conversational-ux

Design dialog flows, error recovery, and persona for voice and chat interfaces.

2.5k|358|Updated Mar 7, 2026
One-click install
npx skills add https://github.com/Owl-Listener/designer-skills --skill conversational-ux
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: conversational-ux
Source: https://github.com/Owl-Listener/designer-skills/tree/main/interaction-design/skills/conversational-ux
Command: npx skills add https://github.com/Owl-Listener/designer-skills --skill conversational-ux

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Voice interfaces and chatbots fail most often at the gaps between what the system expects and what users actually say. This Skill provides the design discipline for scripting dialog turns, handling misrecognition, and defining a consistent persona when the interface speaks and listens rather than being tapped.

Core Features & Use Cases

  • Dialog Flow Design: Structure conversation turns with clear system prompts, confirmation strategies (explicit, implicit, or none), and constrained choices at decision points.
  • Error Recovery Ladders: Handle no-input, misrecognition, out-of-scope, and partial-match errors with escalating reprompts that add information and end in graceful exits.
  • Voice and Chat Specifics: Apply voice-first writing rules (short sentences, no visual-only references, lists under three items) and text-chat affordances like quick replies, typing indicators, and structured cards.
  • Use Case: When designing a smart speaker skill for appointment booking, use this Skill to script the confirmation strategy for dates, design the three-step error reprompt ladder, and write prompts that read naturally when spoken aloud.

Quick Start

Use the conversational-ux skill to design the dialog flow and error recovery for my chatbot's appointment booking conversation.

Frequently Asked Questions about conversational-ux

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I design error recovery for a voice interface?

Use a three-step reprompt ladder: first rephrase the prompt with more context, then offer explicit examples or constrained choices, then provide a graceful exit to a human agent or another channel. Never repeat the same error prompt twice; each reprompt must add information.

How do I write prompts for voice assistants and chatbots?

Write short sentences in active voice that state one clear thing and signal the expected response type. Read every prompt aloud before shipping, spell out abbreviations on first use, and avoid lists longer than three items since users cannot re-read audio.

When should I use a chatbot instead of a form?

Conversation excels at ambiguous, open-ended, or multi-step tasks where users do not know the exact path. It fails at tasks with many required fields or complex parallel selections, where a form, table, or menu serves users better.

What confirmation strategy should a voice interface use?

Use explicit confirmation (repeating details back for verification) for high-stakes or easily confused inputs, implicit confirmation for low-stakes recoverable actions, and no confirmation when misrecognition is rare and recovery is easy.

How do I design for multimodal voice and screen devices?

Ensure the spoken prompt makes sense without the screen, since not all users look at it. Let the screen reinforce and disambiguate rather than replace speech, and make every interactive visual element also activatable by voice.