5 Lessons Learned from Voice Design Experts

This month, AnswerLab hosted five leaders driving the conversation around voice at our San Francisco office to share their best practices in voice design and ways to improve the user experience of voice interactions. Speaking to a packed house, our panelists gave their insights into the future of voice technology and left us with these top five lessons on how to approach designing for voice.

Lesson 1: ‘Doing it Right’ Relies on Context

“When you’re in the car and you’re driving, Cortana will alert you via voice if a text comes in. It’s context-aware. Any voice experience that takes into account whether you’re at home or on the train, and what’s going on or who’s in the room, is doing a good job of using that to enhance the experience so that it changes based on where you are.”

-Cathy Pearl, Head of Conversation Design Outreach, Google

Just like designing for any other user interface, good voice design takes into account what the user is currently doing, any relevant events leading up to the voice interaction, and other data that might give context to the user’s needs. Whether the user is in the car or the kitchen will significantly alter how they interact with the device, and failing to consider this can make or break your user experience. To get the context right, we must get rid of closed systems so each action and data point can build off each other to create a continuous conversation.

Lesson 2: Users Have Higher Expectations for Voice Interfaces

"As a graphical user interface researcher, you have to probe at the end of the task to understand how they feel. You don’t have to do that with voice. You can read it on their face--it is very evident if they’re enjoying this. There’s huge opportunity for delight."

-Matt Howell, UX Researcher, Shopping, Google

Voice assistants are more human-like than any other interface, and with that comes high expectations and high stakes. If you get it right, users are overjoyed, but if the interaction isn’t successful, they give up quickly. At our panel, when asked if they’ve ever gotten angry at their voice assistant, nearly every audience member raised their hand. To combat this, designers must be intentional about putting out a good product based on sound research. Sometimes this means focusing on your one perfect skill rather than attempting too much and degrading the overall experience.

Lesson 3: Voice Comes with Ethical Considerations

"The next generation after Gen Z should be Gen V. This is the Voice First generation. […] We know children are interacting with these little boxes, [...] and how you model your interactions with them is important. As designers and thinkers throughout this space, I feel a heavy sense of obligation for us to go the right way in this."

- Katie McMahon, Vice President & General Manager, SoundHound Inc.

Do we need to build “politeness” into our voice interfaces so children don’t learn to “bark” commands into their voice assistant? Will this affect how they interact with others in their everyday lives? Even if people across cultures and nationalities prefer female voices, should all voice assistants be female? Does this reinforce gender stereotypes? These are some of the questions VUI designers must think about when developing their product. There are no easy answers.

Lesson 4: Voice is One Piece of a Larger Interconnected Puzzle

“If we start thinking about voice as a separate thing, then we start designing only inside of voice. But we shouldn’t. It is part of an end-to-end journey. [….] We need to figure out what context you are in and what you’re trying to achieve, and then respond back in the right way. Do we respond with voice? Do we put something visual on your phone or your TV? How do all of these devices come together to produce an interaction that is seamless to your world?"

-Pree Kolari, Senior Director, Design, eBay

Our panelists urged designers to always focus on adding value to the user, and while some skills work well in voice, others do not. Designing voice experiences in a vacuum can often do the opposite of adding value. Remember to consider which platform is best for showing your content, whether through a voice assistant, mobile app, or on a website, and design accordingly.

Lesson 5: Voice Assistants Have the Power to Reach Broad Audiences

"Every system should when appropriate and contextually relevant have voice input as an option. The options and potential for accessibility is huge. [...] If you have no problem using a mouse and a keyboard and looking at a screen, you don’t realize how much motor skill that requires. We take this for granted. [...] I see huge potential for voice in making apps accessible to far more people."

- Erika Hall, Co-Founder and Director of Strategy, Mule Design

Voice assistants have the power to allow more people to access technology. From people with disabilities who can now join our interconnected world via voice to the elderly who can use voice assistants to help with home healthcare and memory loss, voice technology, when done well, can make technology faster, easier, and more inclusive. Unfortunately, due to limited data sets for training AI, only a narrow range of voice types are recognized, meaning many users feel isolated from voice technology. So, while voice assistants do have the power to reach new users, in reality, we have a long way to go to be truly inclusive.

Want to learn more about how you can ensure you’re launching your voice design with the right intents? Download our smart speaker research report or read more about our voice research for more information.