Markdown edition

# UX practitioners report faster AI tasks, with limits on what a survey can show

> A Nielsen Norman Group survey of 521 UX practitioners finds faster task ratings alongside concerns about AI output and work. The study did not measure completed-work productivity.

By BIGCHANGE.AI Editorial

Published: 2026-10-24T18:49:38.159Z
Updated: 2026-10-24T18:49:38.159Z
Canonical: https://bigchange.ai/blog/ux-ai-work-survey-task-speed-quality-concerns

![Conceptual charcoal illustration of an unbranded phone prototype with blank layout blocks, resting in a metal cradle on a wooden surface.](https://bigchange.ai/api/media/file/ux-phone-prototype-concept-hero-v2.png)
AI-generated conceptual illustration by BIG CHANGE.

Nielsen Norman Group’s [survey of UX practitioners](https://www.nngroup.com/articles/ai-in-ux-survey-26/), published October 23, found that respondents who used AI generally rated individual tasks as quicker. Their views of AI at work were more guarded. The distinction matters for teams deciding whether faster drafting or analysis means a better product or a lighter workload: the survey asked people about their experience, and did not measure either end-to-end delivery time or output quality.

## The big change

- **What changed:** A survey of 521 working UX practitioners puts reported task speed alongside concerns about reviewing AI output, rising expectations and job security. Its task ratings capture practitioners’ perceptions of time, not a timed comparison of finished work.
- **Why it matters:** Respondents described time spent checking and fixing AI output, and 54% selected output quality as an adoption barrier. The survey does not calculate whether those efforts cancel the reported task savings across a completed project.
- **What to watch:** For teams assessing AI use, the relevant decision is which work they count. Tracking a single AI-assisted step leaves the subsequent checking and correction outside the result; assessing the complete deliverable would answer a different question from this survey.

## Faster tasks, as reported by the people doing them

NN/g asked which tasks respondents performed, how often they used AI for each, and how AI changed the time those tasks took. Most people who used AI for a given task said it shortened that task. Among respondents who performed each activity, frequent AI use was highest for writing or editing code (73%), desk research (69%) and analyzing their own research data (67%). The report says around nine in ten AI users of those activities rated them faster. The denominators change by activity: people rated only tasks they did, and the time question applied only where they used AI at least sometimes.

Use was thinner in work directly involving users. Among respondents who performed each task, 61% said they never used AI to interview users and 51% said they never used it to test designs with users. These figures describe this survey group’s practices. They do not establish that AI could shorten an interview or improve a design decision.

The task-time question also has a narrower reach than a productivity claim. It records a rating of perceived change, without timing the task before and after AI use, checking the completed artifact, or adding the time another colleague might spend reviewing it. NN/g’s authors argue in their conclusion for judging the whole process required to ship acceptable work. That is a useful management question, but this survey did not measure a net gain or loss for that process.

## The comments explain the friction, within a smaller group

Respondents’ average feeling about AI at work was 3.57 on a five-point scale where 3 was neutral. Separately, 226 of the 521 respondents answered an optional final question about how AI was affecting their work. NN/g’s sentiment analysis found some negative sentiment in 69% of those **226 comments**. That is not a finding that 69% of all respondents disliked AI. Comments could also be mixed: 49% expressed some positive sentiment, according to the report.

The researchers identified at least one reason in 149 of the 155 comments expressing negative sentiment. Unreliable output and the work of verifying it appeared in 40 coded comments; 24 mentioned mandated or pressured adoption, 18 increased expectations and workload, and 18 job or career insecurity. Those are counts of comments with a coded reason, not prevalence estimates for UX practitioners. A person could give more than one reason, and fewer than half the respondents chose to comment at all.

That difference between the closed questions and the optional comments helps explain the apparently mixed picture. The task ratings describe reported speed, while the comments describe concerns about output quality, mandates and extra work generated elsewhere. The survey records both kinds of account. It does not establish how often they coincide across UX teams or whether AI caused an overall workload increase.

## Job expectations are views, not a forecast

Asked how AI would change their roles, 42% expected their role to be redefined but remain about as important, and 23% expected it to grow or take on more valuable work. Another 12% expected it to stay largely the same; 18% expected it to shrink or be eliminated. These are expectations about the future, not observed changes in employment. On a separate job-security scale from 1 (very insecure) to 5 (very secure), the 21 product managers in the sample averaged 3.81, compared with 2.99 for designers and 2.94 for researchers. NN/g cautions that its product-manager respondents, recruited through UX-focused channels, may differ from product managers more broadly.

The survey ran from August 24 through September 11, 2026. NN/g retained 521 complete responses from 2,814 starts after excluding incomplete answers, bots and people not currently working. It recruited through its practitioner panel, newsletter, selected invite-only professional Slack communities and LinkedIn. This convenience sample gives a detailed view of these respondents, with 246 designers and 145 researchers among them; it cannot supply a representative estimate for the profession. The report says researchers coded the open-ended reasons in Dovetail. For sentiment, two AI models used the same written rules, a researcher rated a random sample of 50 comments, and a researcher reviewed model disagreements.

The practical reading for a UX lead is therefore specific. The report shows where these practitioners *say* AI helps them move through tasks, and where some say its output or workplace use creates new work. It leaves the decisive operational measure open: the time and quality of a completed piece of UX work after all review and correction.

## Sources & further reading

- [Nielsen Norman Group, “AI in UX: Work Is Faster, But Practitioners Have Mixed Feelings”](https://www.nngroup.com/articles/ai-in-ux-survey-26/) (October 23, 2026). Original survey report with task questions, sentiment and comment analysis, role comparisons and methodology. The speed figures are respondent ratings; the optional comments and convenience recruitment limit broader conclusions.

## Sources

- [AI in UX: Work Is Faster, But Practitioners Have Mixed Feelings — Nielsen Norman Group](https://www.nngroup.com/articles/ai-in-ux-survey-26/) — Original survey report. Documents methods, task ratings, optional-comment coding and role views; self-report and convenience sampling do not establish net productivity or profession-wide prevalence.