by Edward Muldrew

VUI (Voice User Interface) is a term that gets banded about the Voice community regularly. The technological hemisphere is plagued with death by acronym. Team meetings filled with abbreviations and buzz words. Leaving the layman needing a translator to figure out what on earth is going on.  Almost like a barrier to entry for the tech geeks to prevent your average Joe from understanding our incredibly technical architecture.

So I want to delve in to exactly what VUI is from a beginners stand point.

What is it?

Put simply it is an interface that allows users to interact using audio and sound to make selections. In tandem with VUI is VUI design. Simply put is the process of designing a voice interface for users to engage with. Similarly how graphic designers will create templates to represent different web screens.

So what are the basics for VUI Design?

Invocation: This is the phrase that opens conversation up. The invocation name should be specific to the functionality of the skill. For example “Launch Magic Eight Ball”
Prompts: Prompts ask questions to the user and can be open-ended questions “What do you think?” or closed “Do you want to play?”. Prompts are used to illicit user input.
Statements: These are used to deliver information, a comment, or an opinion and do not require user input.

These are the building blocks for our VUI design. What makes it a little more complicated for newcomers is that they can be called different names under different platforms.

Design

We want to imitate human conversation therefore you must work hard to think about how customers will interact with your skill. It is recommended not to use a flowchart to represent how a conversation works.  Flowcharts are not conversational. They are not representative of the many twist and turns and conversation can have.

Situational Design

A voice first-method recommends you stay focused on the conversation. Each interaction between you’re a client and skill represent a turn. Each shift has a situation that represents a changing context. For example, what the app knows and doesn’t know based off the required user input should dictate conversation.
https://developer.amazon.com/pt-BR/alexa/alexa-skills-kit/get-deeper/tutorials-code-samples/build-an-engaging-alexa-skill/module-2

Don’t overcomplicate it
The words keep it simple stupid are used regularly in the tech world. Experts and professionals can overcomplicate matters when it comes to conversational design. Don’t overload users with information to complete their task. Fewer words in responses are often better. Reduce the number of steps to complete a task whenever possible to keep the conversation short. Simplify messages whenever possible. These are just some quick tips to change a poor VUI in to a good one.

Characteristics of a well-designed VUI

• Uses natural forms of communication
  When talking to a machine, a user should not be required to learn a new language or remember the rules. A machine must conform to the user’s paradigm, not the other way around.
• Navigate through information easily
  Your skill’s VUI should provide an easy way to cut layers of information hierarchy using voice commands to find important information.
• Creates an eye and hands-free experience
  Voice interfaces must allow a user to perform tasks while their eyes and hands are occupied.
• Creates a shared experience
  Voice experiences allow users to collaborate, contribute or play together through natural conversations. For example, an entire family can play a game on a single Alexa-enabled device.
• Confirm when a task has been completed
  An audio cue whether it be a sound effect or response letting the user know they have completed their task is paramount. Don’t leave your user hanging.
• Strong Error Strategy
  You need a solid error strategy because unlike other apps conversation is unpredictable. Be ready to handle a scenario where the assistant doesn’t understand or misinterprets the user. Handling this will give a professional touch to your voice app. Don’t frustrate your user by giving them a response which is unrelated to the context of the conversation
• Keep the communication simple and conversational
  We don’t want the user to feel as if they are talking to a machine (although they are). The language they use and responses they receive should use natural language. The machine must conform to the user’s paradigm and not the other way round. Consider this example in a quiz game.  Alexa: “Incorrect answer. Let’s continue.”or “Oops. You’re wrong. The right answer is 1494”.
• Be brief
  Finally, keep your responses as short as they possibly can. Reduce the number of steps to complete a task whenever possible to keep the conversation short.

Conclusion

My point is let’s try not to overcomplicate the process of VUI design. Follow basic principles with the end user in mind. The diagram above shows what I would call a high level conversational model. I’m not saying these don’t have value but I’ll go back to what I said earlier Keep it simple, stupid! For many, voice apps are a hobby and design charts can take away the fun in a side-hustle. As your app grows and grows, I can see the value of in them. For now I think it is important to remember some key tips and build from there.

Sources Used:
https://blog.prototypr.io/fundamental-elements-of-vui-design-8630077a7009
https://medium.com/botsociety/exploring-vui-design-the-basics-and-getting-started-223663ec65ee
https://developer.amazon.com/pt-BR/alexa/alexa-skills-kit/get-deeper/tutorials-code-samples/build-an-engaging-alexa-skill/module-2
https://www.toptal.com/designers/ui/designing-a-vui