Principal Designer, AI Experience

Overview

In this role you will shape AI-enabled agent experiences within the AI Transformation team to improve how Microsoft Frontier Company operates. You’ll study current work patterns, prototype with agents, and ship designs that balance autonomy with human oversight. You’ll design for supervision, uncertainty handling, and escalation, ensuring safe, accountable AI-driven workflows. This is a cross-functional, high-ambiguity environment that values practical impact and scalable, systemic design. You’ll contribute to building AI-native operating models and reusable solutions at scale.

Responsibilities

• Understand how people work today and design better workflows with agents and AI capabilities
• Define how agents announce themselves, handle uncertainty, escalation, and human review
• Design supervision and oversight for multiple agents including status, dependencies, and intervention points
• Address failure modes with override, rollback, correction, and recovery when systems malfunction
• Set and prove quality standards for experiences (task success, recoverability, improvement)
• Design for estate growth and governance across catalogue, marketplace, and measurement surfaces

Key requirements

• Master’s or Bachelor’s degree in the specified design fields plus significant product or service design experience or equivalent
• Experience in product or service design with depth in agentic or AI-enabled systems
• Hands-on experience designing systems that act on user’s behalf with appropriate autonomy decisions
• Ability to select appropriate research methods (qualitative/quantitative) and apply responsibly to questions
• Prototyping beyond static screens and conducting user evaluation
• Strong cross-disciplinary collaboration with design, engineering, data science and product teams

• cross-disciplinary collaboration
• influence and alignment across teams
• ability to handle ambiguity
• agentic or AI-enabled systems design
• systems-level craft beyond individual screens
• prototyping and user evaluation of AI experiences

…