As Principal UX Designer, I led the navigation and information architecture strategy for unifying 8+ disparate playground experiences across Azure AI Studio, serving 35,000 daily active users and enabling 5,000 daily model deployments. When our PM lead departed mid-project, I stepped up to lead the design v-team, organized cross-functional reviews, and rapidly pivoted for competitive features. We successfully shipped unified navigation at Microsoft Ignite in November 2024, maintaining business operations throughout.
Situation
Azure AI Foundry served 35,000 daily active users creating 5,000+ model deployments per day, but the playground ecosystem was deeply fragmented. Each of the 8+ playground interfaces (Chat, Assistants, Agents, Images, Video, Real-time Audio, o1-Preview, AI Services, Healthcare) was built by different teams with unique UX patterns—users couldn’t transfer knowledge between playgrounds, and the platform felt like disconnected tools rather than a cohesive product. This fragmentation created adoption barriers, customer confusion, and competitive disadvantage.
Task – Establish experience continuity and consistency across unified playground ecosystem
The fundamental challenge wasn’t just navigation—it was establishing continuity and consistency in the experience as users moved between 8+ different playground interfaces. Each playground served a distinct AI capability, but users needed to feel they were working within a cohesive platform, not jumping between disconnected tools.
GOAL 1
Create consistent experience patterns across 8+ playground interfaces to establish platform continuity
GOAL 2
Establish foundational patterns that enable future platform consolidation into one unified playground
GOAL 3
Respond rapidly to competitive AI model releases while maintaining pattern consistency
ROLE
Lead UX Designer
TOOLS
Figma, Figma prototypes, Copilot, PICT framework (Problem-Initiative-Consumer-Test)
TEAM
3 engineer,
1 UX researcher,
1 PM lead,
Content designer
TIMELINE
July 2024-Jan 2025
Action 1 – Audited fragmented landscape & proposed consistent experience patterns (July – August 2024)
Mapped 8+ playgrounds documenting UX patterns, interaction models and page layouts across all. Proposed consistent experience pattern: playground switcher, pinnable navigation, consistent page templates, shared interaction patterns. Presented future vision concepts showing ultimate goal of one unified playground. Designed “phased” approach balancing immediate delivery with long-term foundation.
Audit of playground and playground features, current and planned, across Azure AI Foundry.
Action 2 – Validated design through rigorous usability testing (August 2024)
Study design
- Partnered with researcher
- 5 participants (current OpenAI Studio or AI Studio users)
- 45-minute moderated usability sessions
Research goals
- Identify which navigation approach best supports playground experiences
- Gather feedback on specific features (playground switcher, pinning, parameter visibility)
Key findings
- [P0] Static navigation won: Users strongly preferred Prototype #3—navigation should remain unchanged when entering playgrounds
- [P0] Dropdown menu positive but needs improvement: Users liked playground switcher but needed better discoverability
- [P0] Don’t hide toolbar: Users did NOT want playground parameters hidden—they need controls visible for experimentation
- [P1] Pinning valuable but hidden: Users liked pinning capability but struggled to discover the feature
Action 3 – Established design system & enabled team adoption (Aug – Oct 2024)
Organized Sept 30 management review:
- PM departed Sept 20; I pulled together stakeholders, documented business problems, presented strategy
- Transformed chaotic situation into structured plan with clear ownership
Telling the story about how a user would figure out what the best playground(s) are for building their first chatbot.
Action 4 – Led through transition while shipping competitive features (Sept – Oct 2024)
Organized Sept 30 management review:
- PM departed Sept 20; I pulled together stakeholders, documented business problems, presented strategy
- Transformed chaotic situation into structured plan with clear ownership
Continuous competitive shipping:
- OpenAI/others launched new models every few weeks requiring immediate response
- Shipped o1-Preview playground when OpenAI announced o1-preview model
- Shipped Real-time Audio playground, writing production JavaScript for audio visualization
- Each new playground refined and applied emerging consistency patterns
Vibe-coded audio feedback bars that I handed off directly to dev for implementation.
Designed “All Playgrounds” homepage:
- Solved discovery problem identified in Sept 30 review
- Created landing page showcasing all playgrounds with descriptions
Working with the dev to fix layout bugs – Nov 2024
Action 5 – Shipped at Ignite & coordinated post-launch (Nov 2024)
All primary playgrounds shipped with unified patterns
Result
We successfully established experience continuity across all primary playgrounds (Chat, Assistants, Agents, Images, Video, Real-time Audio) at Microsoft Ignite November 2024, serving 35,000 daily users with 5,000+ daily deployments maintained throughout. The design system proved scalable—multiple teams independently implemented unified patterns, validating the approach through usability testing. These foundational patterns enabled future platform consolidation toward the unified playground vision.
After our PM’s departure in September, I led the design v-team through the hard Ignite deadline, documented critical alignment issues for stakeholder review, and rapidly designed competitive playgrounds (o1-Preview, Real-time Audio) while maintaining consistency standards—demonstrating strategic leadership and design velocity under pressure.
