by Georgina Titheridge, UX Mastery
September 20, 2017
Voice user interface (VUI) design is booming. Computerised personal assistants such as Apple’s Siri, Amazon’s Alexa, Google Now and Microsoft’s Cortana, are racing each other for the title of best voice assistant on the market.
Since Amazon released its voice assistant device Echo in December 2014, it has sold approximately 8.2 million devices, and uptake of voice search continues to climb. According to MindMeld, 60% of people who use voice search started using it in the last year, and 41% of people started in the last 6 months.
BCC research predicts that the global market for voice recognition technologies will increase from $104.4 billion in 2016 to $184.9 billion in 2021, at an annual growth rate of 12.1%.
This surge is being driven by technological improvements and deep learning, which enables developers to create systems that have exceptional accuracy for tasks such as image analysis, speech recognition, and language analysis.
Last year, Microsoft announced that its latest speech-recognition system had achieved equivalence with human transcribers in recognising human speech.
Voice technology is advancing quickly, and it’s changing the way we interact with our devices.
What does this mean for UX professionals? With companies already struggling to find VUI experts, now is a great time to start developing new skills as this game-changing technology evolves.
If you’re considering upskilling or embarking on a career in VUI, you’ll find your existing UX skill set will stand you in good stead.
While many common UX design processes still apply – including user research, persona creation, prototyping, user flows, usability testing, and iterative design – voice UIs have a few key differences you to keep in mind.
If you’re about to embark on your first voice user interface design project, here are five essential tips to help you along the way.
1. Talking vs typing: make it conversational
It’s crucial to make sure that a voice UI understands natural speech – i.e. it can accept a wide range of different inputs.
We speak differently to how we type, in full sentences or questions, rather than a series of keywords.
Imagine your Sunday morning, you roll over and type in your phone “brunch nearby.” Predictably, a list of results will appear.
However, when we interact with a voice service, you’d be more inclined to articulate your request, “Alexa, can you suggest a good place for brunch nearby?”
Machines must be able to understand and respond to thousands of different commands in order to be successful.
2. Recognition: make it intuitive
No one wants to memorise hundreds of commands to perform specific tasks. Be careful not to design a system that is difficult to use and takes too much time to learn.
It is the machine’s job to remember us and become more efficient with each use. It’s intuitive and retains information from the user’s most recent search history.
Imagine you are asking your device for directions, and this is how the conversation plays out:
“Alexa, can you give me directions to work.”
“Alright, where is your work?”
“You know where I work!”
“I’m sorry, you’ll have to repeat that.”
This, as you can imagine, would create a frustrating experience for the user that is neither successful or satisfying.
If however, the system had retained information about your work address, it would quickly be able to give you a list of directions, most likely a short voice response with a visual component of a map and directions – making the experience a rewarding quick and satisfying exchange.
Intuitive design, as with graphical user interfaces, is something designers and programmers need to get right.
3. Accessibility: consider your user’s needs
Voice interactions rely on two things working successfully: the device understanding the person talking, and the person understanding the device.
This means designers should always consider potential speech impediments, hearing impairments, and any other things that could influence the communication, such as cognitive disorders. Even accent, language, or tone of voice can affect how the device understands them.
This requires you to be smart about where and how to use voice and design it in a way that everybody can use it, no matter how they sound or how they hear.
4. Environment: consider your user’s surroundings
Speaking into your phone on a loud busy train is an example of why it’s important to consider how different environments affect the kind of interface you’ve designed.
If the main use is for driving, voice is a great choice—a user’s hands and eyes are busy, but their voice and ears are free.
If your app is going to be used somewhere noisy, a visual interface might be better, as the surrounding noise will make voice recognition and hearing more difficult.
If your app will be used both at home and on public transit, you may want to have the ability to switch between a voice and a visual interface.
5. Feedback: let the user know they are being heard
How do you show people you’re listening during a conversation? Most people nod their heads, smile, repeat things the other person said… all essential cues that show you’re listening.
It’s important to consider this with your design, so the user knows their device is switched on and is paying attention.
The system should always keep users informed about what is going on, and it’s important to consider how your user will know that the system is awake in a non-invasive way. Will it be a flashing light, a sound effect, or responding with an accurate response? You decide.
Voice UIs are set to revolutionise the way we interact with technology. This is creating a huge need for specialised talent to successfully lead this shift in human-computer interaction. For UX designers looking to add another bow to their design string, CareerFoundry has just launched a new online course for Voice User Interface Design, built in collaboration with Amazon Alexa.
It’s a big shift as our devices start to talk and listen to us more, but UX designers are in prime position to adapt skill sets to this new technology.
Do you have any experience designing for voice? Share any tips in the comments, or over in the forums.