Designing Effective Voice User Interfaces: A Practical Guide

Introduction

With the proliferation of smart devices, voice interaction scenarios are becoming more common. This guide, based on Atlassian design director Justin Baker’s recommendations, outlines the unique design challenges of Voice User Interfaces (VUI) and provides a three‑part framework.

Phase 1: Exploration

Understanding the environmental constraints that affect voice interaction helps designers grasp where and how users speak to devices.

Define Interaction Devices

Device form factor determines the interaction mode. Common categories include:

Mobile phones – iPhone, Google Pixel, Samsung Galaxy; network via cellular, Wi‑Fi, Bluetooth; visual, auditory, and haptic feedback.

Wearables – smartwatches, bands, shoes; strong scenario focus; network via cellular, Wi‑Fi, Bluetooth; varied feedback.

Fixed devices – desktops, smart home controllers, TVs; network via cellular, Wi‑Fi, Bluetooth; used in fixed locations.

Non‑fixed devices – laptops, tablets, car media systems; network via cellular, wired, Wi‑Fi, Bluetooth; primarily non‑voice interaction.

User Case Analysis Table

Identify primary, secondary, and non‑essential use cases for each device. Create a matrix to understand why users engage with the device, which interactions are core, and which are optional.

Phase 2: Input

After exploring constraints, focus on how devices listen to user commands. Key interaction nodes include trigger cues, wake‑up feedback, listening feedback, and end‑of‑listening signals.

Trigger Cues

Voice cue – a spoken phrase activates listening.

Touch cue – a button press.

Gesture cue – a hand wave.

Self‑wake – device wakes based on predefined commands.

Wake‑Up Feedback

Devices should provide immediate auditory, visual, or haptic signals to confirm they are listening.

Listening Feedback

Timely visual feedback (e.g., Siri’s waveform).

Audio playback of recorded speech.

Real‑time transcription.

External visual signals such as LEDs.

End‑of‑Listening Feedback

Signals that the device has stopped listening and is processing the command, with appropriate pause length and flexibility.

Phase 3: Conversation

Simple commands may not require dialogue, but complex tasks do. Design rules include providing clear affirmation, allowing user correction, and showing empathy.

Give explicit confirmation (e.g., “Okay, I’ll turn off the lights”).

Allow users to correct misunderstandings.

Show empathy when the AI cannot fulfill a request.

Advanced Scenarios

Personified Interaction

Adding personality through lights, animations, or synthetic voices builds trust and emotional connection.

Dynamic Interaction

Maintain seamless transitions, lively feedback, and efficient processing cues, especially for complex tasks like cooking guidance.

Conclusion

Voice UI design is multifaceted and still evolving. As digital devices become more pervasive, the time spent interacting with them may surpass human conversation, making VUI a potential mainstream interaction method.