Blog: Information Architecture in VUI
A speculative look at where information architecture is going with virtual assistants, written as an assignment for DesignLab’s UX Academy
In 2016, I received Alexa as a gift on my birthday. Initially, I wasn’t attracted to it. It just sat there waiting for someone to speak to it the whole day. But it wasn’t easy to bring myself to be speaking to anyone other than a human being. The concept of an interactive talking device was reserved for sci-fi movies or some lab innovations — something that I was still not used to.
Fast forward to 2019, I start my day speaking to Alexa — motivational quotes, daily news briefing, daily questions, weather prediction, music, grocery list and so on. The convenience of having a voice interface which can understand me and respond, keeping my hands and eyes free has increased its appeal over the years. That been said, there are countless times I get frustrated when Alexa fails to understand me. The VUI technology has come a long way but we still have a long way to go to improve its design and function, and support users.
In 1996 Philips introduced the first mobile phone with speech recognition used specifically for dialing. Today, most smartphones offer voice control. And since the introduction of Siri, we know that this field has changed dramatically. And then we have Microsoft’s Cortana and Samsung’s Bixby.
Alexa has set the scene for voice first devices. Following closely is Google’s Google Assistant and Apple’s HomePod.
Current Design challenges with VUIs
Voice interfaces pose a substantial number of challenges for usability. In contrast to graphical user interfaces (GUIs), best practices for voice interface design are still emergent
More often than not, I received this response from Alexa — “Sorry, I do not know the answer to that”. I try couple of different ways to frame my question but then haven’t had success. This causes me to abandon questioning Alexa any further and look it up myself, which I could have done in the first place and save time.
With purely audio-based interaction, voice user interfaces tend to suffer from low discoverability.It is difficult for users to understand the scope of a system’s capabilities. Low discoverability often results in users reporting confusion over what they are “allowed” to say, or a mismatch in expectations about the breadth of a system’s understanding. Not much has been done to educate users what more is possible.
- Transcription errors
Oh, this has elicited many a laugh in my home. The most recent one I recall was when Alexa added “20 pounds of living room TV” to my list when the command was “Turn off living room TV”!
While speech recognition technology has improved considerably in recent years, voice user interfaces still suffer from parsing or transcription errors in which a user’s speech is not interpreted correctly.
Role of Designers and Information Architecture
- Considering the long term future of VUIs, designers need to research the linguistic behavior and their contextual expectations of VUIs and determine how content will be organized. Personalization will play a key role and the context will determine how well the command can be understood and executed.
- Over half of the information we process is though visual interaction. So designing a product with no visual interface and still making it as usable and discoverable whilst presenting the content in an easily understandable format without overloading the user, will be a new challenge for designers to overcome
VUIs have a long way to go but with the right design approach which embeds the relevant heuristics and focus on optimal IA practices, and combining equally sophisticated technologies in artificial intelligence, they can be much more powerful products.