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 AI Applications
Summary
The way we build applications has changed dramatically with every major platform shift: mainframe to PC, web to mobile, and now AI. Luke Wroblewksi, author of Web Form Design and Mobile First, will chart the evolution of AI product design from behind-the-scenes AI models and chat interfaces to agent management and beyond. He'll show how designing AI applications isn't about where to add chat—it's about using AI to solve UX challenges in onboarding, search, customization, task management, and more.
Key Insights
-
•
Every major technology wave transforms applications fundamentally—from mainframes to AI-powered conversation and agentic systems.
-
•
Early AI adoption was mostly behind-the-scenes, enhancing existing UIs without changing user interaction paradigms.
-
•
The explosion of chat-based AI like ChatGPT introduced direct user-AI interaction but also the challenge of the 'empty text box' with unclear capabilities.
-
•
Suggested questions and answers significantly improve user onboarding by guiding users on what they can do and how.
-
•
Including citations in AI responses builds trust and helps users verify information sources.
-
•
Agentic AI systems that autonomously select and use tools create complex interaction flows, raising new UI design challenges for transparency versus simplicity.
-
•
There are two main user mindsets toward AI autonomy: those who want to see the system’s thinking and those who just want results.
-
•
Separating AI process steps from results in interfaces (e.g., ChatDB) helps manage mental load and make outputs more comprehensible.
-
•
Context awareness—understanding and controlling what AI pays attention to—is essential for effective AI interactions.
-
•
Software developers lead AI agent adoption, influencing the trajectory and sophistication of AI products across industries.
Notable Quotes
"In the AI era, applications are conversations. AI models understand intent and execute complex tasks without programming for every scenario."
"When there’s a major technology shift, almost everything about an application changes, including how users experience it."
"Google Translate completely reinvented itself by replacing many translation systems with a single neural network capable of zero-shot translations."
"Suggested questions give you a sense of what the heck can I do here, which is really important in empty text box AI interfaces."
"Showing citations in AI responses helps people develop trust and understand where answers come from."
"Agentic systems don’t just respond; they decide which tools to use, analyze results, think between tool calls, and can do multiple things simultaneously."
"People fall into two groups: some want to see what the system is thinking and why, others just want to see the final output."
"Separating process from results in the UI lets you see what the AI is thinking while giving more prominence and clarity to the results."
"Context management lets you tell the system what to look at, whether it’s files, folders, or highlighting specific image regions."
"The evolution in AI UIs is ongoing and rapid, and we’re still figuring out what the final form of AI-native applications looks like."
Or choose a question:
More Videos
"People need to trust governments before they can trust digital governments."
Yalenka Mariën Marie MervaillieDesigning for Digital Inclusion in the Belgian Government
December 8, 2021
"At any given year, 10 to 20% of veterans will suffer from PTSD, compared to only 6% of civilians."
Megan Nipe Lyndsay BoothHuman-Centered Design for Engagement: Maturing from Newsletterville to Personalized, One-to-One Messaging
December 8, 2021
"At the current rate, it will be 130 years before we reach global gender equality in political power."
Dr. Jamika D. Burge Mansi GuptaAdvancing the Inclusion of Womxn in Research Practices
September 15, 2022
"Skin folk aren’t necessarily kinfolk — diversity doesn’t magically prevent inequitable research."
Jemma Ahmed Robert Fabricant Sean McKay Llewyn Paine Kate Towsey Noah BondTheme Panel
March 11, 2025
"We've worked with speakers for months to prepare our conference content, so it's expensive to produce, but really good."
Louis RosenfeldAbout the Rosenverse
March 15, 2026
"I have a tagline my team is tired of hearing: about 50% of work is getting it done and 50% is getting it used."
Nalini P. KotamrajuTwo Jobs in One: Being a “Leader who is a Researcher” and a “Researcher who is a Leader"
March 10, 2021
"An everything chat bot is almost certainly not how you’re going to build it; realistically you’re building three apps in a trench coat."
Sarah BarrettAI in Real Life: Using LLMs to Turbocharge Microsoft Learn
February 13, 2025
"The future of research isn't more research, it's research that matters more to organizations."
Jane DavisStrategic Shifts and Innovations in User Research: Navigating Challenges and Opportunities
March 11, 2025
"Using the umbrella term 'people of color' erases the most disadvantaged and privileges the least burdened."
Alba VillamilStereotyped by Design: Pitfalls in Cross-Cultural User Research
March 30, 2020
