AI speeds up UX work, but practitioners remain divided

Summary

NN/g surveyed 521 user experience (UX) experts and found that access to AI is nearly universal (99%) and saves time on most tasks. However, overall satisfaction is only mildly positive (3.57 out of 5), with 69% of open-ended comments being negative. This negativity is attributed to unreliable outputs, top-down imposed policies, and privacy concerns.

Why it matters

This is one of the most comprehensive structured surveys to examine how AI is transforming UX workflows, independent of vendor marketing claims. The sharp divide between 'access' and 'enthusiasm,' along with the drop in agentic AI adoption from 85% in enterprises to 36% in non-profits, reveals that the industry's AI-UX narrative is far more fragmented than assumed. This directly impacts teams' decisions regarding budgeting, training, and tool management.

Highlights

• Output quality (54%) and data privacy (50%) are the two primary barriers to adoption. Only 9% of experts reported no barriers at all.
• AI is used most heavily in back-office tasks (coding, desk research) and least in roles with direct user contact. This suggests AI supports back-office efficiency more than core UX expertise.
• Project managers and operations roles are the heaviest users of agentic AI and feel the most secure in their jobs. Researchers and designers, however, feel the most displaced.

Critical take

The 'mildly positive' average of 3.57 out of 5 persists alongside 69% negative open-ended comments. The article fails to fully resolve this tension. This discrepancy may reflect self-selection bias (dissatisfied experts are more likely to comment) or a survey instrument that nudges participants toward positivity. Additionally, NN/g sells UX training and consulting. Framing AI as 'helpful but flawed' keeps experts in a continuous loop of skill development while eliminating the possibility of questioning the maturity of the tools themselves.