AI Studio: Mission-Ready Assistant Prototype
Project Overview
This project explored how AI assistants can support analysts and mission planners in high-stakes environments. The existing prototype, developed under the AI Studio’s Defense vertical, was functional but lacked a cohesive user experience. My focus was to modernize the interface through a targeted UI facelift that aligned with contemporary assistant patterns such as ChatGPT, Claude, and Copilot.
The initial prototype was built by developers to validate the AI workflow but lacked UX refinement. The interface felt dense and visually uneven, with minimal hierarchy between chat, file uploads, and results. Users struggled to focus on key interactions during demos, making it clear that the design needed a more intuitive structure.
Early developer prototype used for internal demos. Functional but inconsistent in layout and lacking visual hierarchy.
The facelift introduced a clearer visual system inspired by familiar assistant patterns like ChatGPT and Copilot. I streamlined the layout with a left sidebar, modernized the color palette, and improved spacing to guide the user’s attention. The result is a clean, neutral, and scalable interface that aligns with AI Studio’s vision for secure operational tools.
Redesigned interface showing refined layout, scrollable output pane, and simplified document upload flow.
Figma
UX Designer, UX/CX Consultant, Full Stack Engineers, Product Manager
My Role
UI/UX Designer within AI Studio’s Defense vertical
Focused on modernizing visual hierarchy, usability, and interaction consistency across the chat experience.
Worked closely with the product lead and developers to ensure updates fit within the existing MVP scope and technical limits.
Design Objectives
Modernize the interface without introducing new workflows
Create a familiar chat-first layout for faster user orientation
Improve usability for reviewing and handling long mission documents
Maintain a neutral, professional aesthetic suitable for secure environments
Research & Benchmarking
I reviewed leading AI assistants to understand user expectations:
ChatGPT for conversational flow and inline citations, Claude for document-centric layouts, Copilot for enterprise-ready structure, and Perplexity for compact responses with clear source links.
These insights showed users expect assistants to blend chat, document context, and structured references. The original prototype lacked this familiarity, which caused friction in early demos.
Design Improvements
The facelift focused on elevating usability through visual and structural clarity. I refined the layout to mirror modern AI assistants, introduced a sidebar for navigation, and simplified the upload flow to make interactions more natural. The visual system adopted a neutral palette inspired by Material principles, improving contrast and readability without adding brand bias.
Long-form outputs were reorganized into a scrollable reading pane, and placeholders for inline citations introduced familiarity from tools like ChatGPT and Copilot. Subtle adjustments to hover states, section grouping, and navigation space transformed the prototype into a clean, scalable workspace suitable for operational contexts.
Constraints
Visual facelift only; no new features or backend changes
Neutral, non-branded design for external readiness
Quick turnaround to meet development sprint deadlines
Outcome
The facelift turned a functional prototype into a refined, trustworthy AI workspace.
It improved clarity and usability in stakeholder demos, provided a reusable foundation for other AI Studio assistants, and enhanced user trust through familiar patterns and modern layout behavior.
Lessons Learned
Small, focused design refinements can significantly influence user confidence
Familiar interaction patterns from leading AI tools accelerate adoption
Layout and scroll behavior are critical for readability in long-form AI outputs
Working within constraints sharpened focus on polish, clarity, and interaction flow
Next Steps
Prototype collapsible panels and chat states
Extend facelift framework to other vertical assistants
Validate facelift usability with SMEs through short testing sessions