Log in or create a free Rosenverse account to watch this video.
Log in Create free account100s of community videos are available to free members. Conference talks are generally available to Gold members.
Designing Conversational Interfaces
Summary
There has never been a better time to design conversational UIs! In this informal session followed by extensive Q&A, Greg Nudelman, the author of 5 UX design books, veteran presenter, and designer of conversational UIs on 5 different projects in a variety of industries discussed key aspects of the conversational bot experience and conversational UI best practices. Participants learned: how to pick the right bot framework, how to use basic building blocks of bot dialog (Intents, Entities, Dialogs) to create conversational UIs, and the key importance of context and channels (Mobile, Wearables, etc.) for shaping the conversational experience. We spent the bulk of time on what is "right around the bend" in the near-future of conversational UIs. From practical tips you can put to work immediately, to inspiration for your next project, participants came away saying "Bot Future so Bright, Gotta Wear Ear-Muffs!"
Key Insights
-
•
Treat bots as tools, not humans—avoid anthropomorphizing while setting clear user expectations.
-
•
Microsoft Bot Framework offers a flexible, standalone platform ideal for bot prototyping and development.
-
•
Intent recognition accuracy can be expressed probabilistically to improve UX and debugging.
-
•
Dialog management often follows a waterfall model asking for missing info, enabling partial inputs.
-
•
Context retention across voice interactions remains a challenge for major assistants like Siri and Google.
-
•
Speech Synthesis Markup Language (SSML) can control speech cadence, tone, and pauses to make bots more natural.
-
•
Fallback to human operators (AI Turk model) ensures reliability when bots fail to handle complex queries.
-
•
Bots should never pretend to misunderstand by asking users to repeat louder, as it frustrates users.
-
•
Mapping common user misnomers or accents to canonical entities improves recognition and user satisfaction.
-
•
Early prototyping—even imperfect bots—offers significant UX learning and a valuable AI on-ramp.
Notable Quotes
"Bots is like a perfect on-ramp for UX for AI."
"You gotta assume your bot is going to fail and design from zero."
"Don't name your bot Bob, because it's not alive."
"If a bot can’t handle the query, it goes to a human—it's like Mechanical Turk for AI."
"We really try to push the idea that you shouldn't pretend to be hard of hearing; just say sorry, I didn’t get that."
"The biggest thing for me is that bots bring our human values to AI and help curb abuses."
"Context retention is tough; Siri and Google Assistant completely lose prior contexts after one action."
"SSML is huge for advancing bot voice; tone and cadence carry more meaning than words alone."
"Just start building a quick little service to get a taste of what this technology can do."
"Hands-free control with bots is critical to avoid dangerous distractions, like driving."
Or choose a question:
More Videos
"Two-thirds of our internal team had experienced a late or lost delivery, reflecting real customer frustration."
Anat Fintzi Rachel MinnicksDelivering at Scale: Making Traction with Resistant Partners
June 9, 2022
"We’re not feeling what others feel, we’re feeling what we think they’re feeling in our own way."
Cassini Nazir Meah LinThe Dangers of Empathy: Toward More Responsible Design Research
March 27, 2023
"Langauge is what shapes us, and extracting keywords helps us create narratives to understand how people think."
Andrea GallagherThe Problem Space
May 16, 2019
"When you use more accessible products internally, you make it easier to hire and retain people with disabilities."
Saara Kamppari-MillerTheme Three Intro
October 4, 2023
"Psychological safety is key—if folks aren't in the right state, the conversation isn't productive."
Joshua GravesWe Need To Talk: Managing Ludicrous Requests at Work (Part 3 of 3)
May 12, 2025
"Inclusive design is about one size fits one, not one size fits all, especially in research methodologies."
Sam ProulxSUS: A System Unusable for Twenty Percent of the Population
September 29, 2021
"You can upload any asset like a Figma file or live URL and get AI persona feedback within about five minutes."
Gillian Salerno-Rebic Mark MicheliFrom Insight to Impact: How JourneySpark Used WEVO Pulse + Pro to Drive a 50% Lift in Ad Engagement
June 11, 2025
"We covered critical design gaps without waiting for headcount allocation or hiring delays."
Alicia MootyDesign Staffing Models
September 30, 2021
"Accessibility ops is like design ops but focused on accessibility, including accessible tools and vendor requirements."
Suzan Bednarz Hilary SunderlandAccessibilityOps for All
January 8, 2024
Latest Books All books
Dig deeper with the Rosenbot
Can AI accurately detect participant behavior discrepancies between what is said and what is done?
How can organizations involve non-researchers in research without sacrificing quality?
How can research repositories score or show the impact of research recommendations to enhance organizational learning?