Microsoft AI Science Engine

The AI-powered research platform had poor usability and communication issues in its UI.

Usability Audit 2022 - Failed

Key Problems

  • Overly technical language made interactions unclear.

  • Lack of feedback & inintuitive results were displayed, making it hard for scientists to interpret data.

  • Poor documentation on how AI models worked.

    Impact:

    • Scientists struggled to effectively use AI-driven insights.

    • The platform failed usability audits, affecting adoption.

Process & Solution

🎯 Redesigning AI Interactions

  • Improved UI for key interactions:

    • Search experience: Simplified querying and input prompts.

    • AI feedback visibility: Made system responses more apparent and actionable.

    • Query time transparency: Provided users with better progress updates.

🎨 Design System & Usability Enhancements

  • Refined the Design System for consistency across features.

  • Improved accessibility following Azure Design Systems.

  • Worked closely with developers to ensure pixel-perfect execution.

📖 Error Messaging & Empty States

  • Conducted an audit of error messages to improve clarity.

  • Collaborated with copywriters & illustrators to create empty states.

  • Ensured error handling aligned with user expectations & best practices.

🤝 Collaboration & Workflow Optimization

  • Worked cross-functionally with Scientists, PMs, Developers, and Researchers.

  • Led handoff improvements between designers & engineers to streamline implementation.

  • Mentored designers & supported onboarding efforts.

Impact & Results

✅ Audit Success

  • Before: 80% of user journeys failed usability tests.

  • After: 99% of user journeys passed in the final audit.

📈 Improved Adoption & Efficiency

  • Modernized the UI to align with Microsoft’s standards.

  • Enhanced feedback mechanisms, improving user trust in AI outputs.

  • Strengthened internal communication between designers, scientists, and engineers.

💡 Lessons Learned

  • Handling AI-driven design is complex—balancing transparency and usability is key.

  • Accessibility-first design ensures better adoption across all users.

  • Collaboration across disciplines (AI, science, UX) is crucial for success.