VUI: The Future of Voice Interface Design Explained
How Common is Voice Interface Design?
The overall trend toward VUI has been accelerating as processors get smaller and more Internet of Things (IoT) devices come online. All of the “Big Five” tech firms – Google, Apple, Microsoft, Amazon and Facebook – have invested heavily in voice interface as the way forward.
In 2019, an estimated 59 percent of online searches came through voice, with some industry specific searches posting much higher (for example, food and restaurant search reached 68 percent). The speech and voice interface market sector is on track to hit $24.9 billion by 2025.
Voice interface design uses speech recognition to allow users to engage with technology using voice commands. As the world becomes increasingly fast-paced and information-dense, voice technologies are challenging the dominance of the graphical user interface and can make the experience much smoother. Keyboards can work for text-intensive tasks like writing a blog, but for simple tasks like searches, post updates, adjusting controls, and getting directions, people would rather just talk. They can save many steps, for example, typing in an address on a small screen, but the experience must function flawlessly. In reality, voice interface design works together with tap and swipe technologies to create a better user experience.
How to Create Effective Voice Interface Designs
There is only one thing users want from a voice interface. They want their devices to recognize their commands instantly, without mistakes. Humans are astoundingly adept at understanding voice messages despite background noise, missing syllables, variations in volume, accents and word meanings that are constantly evolving. No AI can match that ability to understand spoken voice commands yet, so it is up to designers and developers to think of every contingency. Here are a few use cases of excellence in the field.
VUI Use Cases
Any device or app could work with speech, but that doesn’t mean they all should. Intelligent voice user interface design begins with an examination of where speaking will make life easier for the user. For example, devices that will be used by the public outside, like at a drive-up window, present too many problems in terms of voice variation and environmental noise. Also, whenever the user is dealing with sensitive information like health data or confidential information, speaking out loud is the last thing they want. VUI is perfect for any tasks that benefit from hands-free use (like cooking or driving apps), smart home devices, and apps that have an emotional component (like exercise or productivity). User studies will let you know exactly where the points of friction are and how VUI could improve the UX.
How VUIs Process Information
Humans can process voice information at about 39 bits per second, so VUIs must strive to process information as fast as possible without sacrificing ambiguities of meaning. In most cases, designers have built conversational AI for information processing or outsource this portion to open source speech recognition engines. Users become impatient when the device requires them to speak slowly or one word at a time. They also may be unfamiliar with core concepts based on their generation. For example, expressions like “clockwise” and “broken record” might not mean anything to younger users. Spend as much time as possible getting familiar with the way your users converse and design the VUI’s personality around that research. Build a rich library of utterances, commands, value and intentions to streamline information processing.
What to Consider When Designing VUIs
The first and most important element in a design plan for VUI is the “customer journey map,” where designers specify what the user needs most at various stages of engagement. If there are competitors that have already introduced working versions of their product, half of that work can be captured easily through competitive intelligence. Otherwise, it will take a significant amount of time in gathering user feedback.
With a map in place specifying what VUI must do and expected outcomes, the burden shifts to defining user personas, how the VUI will interface with various devices and a library of triggers.
Creating a persona is a critical first step to keep the team on the same page about what the user wants to accomplish. The VUI user persona expands the usual attributes, such as name, role, expectations, and so on with additional details on their tone of voice and typical word choices. Ideally, your VUI should be robust enough to automatically recognize different people or different setups for the same person. For example, a smart thermometer might have different user personas for “work at home” and “entertaining guests.” A smart car app could have different radio and navigation settings based on who is driving that day.
There are four main types of devices that depend on voice interaction and each comes with its own set of constraints. Consider these factors when designing for each.
- Phones – carefully review design specs, consider how changes in the environment affect the VUI, graphics can help guide user inputs
- Mobile devices other than phones – car-based devices and laptops will have intermittent connectivity, voice is usually not the only or primary UI
- Wearables – most are single-use devices with limited functionality but may be paired with a phone or laptop app
- Fixed devices – probably must adapt to multiple users, interacts with other devices in the area
There are also four main types of triggering events to consider:
- Voice – the primary trigger for both launching the VUI and providing input
- Haptic – pushing a button like a microphone or setting controls like turning a dial
- Motion – often wearable and fixed devices take actions when a person enters the room or waves their hand
- Periodic – VUIs can be set to remind the user at specific times or when a goal is reached like desired
Some car-based VUIs have also begun to incorporate augmented reality data like reminding users to shop for needed items when they drive near a preferred retailer.
UX and Voice Interface Designs
Designing a voice UI that meets user expectations can be very challenging, even for experienced UX designers. Users have been accustomed to making adjustments when an app or device needs more input, but the psychology of repeating a spoken command is different. When a VUI asks for a number of tickets, people tend to respond with answers like “A couple,” or “Just me and my wife.” It’s impossible to plan for every contingency, so designers must come up with engaging scripts that guide users in a friendly, conversational manner. One handy reference is Google’s guide to conversational development, based on thousands of hours of research.
Designing Adaptable and Accessible Interfaces
Specs are always evolving and radically new devices are coming online all the time, but the human voice is a user interface you can rely on for the foreseeable future. Processors are shrinking and IoT is bringing everything to life. In the world that’s emerging, there is little room for any other interface other than sound. The immediacy and versatility of voice form a direct link from user desire to device response. To reap the benefits of VUI, though, designers must start from detailed personas, work around potential limitations of specific devices, and make the best use of all potential action triggers.
Develop, Improve, Repeat
No matter what project you have in mind, iterative improvement is the foundation of great design. When your design team has the tools to stay coordinated and aligned, they can improve the code faster, make smarter design choices, and amaze end-users with experiences they’ve never seen before. UXPin merges top tools for design and engineering into one world of better, faster product development. Sign up today for UXPin 14-day free trial and see how it feels to collaborate flawlessly across distances and departments.