102.3K+Download
Skip to the usage guide
Conversational UXby Owl-Listener

Conversational UX Skill.

Design what the system says, what it expects back, and how it recovers when those two do not match.

  • UX
  • Design Guidance
Download OpenDesignView Usage Guide
Source Review · Conversational UXSKILL.md
Task UX Produces • Dialog flows with meaningful branches • Distinct retry prompts and escalation steps for different errors • Persona and voice/text-specific interaction rules Before You Start Provide supported tasks, channel constraints and confirmation requirements. Voice flows also need device and speech-recognition testing; a written dialog alone cannot establish recognition quality.
Editorial summary · source reviewed 2026-10-07 · not a generated example

01 What It Does

Conversational UX: Design Turns and Recovery

Voice and chat turn design, confirmation strategies, error ladders, persona, and channel-specific constraints.

Source-based workflowNot an execution benchmark

The Working Approach

Conversational UX treats a conversation as a sequence of turns rather than a screen layout. It covers the messages the interface says or displays, user responses, acknowledgements, and the choice between explicit, implicit, or no confirmation. Voice and text share that structure but offer different ways to interact.

The source puts unusual emphasis on unhappy paths. No input, misrecognition, out-of-scope requests, and partial matches require different responses. Recovery should add information rather than repeat the same message, and eventually offer an exit or another channel. Conversation is not automatically better than a form or menu.

Expected Deliverables

  • Dialog flows with meaningful branches
  • Distinct retry prompts and escalation steps for different errors
  • Persona and voice/text-specific interaction rules
See the Example Brief

From Brief to Review

  1. 01
    Choose the Channel

    Identify voice-only, text chat, or a hybrid interface and the task it should support.

  2. 02
    Write the Turns

    Specify prompts, expected response types, acknowledgements, and confirmation based on action risk.

  3. 03
    Design Recovery First

    Handle silence, misunderstanding, partial information, and unsupported requests with increasingly useful help.

  4. 04
    Rehearse the Experience

    Read voice copy aloud and test real utterances, pacing, exit, restart, and escalation paths.

02 When to Use It

When to Use It

Provide supported tasks, channel constraints and confirmation requirements. Voice flows also need device and speech-recognition testing; a written dialog alone cannot establish recognition quality.

A Good Fit

  • Ambiguous User RequestsConversation can support open-ended tasks where people do not know a fixed navigation path.
  • Voice-Primary ProductsUseful when users cannot rely on scanning a visible interface.

Limits to Check

  • Not Every Workflow Should ChatThe source warns against using conversation for tasks better served by forms, tables, or menus.
  • Scripts Do Not Prove RecognitionReal-device speech and actual user utterances must test the designed dialog.

Bring These Inputs

  • Supported TasksList what the assistant can genuinely do and what remains out of scope.
  • Action RiskIdentify actions needing explicit confirmation or human escalation.
  • Channel and ToneState voice, chat, or multimodal constraints and the intended register.

03 Decisions That Matter

Decisions That Matter

  1. Recovery Adds Information

    A reprompt ladder gives users a new route forward instead of looping the same failure.

  2. Voice Is Ephemeral

    Short prompts and limited spoken choices account for the inability to reread audio.

04 Try a Focused Brief

Try a Focused Brief

The example brief below has not been tested. Check the skill’s setup before using it.

Review the Setup

Load this task's SKILL.md from the designer-skills repository in your agent's supported skill location. No universal installer command is established by the entry file.

Upstream Source File
https://github.com/Owl-Listener/designer-skills/blob/9a6930cf84a822eb458624bd11c61aac5bbdf224/interaction-design/skills/conversational-ux/SKILL.md

Provide supported tasks, channel constraints and confirmation requirements. Voice flows also need device and speech-recognition testing; a written dialog alone cannot establish recognition quality.

Start With One Deliverable

Adapt the scope, source material and constraints in this brief before running the skill.

Example Brief
Use conversational-ux to design a voice appointment-rescheduling flow. Handle no input, an unknown date, partial understanding, and an unsupported request. Confirm the final change explicitly, offer a clear exit, and rehearse the script aloud without assuming a screen is visible.

Inspect the Result

Check the actual output against these observable criteria.

  • Failure types receive distinct recovery prompts.
  • High-stakes actions have deliberate confirmation.
  • Users can stop, restart, or reach another channel.

OpenDesign provides a local design workspace for coding agents. This listing does not mean the skill is bundled or that its external tools, credentials and software dependencies are included.

Download OpenDesignRead the Quickstart

05 Source and License

Source and License

This page describes the pinned source revision linked below. Upstream behavior and requirements can change.

Source Owner
Owl-Listener
Repository
Owl-Listener/designer-skills
Skill File
interaction-design/skills/conversational-ux/SKILL.md
License
MIT
Reviewed Snapshot
2026-10-07 · 9a6930cf84a8
Editorial Source Summary
Voice and chat turn design, confirmation strategies, error ladders, persona, and channel-specific constraints.

Original summary by OpenDesign, based on the linked source. Repository stars are a dated popularity signal, not skill usage or a quality score.

06 Questions Before You Start

Questions Before You Start

Is a chatbot just a form with messages?

No. The source treats turn structure and recovery as distinct design concerns and advises using forms when appropriate.

Should the same error prompt repeat?

No. Each recovery attempt should add information, with a graceful exit after repeated failure.

Can voice depend on the screen?

The source says the spoken interaction should make sense without looking at a display.

OpenDesign

Make Room for the Next Design Iteration.

Keep your brief, source material and design work together in a local workspace. Check the skill’s own setup before adding it to your workflow.

Download OpenDesignBrowse Design Skills
Upstream Source File
https://github.com/Owl-Listener/designer-skills/blob/9a6930cf84a822eb458624bd11c61aac5bbdf224/interaction-design/skills/conversational-ux/SKILL.md