The Silence of the…? Not “Lambs,” but “Visuals.”
A sneak peek at the design challenges faced by user interaction designers of voice assistant products.
Did you know that by 2024, the number of digital voice assistants will reach 8.4 billion units? A number higher than the world’s population (according to the Statista Research forecast). It is fascinating to see the growing demand for voice assistant products, which in turn forces the designers who work on these products to turn out top-notch interfaces that users can easily use.
A decade ago I was staring at gadgets in sci-fi movies, but now we can own one :)
When I first used a voice assistant product and after I was done with the lame questions like, “Alexa, tell me a joke,” or “How old are you?” (Yes, I tried these as well.) Some serious questions also popped into my mind. Especially with my background as a software designer where visual interface is the main driving factor, I wondered about the design challenges faced by the voice assistant products whose visual cues are almost null aside from some light emissions.
I assumed the designers behind these products might have thought through the user challenges as if they are blindfolded and living in a voice-only world while accessing the product.
I started to learn a bit more about the design challenges these products have and understand the basic design concepts of voice-only interface.
Understanding the Basics
The voice assistant or the voice user interface (VUI) is what the user interacts with while giving a command or asking a question. The two major technologies behind VUI are:
1. Automated Speech Recognition (ASR)
2. Natural Language Processing (NLP)
Natural Language is any common language that has evolved within humans and are expressed vocally. For any conversation, the speakers should communicate cooperatively, be it a human or a product. If they miss the basics like relevance, quality, memory, and understandability, the conversation becomes less useful.
The voice assistants recognize the user’s natural language through speech recognition and NLP and translate what the user says. It also can follow commands to produce the output.
Challenges of Voice User Interface
Challenges arise when the user fails to discover the possible actions, fails to understand what the interface says OR if the product fails to deliver the expected result.
Context and memory are crucial for the voice interface.
If the voice user interface (VUI) does not include context and memory, the usability will be limited. Unlike visual interface, the voice interface is more engaged with the user through conversations, and remembering what the user said earlier is crucial to connect the dots for a contextual response.
Voice interface is less scannable than the graphical user interface (GUI).
Unlike visual interfaces, where the user can scan through the instructions or skip, the voice interface needs complete user attention.
Key Design Concepts of Voice User Interface
1. Keep it Short.
Unlike the visual interfaces, the voice interface does not give the user the flexibility to scan through or skip instructions. Users are in a place to listen to what the voice interface says or asks. The voice interface should consider the user’s memory and avoid giving long instructions to the user.
Make sure the interface delivers the information briefly but clearly.
Only provide important or relevant information.
2. Keep it Lively.
Many of the voice assistant products use commands to begin a conversation (like “Alexa…” “hey Google…” ) and occasionally ending commands. But make sure that these commands are only used to begin the conversation and not necessary to use in every sentence that follows to keep it more conversational.
Make it more natural, apply a turn-taking approach.
3. Provide Context.
In a conversation, it is critical that the product remembers the context and responds accordingly. It must also prompt the user and help inform them on how they should respond and take action.
Give hints, direct the user.
4. Avoid Confusion. Clarify.
If there is insufficient information for the interface to respond, the interface needs to request for more input. Sometimes it is necessary to narrow down certain instructions to clarify, however, the interface must also make sure that the user understands the conversation. Design confirmations to help clarify; however, be cautious and avoid using them often, which might frustrate the user.
Clarify and give or get confirmations.
Final Thoughts
To provide a smooth user experience, a voice interface designer should avoid friction as much as possible by applying these design concepts. Make it a great product by conveying each feature’s value and emphasizing how it helps the user save time, improve their lifestyle, and more.
Teaching the user to use the product with the learn-by-doing approach has shown a promising user retention rate.
Next time when you use your intelligent voice assistant products, I am sure you’ll remember some of the issues and key design concepts behind the product.
Alexa… Thank You! Go Home!
