Computer Using Agents
Computer Using Agents
Computer Using Agents
Designing the maker experience behind intelligent agents.
Designing the maker experience behind intelligent agents.
Designing the maker experience behind intelligent agents.
Over a 12-week internship at Microsoft’s Business & Industry Cloud Design + Research team, I worked on the 'maker' experience for Computer Using Agents... a future-vision initiative exploring how people create, configure, and reason about intelligent systems in enterprise environments. My work focused on designing frameworks, visual models, and storytelling artifacts that made this emerging space tangible for designers, researchers, and engineering partners.
Over a 12-week internship at Microsoft’s Business & Industry Cloud Design + Research team, I worked on the 'maker' experience for Computer Using Agents... a future-vision initiative exploring how people create, configure, and reason about intelligent systems in enterprise environments. My work focused on designing frameworks, visual models, and storytelling artifacts that made this emerging space tangible for designers, researchers, and engineering partners.
Client
Microsoft Design + Research (BIC)
Services
Product Design, Design Strategy, Future Vision, Storytelling
Industries
Enterprise AI
Date
May – Aug 2025
Due to confidentiality, contents in this project are abstracted or blurred.
How do we make AI agents more understandable & human alligned?
How do we make AI agents more understandable & human alligned?
Computer Using Agents (CUAs) are intelligent systems that can act on your behalf autonomously... operating apps, reading context, and automating tasks. But as they become more capable, they also become harder to understand, trust, and control. My work focused on exploring what it means to design for the makers, the people who shape, teach, and refine these systems. The goal was to reimagine how creation and collaboration with intelligent agents could feel more transparent and intuitive.
Computer Using Agents (CUAs) are intelligent systems that can act on your behalf autonomously... operating apps, reading context, and automating tasks. But as they become more capable, they also become harder to understand, trust, and control. My work focused on exploring what it means to design for the makers, the people who shape, teach, and refine these systems. The goal was to reimagine how creation and collaboration with intelligent agents could feel more transparent and intuitive.
My Goals?
My Goals?
Clarify the Maker’s Mental Model
Clarify the Maker’s Mental Model
Map out the journey of people building and testing CUAs. From crafting prompts, observing behaviors, and to identify where breakdowns in understanding occur.
Map out the journey of people building and testing CUAs. From crafting prompts, observing behaviors, and to identify where breakdowns in understanding occur.
Design for Collaboration and Trust
Design for Collaboration and Trust
Explore how CUAs could communicate their reasoning and adapt through guided interaction, helping makers build confidence through transparency.
Explore how CUAs could communicate their reasoning and adapt through guided interaction, helping makers build confidence through transparency.
Visualize the Future Maker Experience
Visualize the Future Maker Experience
Create speculative frameworks and storytelling artifacts that show how designing, debugging, and teaching agents might look in the near future.
Create speculative frameworks and storytelling artifacts that show how designing, debugging, and teaching agents might look in the near future.
What did I learn from this?
What did I learn from this?
Enough to fill a few NDA pages and a few FigJam boards
Enough to fill a few NDA pages and a few FigJam boards
Let’s just say this project changed the way I think about designing for people who build with AI... from how they trust their tools to how they teach them. That’s for you to find out when we talk 🔒
Let’s just say this project changed the way I think about designing for people who build with AI... from how they trust their tools to how they teach them. That’s for you to find out when we talk 🔒
The early phase focused on understanding what makes agents feel trustworthy. I studied existing AI workflows, developer experiences, and maker tools to uncover patterns of friction. Then, building on these insights, I created concept models for how makers might compose, inspect, and adjust CUAs. The goal wasn’t to design the final UI, but to frame how makers could actually make use of CUAs.
The early phase focused on understanding what makes agents feel trustworthy. I studied existing AI workflows, developer experiences, and maker tools to uncover patterns of friction. Then, building on these insights, I created concept models for how makers might compose, inspect, and adjust CUAs. The goal wasn’t to design the final UI, but to frame how makers could actually make use of CUAs.
So... Whats my impact?
So... Whats my impact?
During this project, I helped define early maker experience principles that guided CUA exploration and informed how the team began designing tools for the people building intelligent systems. My work also contributed to internal discussions around trust, collaboration, and explainability, bridging conversations between design, research, and engineering.
I presented these findings at the Intern Design Showcase, where the project was recognized for its clarity and storytelling. I believe the concept frameworks I developed continue to influence how the team at Microsoft envisions the relationship between humans and AI agents… not as tools, but as creative collaborators.
During this project, I helped define early maker experience principles that guided CUA exploration and informed how the team began designing tools for the people building intelligent systems. My work also contributed to internal discussions around trust, collaboration, and explainability, bridging conversations between design, research, and engineering.
I presented these findings at the Intern Design Showcase, where the project was recognized for its clarity and storytelling. I believe the concept frameworks I developed continue to influence how the team at Microsoft envisions the relationship between humans and AI agents… not as tools, but as creative collaborators.
During this project, I helped define early maker experience principles that guided CUA exploration and informed how the team began designing tools for the people building intelligent systems. My work also contributed to internal discussions around trust, collaboration, and explainability, bridging conversations between design, research, and engineering.
I presented these findings at the Intern Design Showcase, where the project was recognized for its clarity and storytelling. I believe the concept frameworks I developed continue to influence how the team at Microsoft envisions the relationship between humans and AI agents… not as tools, but as creative collaborators.
My final work connected design strategy with storytelling. Making the 'maker' experience as a crucial part of Microsoft’s AI future. The work emphasized that designing for understanding and agency is as important as designing intelligence itself. In the end, it wasn’t just about what agents could do... but what people could create with them.
My final work connected design strategy with storytelling. Making the 'maker' experience as a crucial part of Microsoft’s AI future. The work emphasized that designing for understanding and agency is as important as designing intelligence itself. In the end, it wasn’t just about what agents could do... but what people could create with them.