Pontus Wärnestål | Compass UX Session 2025
summary
Argues that a Conversational Design System (CDS) adapts visual design systems to dialogue by defining reusable dialogue components, pattern libraries and meta tokens to lock brand constraints; it sits as a filter between user input and optional LLMs to prevent tone drift and reduce hallucination risk. Shows an automotive Seeker example, uses Wizard‑of‑Oz testing and months of iteration to craft and QA apologies, clarifications and confirmations, and exposes style tokens and conversational primitives for instantiating multiple brands and personalization.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:03Pixels to conversational design
- 2:09Interaction design in UX
- 4:19Natural language dialogue challenges
- 6:26Conversations as visual interfaces
- 8:32Tools and guardrails in design
- 9:15Limitations of raw LLM solutions
- 10:58Introducing conversational design system
- 13:05Design system framework
- 16:25Goals and meta tokens
- 19:12Components and patterns
- 21:45Dialogue quality crafting
- 23:10Prototyping with Wizard of Oz
- 24:33Impact on interaction design
- 26:24Q&A
- 36:19Final Q&A and closing remarks
related talks
54mConversational UX—Human Speech Patterns Meet UX
44mNavigating the Hurdles of Conversation Design
43mHi, Want to Chat? When Conversation Is the Interface, the Challenge Is Designing Personalities
10mThe Role of the UX Designer in Creating Conversational Interfaces
03mDesigning for humanlike interaction
32mParty like it's 1970: Conversational interfaces are back (into your Android app)
46mPresenting Your UX Work Using Spontaneous Talks Frameworks
46mDesign Patterns For AI Interfaces