Summary
As a designer research working on Conversational AI products, I’ve been collaborating with AI Researchers (Computer Scientists) and ML engineers. The cross-pollination of our two research disciplines – i.e., I bring my design thinking, storytelling, qualitative research and thick data analysis lens, while the Computer Scientists bring their logical reasoning, modeling and coding, and big data analysis lens – has resulted in a much smarter and more empathetic AI product, as well as innovations in the Cognitive AI domain. I’ll share three use cases of how we human researchers collaborate with the AI researchers and the lessons learned.
Key Insights
-
•
UX research often stays on the surface of usability, missing deeper impacts related to data and AI models beneath the iceberg.
-
•
Natural language processing models require extensive and specific training data to understand diverse user utterances.
-
•
AI systems struggle to identify causality, often confusing correlation with cause and effect in user queries.
-
•
Users often engage in investigative dialogue before reaching their true question or desired outcome.
-
•
Human researchers can pinpoint why users interact with AI in unexpected ways by analyzing behavior beyond simple metrics like button clicks.
-
•
Collaboration between human researchers and AI scientists through shared observations and joint sense-making enhances AI development.
-
•
AI planning, such as the monkey and banana problem, offers a framework connecting user goals with efficient AI action sequences.
-
•
Human-centered research can inform training data selection, helping AI models to better predict and respond to user needs.
-
•
Effective AI product research requires understanding the technology enough to communicate meaningfully with technical teams.
-
•
Human researchers increase impact by shifting from passive observers to active participants in ideation, design, and strategy discussions.
Notable Quotes
"After more than 10 years of doing UX research, I was still only working above the iceberg, focusing on the usability and feature level."
"Machines can identify correlation via association, but it’s extremely difficult for them to reason and identify causality."
"People don’t always get straight to their question; they might need to do some investigation first before they figure out what to ask."
"The AI planning monkey and banana problem made me question whether getting the bananas is really the end goal."
"By identifying the triggers that lead customers to ask questions, I helped data scientists figure out what training data to explore."
"It’s not enough just sharing customer stories and insights; we need to get into the weeds and collaborate with product and tech partners."
"We need to be genuinely curious not just about our customers, but also about our colleagues and the technology behind AI."
"Our foundational interview, observation, and sense-making skills are evergreen and indispensable in the AI world."
"To be truly impactful, human researchers need to actively participate in product design sprints and strategy meetings."
"This emerging AI technology has given us a golden opportunity to combine the heart and brain of technology and make meaningful impact."
Dig deeper—ask the Rosenbot:
More Videos
"Connecting strings of Christmas lights is a mental model for connecting teams’ insights without superseding their systems."
Etienne FangPower of Insights: Why sharing is better than silos with Uber’s Insights Platform
December 16, 2019
"Our apprentice program is free and intense, combining foundational training with real-world projects over six months."
Victor M. GonzalezPracticing Learners and Learning Practitioners
March 10, 2021
"Storytelling isn’t just for qualitative research; it’s also powerful for framing quantitative data to drive decisions."
Roberta Dombrowski Sam Duong WoloszynskiMaking Research a Team Sport
March 11, 2022
"BlueStar lowered users' A1C by 2 points, outperforming the most common drug metformin which averages 1.5."
Daniel J. RosenbergDigital Medicine Design
September 26, 2019
"Document your design decisions and architecture visually so new team members can quickly understand your system."
Jaime CreixemsBest Practices when Creating and Maintaining a Design System
June 7, 2023
"We are stuck in the user trap because these platforms are central to being human today."
Tricia WangFrom Users to Shapers of AI: The Future of Research
March 25, 2024
"Building relationships beyond roles, as people, makes everything else fall into place."
Shahrzad SamadzadehWhat Is My Value? Two Takes and Some Mistakes
January 8, 2024
"Designers set out to resolve ill-defined problems using abductive thinking — making things that we know are kind of wrong to test solutions."
Jorge ArangoDesign as an Antidote to VUCA
May 9, 2019
"The Rosenverse is a weird mix: it’s a for-profit business but also a neutral community resource."
Louis RosenfeldThe Rosenbot and the Rosenverse: An AMA with Lou Rosenfeld
June 5, 2024
Latest Books All books
Dig deeper with the Rosenbot
How does reframing disability from a medical to a social model improve accessibility adoption?
What lessons did Sean and Sarah share about sustaining knowledge management beyond the closure of a government design team?
How did async feedback loops reduce review times and improve executive alignment at Zapier?