7.5% ↑ in Confidence Level
Simplifying Complex Decisions through a Multi-agentic AI tool for Manufacturing Leaders
Simplifying Complex Decisions through a Multi-agentic AI tool for Manufacturing Leaders
ROLE
Product Designer
UX Designer
DURATION
6 months
CLIENT
Honda Research
Institute
RESPONSIBILITIES
User Research & Analysis, Futures Video Filming & Editing, User Testing, Design, Product Management
TOOLS
Figma, DScout, Dovetail, AfterEffects, PremierPro, JIRA
PROJECT BRIEF
How might non-technical users reflect their evolving values & ethics in
AI systems beyond the domains of transportation in the next 10-15 years*?
How might non-technical users reflect their evolving values & ethics in AI systems beyond the domains of transportation in the next 10-15 years*?
*In an ambiguous brief as such, I started off by defining given constraints, as highlighted in bold
*In an ambiguous brief as such, I started off by defining given constraints, as highlighted in bold
NAVIGATING AMBUITY
I first prioritized answering who and why — unknowns that shaped our research direction in identifying opportunities in a manufacturing & robotics domain
I first prioritized answering who and why — unknowns that shaped our research direction in identifying opportunities in a manufacturing & robotics domain
Lit Review
Industry 5.0 Shift
Transition to Industry 5.0 pushes for human-centered practices, making it a meaningful and timely area to apply ethical AI .
Interviews
Familiarity with AI
Manufacturing leaders face fewer AI adoption barriers, offering a smoother path for AI deployment.
Interviews
Familiarity with AI
Manufacturing leaders face fewer AI adoption barriers, offering a smoother path for AI deployment.
Interviews
Complexity
Manufacturing’s decision complexity is a real-world proving ground that will push our solution to perform at its best.
Interviews
Complexity
Manufacturing’s decision complexity is a real-world proving ground that will push our solution to perform at its best.
Lit Review
Industry 5.0 Shift
Transition to Industry 5.0 pushes for human-centered practices, making it a meaningful and timely area to apply ethical AI .
Lit Review
Industry 5.0 Shift
Transition to Industry 5.0 pushes for human-centered practices, making it a meaningful and timely area to apply ethical AI .
Pain points
Pain points
Challenges that
manufacturing & robotics leaders face
Challenges that
manufacturing & robotics leaders face
Challenges that
manufacturing & robotics leaders face
Based on interviews and surveys with 400+ participants
Based on interviews and surveys with 400+ participants
Based on interviews and surveys with 400+ participants
Challenge 1
Challenge 1
Leadership lacks frontline worker input on implementation realities
Leadership lacks frontline worker input on implementation realities
Challenge 2
Challenge 2
Critical data is scattered across multiple systems
Critical data is scattered across multiple systems
Challenge 3
Challenge 3
There is loss of institutional knowledge due to generational gap
There is loss of institutional knowledge due to generational gap
REFRAMED PROBLEM STATEMENT
REFRAMED PROBLEM STATEMENT
HMW support manufacturing and robotics leaders
to make informed decisions, so that
critical perspectives like floor workers are considered?
HMW support manufacturing and robotics leaders to make informed decisions,
so that critical perspectives like
floor workers are considered?
HMW support manufacturing and robotics leaders
to make informed decisions, so that
critical perspectives like floor workers are considered?
SOLUTION
SOLUTION
One system, many perspectives.
One system, many perspectives.
One system, many perspectives.
We developed multi-agent AI Decision Support System* with key stakeholders that surfaces multiple stakeholder perspectives (including floor workers) at once
We developed multi-agent AI Decision Support System* with key stakeholders that surfaces multiple stakeholder perspectives (including floor workers) at once
We developed multi-agent AI Decision Support System* with key stakeholders that surfaces multiple stakeholder perspectives (including floor workers) at once
*The collaborative intelligence was inspired by our previous interviews with healthcare professionals where seven to eight medical experts collaboratively shape a patient's care plan.
*The collaborative intelligence was inspired by our previous interviews with healthcare professionals where seven to eight medical experts collaboratively shape a patient's care plan.
*The collaborative intelligence was inspired by our previous interviews with healthcare professionals where seven to eight medical experts collaboratively shape a patient's care plan.
AI that remembers context.
AI that remembers context.
AI that remembers context.
It captures historical contexts through a RAG pipeline* to generate responses that are grounded to the specific context of the organization
It captures historical contexts through a RAG pipeline* to generate responses that are grounded to the specific context of the organization
It captures historical contexts through a RAG pipeline* to generate responses that are grounded to the specific context of the organization
*Retrieval-Augmented Generation (RAG) pipeline provides context-aware responses by retrieving relevant, real-time information from external data like policy documents, past case data and regulations.
*Retrieval-Augmented Generation (RAG) pipeline provides context-aware responses by retrieving relevant, real-time information from external data like policy documents, past case data and regulations.
*Retrieval-Augmented Generation (RAG) pipeline provides context-aware responses by retrieving relevant, real-time information from external data like policy documents, past case data and regulations.
Potential side effects made visible.
Potential side effects made visible.
Potential side effects made visible.
It surfaces potential unintended consequences and highlights how their decisions can impact different people involved
It surfaces potential unintended consequences and highlights how their decisions can impact different people involved
It surfaces potential unintended consequences and highlights how their decisions can impact different people involved
Solutions and Trade-offs.
Solutions and Trade-offs.
Solutions and Trade-offs.
AURA provides users with actionable solutions and trade-offs to mitigate the consequences they choose to address
AURA provides users with actionable solutions and trade-offs to mitigate the consequences they choose to address
AURA provides users with actionable solutions and trade-offs to mitigate the consequences they choose to address
LIMITATION
LIMITATION
However, I learned that buttons oversimplified user's decision-making space
However, I learned that buttons oversimplified user's decision-making space
Using buttons is a clear way to capture user input
Using buttons is a clear way to capture user input
Expressing values and ethics in decision-making cannot be minimized to a "click"
Expressing values and ethics in decision-making cannot be minimized to a "click"
IT'S TIME TO LOOK AHEAD
IT'S TIME TO LOOK AHEAD
So, how did I envision* AURA evolving
to overcome these limitations?
So, how did I envision AURA evolving to overcome these limitations?
*I filmed, directed and edited a demo video using AfterEffects and PremierPro to conceptualize the future human-AI interactions based on the research findings
*I filmed, directed and edited a demo video using AfterEffects and PremierPro to conceptualize the future human-AI interactions based on the research findings
Multi-modal interactions beyond buttons
Multi-modal interactions beyond buttons
Multi-modal interactions beyond buttons
With speech, AURA reads hesitation, tone, and emotion to understand users’ personal stakes, while touch makes exploration feel more fluid and intuitive than clicking through a set of buttons.
With speech, AURA reads hesitation, tone, and emotion to understand users’ personal stakes, while touch makes exploration feel more fluid and intuitive than clicking through a set of buttons.
With speech, AURA reads hesitation, tone, and emotion to understand users’ personal stakes, while touch makes exploration feel more fluid and intuitive than clicking through a set of buttons.
Articulating what matters most with spatial distance.
Articulating what matters most with spatial distance.
Articulating what matters most with spatial distance.
Using spatial distance, users can dial down certain trade-offs so AURA can highlight the ones that are aligned more closely with their values.
Using spatial distance, users can dial down certain trade-offs so AURA can highlight the ones that are aligned more closely with their values.
Using spatial distance, users can dial down certain trade-offs so AURA can highlight the ones that are aligned more closely with their values.
ADDITIONALLY
I also contributed to the storytelling and the creation of conceptual models used to visually simplify complex theories in the presentation ↗, defined the long-term success metric and performance tracking of AURA in the user guide ↗, and usability test findings and insights in the research report ↗.
I also contributed to the storytelling and the creation of conceptual models used to visually simplify complex theories in the presentation ↗, defined the long-term success metric and performance tracking of AURA in the user guide ↗, and usability test findings and insights in the research report ↗.
I also contributed to the storytelling and the creation of conceptual models used to visually simplify complex theories in the presentation ↗, defined the long-term success metric and performance tracking of AURA in the user guide ↗, and usability test findings and insights in the research report ↗.
IMPACT
To measure success, we surveyed manufacturing leaders on their confidence levels before and after using AURA. In unmoderated usability testing, users appreciated how the system "surfaced blind spots [they] hadn’t considered" and gave them "a clearer view" of their decision space and its impacts.
To measure success, we surveyed manufacturing leaders on their confidence levels before and after using AURA. In unmoderated usability testing, users appreciated how the system "surfaced blind spots [they] hadn’t considered" and gave them "a clearer view" of their decision space and its impacts.
+ 7.5%
Confidence Level
TAKEAWAY
Leading this 6-month Capstone Project as a UX Designer, navigating a highly ambiguous space and speculating on the future of human–AI interaction, challenged me in ways that required persistence and growth. I had the privilege of leading and designing alongside an incredibly motivated and passionate team, and I’m deeply grateful for the opportunity to have worked on such a complex and creatively demanding problem.
Hard to pick, but these are the biggest takeaways I had.
Leading this 6-month Capstone Project as a UX Designer, navigating a highly ambiguous space and speculating on the future of human–AI interaction, challenged me in ways that required persistence and growth. I had the privilege of leading and designing alongside an incredibly motivated and passionate team, and I’m deeply grateful for the opportunity to have worked on such a complex and creatively demanding problem.
Hard to pick, but these are the biggest takeaways I had.
Leading this 6-month Capstone Project as a UX Designer, navigating a highly ambiguous space and speculating on the future of human–AI interaction, challenged me in ways that required persistence and growth. I had the privilege of leading and designing alongside an incredibly motivated and passionate team, and I’m deeply grateful for the opportunity to have worked on such a complex and creatively demanding problem.
Hard to pick, but these are the biggest takeaways I had.
Finding a starting point in ambiguity
Within the inherently subjective context of ethics with time constraint, my strategy was to go broad individually, then converge as a team.
We initially started with literature reviews, then shared out our findings to then align on a shared definition of “ethics” as personal principles and moral codes that guide decision-making.
Guiding the team through uncertainty
Gaining conviction through action, rather than overthinking the perfect next step, empowered the team to critically assess and refine the direction of our work, making it easier to pivot when necessary.
Balancing different ways of working within the team
Working on speculative projects under time constraints meant I had to learn how to effectively balance the team's creativity with grounded work. Deliberately planning activities like Crazy 8s and sci-fi movie nights while developing our research plans and survey questions helped the team uncover ambitious unknowns that was outside the box, while following up with grounded discussions helped surface new insights that directly supported our project goals.
Finding a starting point in ambiguity
Within the inherently subjective context of ethics with time constraint, my strategy was to go broad individually, then converge as a team.
We initially started with literature reviews, then shared out our findings to then align on a shared definition of “ethics” as personal principles and moral codes that guide decision-making.
Guiding the team through uncertainty
Gaining conviction through action, rather than overthinking the perfect next step, empowered the team to critically assess and refine the direction of our work, making it easier to pivot when necessary.
Balancing different ways of working within the team
Working on speculative projects under time constraints meant I had to learn how to effectively balance the team's creativity with grounded work. Deliberately planning activities like Crazy 8s and sci-fi movie nights while developing our research plans and survey questions helped the team uncover ambitious unknowns that was outside the box, while following up with grounded discussions helped surface new insights that directly supported our project goals.
Finding a starting point in ambiguity
Within the inherently subjective context of ethics with time constraint, my strategy was to go broad individually, then converge as a team.
We initially started with literature reviews, then shared out our findings to then align on a shared definition of “ethics” as personal principles and moral codes that guide decision-making.
Guiding the team through uncertainty
Gaining conviction through action, rather than overthinking the perfect next step, empowered the team to critically assess and refine the direction of our work, making it easier to pivot when necessary.
Balancing different ways of working within the team
Working on speculative projects under time constraints meant I had to learn how to effectively balance the team's creativity with grounded work. Deliberately planning activities like Crazy 8s and sci-fi movie nights while developing our research plans and survey questions helped the team uncover ambitious unknowns that was outside the box, while following up with grounded discussions helped surface new insights that directly supported our project goals.
IMPACT
To measure success, we surveyed manufacturing leaders on their confidence levels before and after using AURA. In unmoderated usability testing, users appreciated how the system "surfaced blind spots [they] hadn’t considered" and gave them "a clearer view" of their decision space and its impacts.
7.5% +
Confidence Level
Copyright © 2025 Lucy Ji Soo Choi
LIMITATION
However, buttons oversimplified user's decision-making space
Using buttons is a clear way to capture user input
Expressing values and ethics in decision-making cannot be minimized
to a "click"
IT'S TIME TO LOOK AHEAD
So, how did I envision AURA evolving
to overcome these limitations?
Copyright © 2025 Lucy Ji Soo Choi
7.5% ↑ in Confidence Level
Simplifying Complex Decisions through a Multi-agentic AI tool for Manufacturing Leaders
ROLE
UX Designer
DURATION
6 months
CLIENT
Honda Research
Institute
RESPONSIBILITIES
User Research & Analysis, Futures Video Filming & Editing, User Testing, Design, Product Management
TOOLS
Figma, DScout, Dovetail, AfterEffects, PremierPro, JIRA
PROJECT BRIEF
How might non-technical users reflect their evolving values & ethics in AI systems beyond the domains of transportation in the next 10-15 years*?
*In an ambiguous brief as such, I started off by defining given constraints, as highlighted in bold
NAVIGATING AMBUITY
I first prioritized answering who and why — unknowns that shaped our research direction in identifying opportunities in a manufacturing & robotics domain
Lit Review
Industry 5.0 Shift
Transition to Industry 5.0 pushes for human-centered practices, making it a meaningful and timely area to apply ethical AI .
Lit Review
Industry 5.0 Shift
Transition to Industry 5.0 pushes for human-centered practices, making it a meaningful and timely area to apply ethical AI .
Interviews
Familiarity with AI
Manufacturing leaders face fewer AI adoption barriers, offering a smoother path for AI deployment.
Interviews
Familiarity with AI
Manufacturing leaders face fewer AI adoption barriers, offering a smoother path for AI deployment.
Interviews
Complexity
Manufacturing’s decision complexity is a real-world proving ground that will push our solution to perform at its best.
Interviews
Complexity
Manufacturing’s decision complexity is a real-world proving ground that will push our solution to perform at its best.
Copyright © 2025 Lucy Ji Soo Choi