Voice-first design shouldn’t simply copy web or mobile patterns. A voice user interface is the scripted, turn-by-turn conversation between a person and a device, and the medium—spoken language—changes the rules. Unlike GUIs that spread information across space and let users scan or revisit content, VUIs unfold in time, one beat after another. Voice also differs from chat: speech carries pauses, fillers, and mispronunciations, and requires tighter control of turn-taking. Effective VUIs acknowledge these constraints, anticipate the many ways users might speak, and respond in ways that keep the dialog efficient and natural.
Grice’s cooperative principle offers a practical compass for voice design. Quantity: say just enough—no sprawling IVR menus or bloated replies—because listeners can’t skim. Quality: avoid implying capabilities you don’t support; hedge when uncertain and prevent “broken link” moments in conversation. Relation: stay relevant by correctly mapping requests to intents and by using prompts that gently lead to replies your system handles well. Manner: be clear and orderly, writing for speech rather than for print—shorter sentences, familiar words, and messages that acknowledge ambiguity (especially in errors). Read responses aloud and test with synthesized voices to catch phrasing that looks fine on the page but lands poorly in the ear.
Plan conversations before you code. Define the primary intent, sketch likely requests and responses, and explore edge cases you’ll support or decline. Use Wizard of Oz sessions—acting as the assistant with real people—to discover genuine user phrasing, refine prompts, and reduce tasks that are awkward by voice (like long data entry). To avoid a wooden, repetitive feel, add safe variety: rotate equivalent phrasings and tune tone to context (playful for sports, restrained for banking). Overall, design for the time-bound nature of speech, apply the maxims to every turn, anticipate and shape input, and deliver concise, truthful, relevant, and clear responses that move users smoothly toward their goals.
Summary
A VUI is its own brand of conversational interface, distinct from chatbots and GUI, with a reliance on brief, spoken information
A VUI should be "liberal in what it accepts" and account for the different manners in which a user will speak to a skill, from the fluid to the terse
Interfaces relying on voice must provide the most relevant information that is directly related to the user’s requests, of the right quantity and with the appropriate manner
Playing the part of the assistant with a new user to your application provides information about how your future users will speak to the application, how it should respond, and which functionality to include
People crave variety in their conversations, including with voice applications
FAQ
What is a voice user interface (VUI)?A VUI is the set of words, turns, and speaking patterns that form a conversation between a person and a device to help the user efficiently achieve a goal—think of it as the script for human–computer dialogue.How is a VUI different from a GUI?GUIs are spatial (two dimensions) and let users scan, revisit, and ignore extra on-screen elements. VUIs are temporal (unfold over time), so responses must be just enough for the current moment—there’s no place to “park” extra details the user can visually scan later.How do voice interactions differ from text/chat interactions?Text lets people send dense information, skim, and reply asynchronously. Voice relies on turn-taking, pacing, and spoken quirks (pauses, fillers, mispronunciations). This changes how much to say, how fast to say it, and how to structure each turn.What is the cooperative principle (Grice’s Maxims) and why does it matter for VUIs?It’s a set of conversational assumptions that keep dialogue efficient: Quantity (say enough, not too much), Quality (be truthful and well-founded), Relation (be relevant), and Manner (be clear and orderly). Users expect assistants to follow these, so VUIs should be designed around them.How do I apply the maxim of Quantity in voice design?Give only the information needed for the user’s request—no more, no less. Iteratively trim responses, use context to omit implied details, and avoid long “IVR-style” menus. Aim for concise, complete turns that move the conversation forward.How do I uphold the maxim of Quality in a VUI?Don’t imply capabilities the skill doesn’t support and avoid dark patterns. If certainty is low, hedge honestly. Treat misunderstandings like broken links—detect and correct them rather than pretending success.What does the maxim of Relation mean in practice?Every reply should be relevant to the user’s last turn. Many failures come from misrecognition or poor intent mapping. Shape user replies with well-posed prompts (for example, yes/no questions) and, if necessary, simplify or remove features that cause confusing routing.How does the maxim of Manner affect writing prompts and error messages?Use clear, specific, orderly language suited for speech. Avoid jargon and “written-sounding” sentences. Read prompts aloud and test with the platform’s TTS. In errors, be helpful without guessing wrong; guide the user on what to say or try next.What’s a good process for planning a VUI before coding?Start with the primary intent, brainstorm likely user utterances, sketch dialog flows, then run “Wizard of Oz” sessions where you play the assistant with real users. Capture phrasing for NLU training and refine both prompts and functionality based on what you hear.Why add variety to responses, and how?Repeating the same phrasing feels robotic and fatiguing. Prepare multiple paraphrases per situation and select among them (for example, at random), while keeping the underlying action consistent. Match tone to context—light for entertainment, measured for sensitive tasks.
pro $24.99 per month
access to all Manning books, MEAPs, liveVideos, liveProjects, and audiobooks!
Voice Applications for Alexa and Google Assistant ebook for free
