Summary
Join Katie Johnson, Head of Consumer Insights at PanasonicWELL, to explore the future of developing products and services atop Generative AI technology. This talk will cover how to think about experimenting with users today for use cases that don’t yet exist at scale, and how to bring insightful findings to leadership to make strategic decisions quickly. Katie will shed light on methods and strategies she’s employed in building products with cutting-edge technology throughout her career in agency life, blockchain, Google’s 0 to 1 environment in Assistant, and now at PanasonicWELL.
Key Insights
-
•
Building AI products requires giving up control over exact outputs since LLMs behave as black boxes even for their creators.
-
•
Traditional product orthodoxies like fixed user experiences and small-sample usability tests need re-evaluation for AI-driven products.
-
•
User interactions with LLM-based products diverge uniquely, creating personalized 'relationships' rather than uniform experiences.
-
•
Longitudinal and repeated-user studies are critical to understand AI product dynamics over time rather than one-off usability sessions.
-
•
Wizard of Oz testing, where humans simulate AI bots, enables early learning about conversation tone, latency, and asynchronous interaction challenges.
-
•
Model drift causes AI chatbots to unpredictably change behavior, requiring interventions like conversation turn capping and memory features.
-
•
Users often lack mental models for delegating tasks to AI assistants, demanding product designs that accommodate multiple simultaneous learnings.
-
•
Failures in AI product development are inevitable and must be embraced early and systematically to avoid costly large-scale issues later.
-
•
The stochastic nature of AI output means users may never reach full mastery of the product in the traditional sense.
-
•
AI products should augment uniquely human abilities—creativity, accountability, authenticity—by automating routine synthesis.
Notable Quotes
"Building on AI means giving up control."
"If the folks who design this tech don’t know how it works, it’s okay if you don’t know either."
"Orthodoxies all have expiration dates, especially when new technology upends the way we work."
"With LLMs, the product is no longer fixed; different users have fundamentally different experiences."
"Five to eight users interacting with an LLM will have five to eight completely different experiences."
"Longitudinal studies are essential because relationships evolve over time."
"You learn a lot even from Wizard of Oz testing—humans playing the bot."
"Model drift is hugely problematic; after enough turns the chatbot can start misbehaving."
"Users building mental models for AI assistants are often learning to delegate for the first time."
"Creativity, accountability and authenticity will be the new markers of humanity in this AI age."
Dig deeper—ask the Rosenbot:















More Videos

"Benchmarking can be used repeatedly to track whether optimizations have the desired effect over time."
Daniela Magaña Flores Ariane Rahn Jeff Ephraim BanderAhead of Competition: Learn What UX Benchmarking Can Do for Your Business Today
March 10, 2022

"Prompt chaining is one of the most underrated techniques in prompt engineering, turning chains of prompts into designed information processing workflows."
Savina HawkinsHarnessing AI in UXR: Practical Strategies for Positive Impact
March 26, 2024

"Innovation in conservation practices is desperately needed to keep pace with environmental changes."
Alex Hurworth Bonnie John Fahd Arshad Antoine MarinDesigning a Contact Tracing App for Universal Access
October 23, 2020

"Building a custom repository is resource intensive, but offers control necessary for large organizations like Microsoft."
Matt DuignanHITS, Microsoft's internal human insight system: From research library to living body of knowledge (Videoconference)
July 16, 2019

"Make it easy to do the right thing, and as a side effect, make it harder to do the wrong thing - guiding principle from Salesforce Lightning Design System."
Craig VillamorDesign Systems for Ethical Design (Videoconference)
January 26, 2023

"The goal isn't to get people to understand design or research; the goal is to get human-centered design into products that ship."
Ovetta SampsonTurning UX Passion into Real Product Influence
June 7, 2023

"Numbers drive conversations, numbers drive decisions. That’s the executive language."
Rima Campbell Amrit S BhachuIncrease Productivity and Drive Business Impact
September 24, 2024

"Bring blockers into the process so they see firsthand the value of what you’re trying to do."
Amy Gawronski ZuccaroAdvice for DesignOps Employee #1
September 29, 2021

"Whiteness isn’t just the color or race; people of all colors can reinforce white supremacy."
Victor UdoewaBeyond Methods and Diversity: The Roots of Inclusion
March 26, 2024