Why designing great voice interfaces matters

Most of us have asked Siri to crack a joke, told Google Assistant to set a cooking timer, or demanded that Alexa play Mysterious Girl by Peter Andre (maybe that last one is just me), but Voice User Interfaces (VUIs) are everywhere, and they’re not going away with more than 1 billion voice searches taking place each month. But what exactly is a VUI? Simply put, it’s an interface where users interact through speech rather than clicks, taps or swipes.

VUIs mark a massive shift in how humans interact with computers. For decades, interfaces were all about graphical user interfaces (GUIs) — buttons, icons and menus. But over the last 10 years, voice interaction has boomed, bringing its own set of opportunities and challenges.

We’ve all experienced frustrating automated phone systems that make us want to throw our phone at a nearby wall, but those systems fail not because of bad tech but because of bad design. So, how do we design VUIs that don’t just work but feel seamless and human? Jakob Nielsen’s usability heuristics, a set of principles for designing great user experiences, were first introduced in the 90s. By adapting these principles for VUIs, we can build better voice-first experiences. Let’s dive into three of these principles and what they mean for VUI design.

What is VUI?

Do you have an Alexa, or Google dot, or do you use Apple’s Siri? Well, these are all Voice User Interface devices. It’s a type of technology found within these devices or tools that allow us to use our voice to interact; using commands or prompts in order for the VUI to answer questions and carry out actions. The technology uses speech recognition in order to understand these, and as time has gone on this technology has improved, even being able to respond to multiple voices, regardless of the tone, dialect or accent. 

Feedback is key 

Imagine grabbing coffee with a friend. Halfway through the conversation, they simply… stop responding. You’d probably wonder if they’re upset, distracted or just didn’t hear you. The same confusion happens with VUIs, without clear feedback, users are left wondering if their command was heard, understood or simply ignored.

When you shout a command like “Alexa, play my playlist!” and nothing happens, it’s a dead-end. Unlike typing into a chatbot, where you can see the text, VUIs don’t have the luxury of constant visibility, so this makes feedback crucial.

Good feedback informs users about the system’s state at all times. Amazon Alexa for example uses colored lights to show if it’s listening or struggling to understand. Similarly, on-screen messages or audio cues can fill the visibility gap. For VUIs, there are typically seven states to communicate:

1. Waiting

2. Listening

3. Thinking

4. Replying

5. Incomprehension

6. Confirmation

7. User speaks

The goal is simple: keep users informed so the interaction flows as naturally as chatting with a friend.

Let users stay in control

VUIs need to give users a sense of control. This starts with making options clear: What can they do next? How can they backtrack or exit a conversation? For example, Alexa’s “stop” or “cancel” commands let users end interactions they don’t want to continue. These “out-of-turn” interactions are vital for helping users feel in charge, not at the mercy of the system.

Another way to improve control is by confirming actions, for instance, if a user says, “Order flowers,” the system can respond, “Did you mean a bouquet of roses?” before proceeding. This small step helps avoid mistakes and keeps users feeling empowered.

Always have a plan B when things go off-script

Mistakes happen. It’s part of being human. In GUIs (Graphical User Interfaces), users can click “undo,” “cancel,” or hit the back button to fix errors, but what about voice? If you accidentally order a bouquet of flours instead of a bouquet of flowers, how do you fix it without spiralling into frustration? That’s why VUIs need to be ready for anything. Misunderstandings will happen — commands will be misheard, accents might trip up the system and users will say things in ways the VUI wasn’t designed to handle.

The key is handling errors gracefully. When something goes wrong, users should never feel stuck, instead of saying, “Sorry, I didn’t get that,” the VUI could ask a clarifying question: “Did you mean play your workout playlist?” This keeps the interaction moving and reduces frustration.

Even better? Avoid errors altogether. If the system needs more context, it should proactively ask for it, for example, if a user says, “Order coffee,” the VUI could follow up with, “What size would you like?” instead of assuming and potentially getting it wrong.

Key takeaways

Designing for voice isn’t about reinventing the wheel — it’s about adapting proven principles to a new medium. By focusing on clear feedback, giving users control, and planning for errors, we can create VUIs that feel natural, intuitive, and even enjoyable.

Because at the end of the day, whether you’re asking Alexa to play cheesy 90s hits or dictating a shopping list to Google Assistant, a great voice interface should feel less like talking to a robot and more like chatting with a helpful friend.

If you’d like to chat more about this topic or have an idea in mind, we’d love to hear from you! Our experts at LAB go beyond design and can guide you through your creative and customer experiences.