Principal UX Researcher, AI

Position: Principal UX Researcher, AI Experiences
Job d:

Location:

United States, California, Mountain View Salary: USD $142,800 - $274,800 per year

Employment type:

Full-Time Work site: 4 days / week in-office

Role type:
Individual Contributor Travel:
Less than 25%
Profession:
Design & Creative Discipline: UX Research Company:
Microsoft

OverviewAI is moving from answering questions to doing work. With Copilot Cowork, people hand off multi-step tasks and expect finished work back. This is a new kind of experience. We are still learning who uses it, what they use it for, which capabilities matter most, where it breaks down, and what makes people come back. We are looking for a Principal UX Researcher to lead mixed-methods research on Copilot Cowork.

You will study how people use it, uncover unmet needs and friction points, and turn that understanding into a better experience. You will bring strong quantitative skills in surveys, behavioral data, and experiments, along with the qualitative craft to see needs and friction firsthand. You will also help run agentic evaluations for Cowork, making sure we measure agent quality against competitors on the work people actually delegate.

This is a hands-on individual-contributor role with broad influence across Product, Design, Engineering, and Data Science.

What success looks like in the first year:

Product teams set priorities from an evidence-based picture of Cowork's users, top use cases, and the drivers of retention and drop-off. A sized, prioritized map of friction points and unmet needs is in place, and the top issues show measurable improvement. An experience benchmark tracks Cowork across releases using both survey and behavioral metrics. Agentic evaluations for Cowork run regularly, report competitive win rates, reflect real usage, and inform the quality roadmap.

AI Experiences employees who live within a 50- mile commute of a designated Microsoft Hub in the U.S. are expected to work from the office at least four days per week. This expectation is subject to local law and may vary by jurisdiction.

Responsibilities Understand how Cowork is used. Combine telemetry, surveys, and interviews to map who uses Cowork, which tasks they use it for, how usage changes over time, and what separates people who return from those who try it once and leave. Uncover unmet needs and new opportunities. Lead generative research such as contextual inquiry, diary studies, and in-depth interviews with knowledge workers.

Identify the jobs Cowork could own, the capabilities people expect, and the direction the product should take next. Find and size friction. Identify where people get stuck: discovering Cowork, moving from chat into a task, delegating work, following progress, reviewing outputs, and recovering from mistakes. Quantify how common and costly each friction point is so teams fix the most important problems first.

Assess capabilities against expectations. Measure how well Cowork's capabilities meet what users expect across their real workflows, and identify which capability gaps most limit trust, adoption, and repeat use. Improve the experience. Partner with Design and Product on concepts, prototypes, and usability studies. Establish an experience benchmark built on attitudinal and behavioral metrics, and measure whether shipped changes actually improve outcomes.

Help run agentic evaluations for Cowork. Contribute to interactive side-by-side evaluations in which users complete representative tasks in Cowork and competing agents. Help design task sets, rubrics, and analysis plans. Ground evaluations in real usage, and turn losses into prioritized fixes that feed the quality roadmap. Bring quantitative rigor to every question. Design surveys and experiments with sound sampling and enough statistical power.

Segment users, model drivers of satisfaction and retention, and connect what people say with what they do. Use AI-assisted analysis to scale research where it helps. Lead at principal scope. Set the Cowork research agenda with partners, influence roadmap and quality decisions with senior leaders, and communicate uncertainty and tradeoffs clearly. Stay hands-on while mentoring researchers and raising the team's mixed-methods and evaluation craft.

Qualifications

Required Qualifications:

Doctorate in Human-Computer Interaction, Human Factors Engineering, Computer Science, Technical Communications, Information Science, Information Architecture, User Experience Design, Behavioral Science, Social Sciences, or related field AND 3+ years User…