Accessible Azure AI-powered tool app that optimized everyday workflow

70%

Usability Test

+1mins/day

Saved Time

Accessible AI

Azure AI Foundry

Corporate Sponsor

Microsoft Azure AI

Timeline

9 months

(Dec 2024-Aug 2025)

Team

Microsoft Principal PM

Microsoft Senior UX Researcher

+3 UW MHCI+D UX Designers

(Me) UX Designer + PM

HIGHLIGHTS

1. Overview

2. Problem

3. Final Solution

4. Key Features

5. Impact

PROCESS

6. Research

7. Magic Mouse

8. Activity History

9. AI System Instruction

10. Reflection

Impact

1. Executive Summary

Flow: Accessible Azure AI desktop tool-app that drives Productivity & Adoption

WHAT MICROSOFT NEEDED

When Azure AI Foundry launched in 2024, it offered strong technical capabilities but lacked inclusive, real-world use cases to prove value and drive adoption. Microsoft needed accessibility-driven use cases that showed how Foundry could deliver productivity gains for professionals and strengthen its role in the Microsoft 365 ecosystem.

OUTCOME

I designed Flow, a 0–1 desktop app with three neurodivergent-inspired features (Magic Mouse, Activity History, and Context Bookmark) and a custom AI system instruction. These solutions reduced context switching, eased communication stress, and lowered cognitive load, delivering ~25 minutes saved per user daily and 100% pilot adoption.

OUTCOME

I translated Azure AI Foundry’s abstract capabilities into an accessibility-driven, dual-layer product that saved time, eased daily friction, and proved how accessibility can drive enterprise adoption.

My Contributions

Increased productivity retention by 40%

✅ 100% user-adoption,

✅ Led end-to-end design and prototype for three key AI features

✅ Led as Azure AI Foundry SME

2. Context

The Gap: Proving Azure AI Foundry’s Value

What is Azure AI Foundry

Launched in 2024, Azure AI Foundry is Microsoft’s platform for building and managing generative AI apps and agents across the enterprise ecosystem.

The Challenge

Strong technical foundation, but no inclusive, real-world use cases. Without adoption stories, enterprise uptake slowed and business impact remained unclear.

Why Neurodivergence

Designing for cognitive diversity revealed universal workflow challenges that hurt productivity for all professionals.

Project Goal

Translate Azure AI Foundry’s capabilities to design neurodivergent-friendly features that help professionals in daily work, making AI more accessible and driving adoption.

3. Problem

Workflow Challenges Professionals Struggle With

Research revealed three consistent pain points in professional workflows: constant context switching, communication stress, and difficulty maintaining focus. These issues drain productivity and create friction in daily work, making them the most critical opportunities for AI to solve.

This insight shaped the design of Flow, where I translated Azure AI Foundry’s capabilities into features that directly address these everyday challenges.

↓ Go to Research Process

Context-Switching

Teams ping or meeting pulls focus.

Communication Stress

Teams ping or meeting pulls focus.

Scatted Focus

Teams ping or meeting pulls focus.

text

Learn which key performance indicators matter most when scaling your SaaS product with confidence.

text

Learn which key performance indicators matter most when scaling your SaaS product with confidence.

text

Learn which key performance indicators matter most when scaling your SaaS product with confidence.

4. Research Process

Reframing Neurodivergent Pain-Points into Universal Workflow Challenges

RESEARCH GOALS

What are their current challenges?

What are their current methods?

📝 (n=43) Survey responses for the quantitative data and responses

🧠 (n=5) In-depth Semi-Structured Interviews to dive deeper into behaviors

🎤 (n=3) SME Interviews with Clinician, ADHD Coach, ADHD-productivity tool startup to understand existing solutions

WHAT DO I NEED TO SOLVE?

Research with neurodivergent professionals revealed four recurring challenges: communication overload, fluctuating executive function, difficulty balancing focus, and fragile momentum. When mapped against workplace studies, these translated into universal workflow barriers that needed to be addressed in our proposed features.

📍 See the detailed Research Report

WHAT AZURE AI CAPABILITIES CAN I LEVERADGE TO SOLVE THIS?

In parallel, I led Azure AI sessions to deepen the team’s understanding of the technical environment and mapped out what specific Azure AI Capabilities can be used to solve these pain-points.

📍I created a detailed Azure Cloud Basics Session & Azure AI Foundry deck to guide the team.

5. Design Process

Dual-layer Solution: UI features + Custom AI system instruction

WHY DUAL-LAYER?

In parallel, I led Azure AI sessions to deepen the team’s understanding of the technical environment and mapped out what specific Azure AI Capabilities can be used to solve these pain-points.

UI Features

Reduces Visible Workflow friction

AI System Prompt

Addresses backend invisible communication stress

VISIBLE LAYER

01. UI features

1. Feature Prioritization

I started with 7 potential features and prioritized them against three criteria below. This ensured I selected not just interesting ideas, but three adoption-ready features that balanced user needs with platform capabilities

Solve the Right Problem

Does this feature directly address the biggest workflow barriers — context switching, communication stress, or focus breakdown?

Leverage Azure AI Strengths

Does it map to Foundry’s core capabilities — prompt orchestration, context input, and memory?

Drive Everyday Adoption

Will professionals see immediate value and use it seamlessly inside Microsoft 365?

This down-selection was critical in translating Azure AI Foundry’s abstract capabilities into real workflows professionals would adopt.

Iteration Highlights: Activity History

User testing was critical to move Flow’s features from promising concepts to tools professionals would actually use.

EARLY CONCEPT

Initial log design — linear and manual, requiring users to scroll through actions.

USABILITY INSIGHTS

After conducting five usability testing with users, I have found that the

Problem

3 out of 4 users preferred Copilot chat interaction over manual timeline

Refinement

Integrated Copilot as the default entry point for Activity History

Impact

Reduced friction and made the timeline interaction natural

Problem

Users wanted not just what they did, but why it mattered

Refinement

Added summary output explaining task context

Impact

Boosted clarity and relevance, making Activity History adoption-ready

FINAL VERSION

Refined Activity History transformed from a passive log into an actionable timeline, cutting time-to-resume tasks and reducing cognitive load by X%.

INVISIBLE LAYER

02. AI System Prompt

1. Problem

Shittyz Tone

Drive Adoption

Show Enterprise Scale

Off-the-shelf AI interactions often felt rigid and inconsistent, creating communication stress for neurodivergent professionals. (show the comparison output before v. after)