Generative AI is speeding up many user-experience tasks, but the people using it remain divided about its value at work, according to a Nielsen Norman Group survey of 521 UX practitioners. The study found near-universal access to AI tools while documenting significant reservations about output quality, data handling and pressure from management.
More than 99% of respondents said they could access generative AI at work. Access did not always mean unrestricted choice: 59% said they could use only tools approved by their employer, while 12% were free to choose any tool. Two-thirds reported using at least one agentic product, a category that included general agents as well as coding agents and app-building tools.
Adoption varied sharply by workplace. Agentic tools were reported by 85% of respondents at startups and 75% at small or midsized businesses, compared with 42% in government and 36% at nonprofits. The researchers suggested that approval processes, security requirements, cost and technical capacity could help explain those differences, but the survey does not establish a single cause.
Respondents generally described AI as a time saver. On a five-point scale, where three meant no change and five meant a much shorter task, the average rating was 4.2. Usage was strongest for writing or editing code, desk research and analysis of a practitioner's own research data. About nine in ten users of AI for those activities said it reduced the time required.
Tasks involving direct contact with users showed much less adoption. Most respondents said they never used AI to interview users, and just over half never used it to test designs with users. Only 11% reported using AI interviewers, with the same share using synthetic users. AI could accelerate preparation around a session, the report said, without eliminating the time needed to conduct the interaction itself.
The efficiency gains did not produce broad enthusiasm. Output quality was an adoption barrier for 54% of respondents, while privacy or data-security concerns were cited by 50%. Ethical or sustainability concerns were selected by 34%. Only 9% reported facing no major barrier, and most chose more than one.
The results describe a self-reported survey rather than a controlled measure of productivity. They nonetheless show why workplace AI adoption cannot be reduced to access or speed. UX professionals are finding practical uses for the tools, especially in work involving code and data, while still spending time checking results and negotiating rules, risks and expectations inside their organizations.
