might have a red light pulse if the command wasn’t understood. Or alternatively, you could have the light change depending on if the application is listening or processing information.

It’s also important to design audio cues as well around when your command has been received and you’re getting the answer. Otherwise, users won’t know if they’ve done it right or not!

6. Design in Conversations, Not in Flows

Any voice application is fundamentally about a conversation between a robot and human. Basically, it’s a two-way street. 

So that’s how you need to design your voice application. Think through questions like: ‘When a user asks, X, what will the voice app’s response be?’

This isn’t to say that flows are not necessary — they’re often a good way to represent the technical side of things. But from a UX perspective, switch your thinking to be conversation first rather than flow-first. 

Wrap Up

Voice is increasingly a major interaction point for people. It’s going to emerge as the number one way humans engage with the internet, extract information, and make requests and decisions. 

Fortunately, the 20+ years we’ve all spent getting really good at designing for other modalities will come in handy — many of the same principles apply. 

But there are some quirks, and designers need to be aware of them if they want to build functional voice apps.

Plan your project right - a step-by-step guide to ensure a successful digital project launch. Read now.
Guide: Achieving Success On Your CMS Project


See all ≫ ≪ Hide all

Subscribe to Enginess Digital Insights

Share the insights /