User Experience Designers are often working with desktop, mobile, or even kiosk screens, but the UX design process can be applied to any interaction between a human and a technology system. The Voice User Interface or, VUI, facilitates the interaction between voice devices and users. Common examples are the Amazon Echo 'Alexa', Google Home, or Siri on the iPhone. A voice interface can assist those with disabilities such as sight impairment or manipulation of a screen or mobile phone. The interface allows these users to achieve greater independence. All reflect current states of the VUI.
The UX designer has many things to consider when creating for a VUI. They must first consider the overall strategy. What are the business goals? How does the VUI solve a problem for the user, and does a voice solution really fit in a meaningful way? Once these have been resolved, VUI UX design moves forward with user research, insights, ideation, and prototyping in much the same way as for any design challenge. But there are differences. Research may focus more on language, and prototypes will likely be voice devices or voice systems. Prototypes are created using tree scripts and software such as VoiceFlow, with publishing to devices such as the Amazon Alexa or a mobile phone for testing.
There are also best practices to consider.
- Time Efficiency - determine as much as you can upfront so user does not need to supply this information.
- Keep it Short - use filtering to narrow options
- Provide Context for your user - help them understand what voice can (and can’t) do.
- Conversation not Commands - the goal in voice is for the user to experience a ‘conversation’, a back and forth that feels natural.
A good example of these put into practice might be for a voice agent (VA) in a Medical Health System. There would be benefits to the business:
- A VA can handle more customers with fewer human resources, increasing revenue.
- Customers can easily make appointments 24/7. They will return more often.
- Training is no longer required. Money saved can go to research.
This voice agent would be a critical touchpoint for the many diverse users of the medical system. For these customers there are significant advantages:
- Ability to call and resolve goals at a time convenient for them.
- The customer does not have to repeat information. The VA can retrieve data like address, phone, physician and history instantly.
- The VA is infinitely patient and can wait while a customer searches for information.
The artificial intelligence (AI) used in such a voice system is built on recognizing ‘Intents,' or user goals. Are users trying to make an appointment, fix a payment issue, or schedule surgery? As the AI voice system learns more about these and other goals, the designer and the developers must work together to improve the interface and evolve the experience.
Of course this scenario, and others, depend on a high level of AI and VUI sophistication. This is evolving quickly in several key areas; the conversation between a human and a system must move beyond one ‘turn,' and be a back and forth conversation, the VUI must have contextual awareness, understanding the person and the environment, and the VUI must have memory of previous sentences and objects within the same conversation.
VUI UX designers must understand the strengths and weaknesses in the existing technology. They must be able to analyze the data that they receive to know where the system is failing and how to improve it. And they have to consider how the option for a ‘hand-off’ to human intervention must occur within the design.
In all, the use of voice technology is moving forward rapidly, and the designer must prioritize the user, ensuring their goals are met and their experience is easy and productive.
