NewsResearchFor: HR and recruiting · Business operations

Indeed: AI users claim 2.1 hours saved a day, down from 2.3, as reviews sour

Indeed and Glassdoor report self-reported AI time savings slipping and 45% of AI mentions in staff reviews positive. HR should measure before the next rollout.

US workers who use AI at work at least once a week say the tools save them 2.1 hours a day, down from 2.3 hours a year earlier, according to [a report chapter published on 18 November 2026](https://hiringlab.indeed.com/2026/11/18/there-is-no-ai-jobpocalypse/) by Indeed Hiring Lab and Glassdoor Economic Research. In the same survey, 56% of employed US respondents did not agree with the statement "AI makes or could make me more efficient".

Glassdoor's reviews point the same way. Of employee reviews that mention AI, 45% framed it positively in 2026, down from 82% in 2019, and 51% now mention AI only in the "cons" section, the report says.

What was counted

The survey figures come from Indeed's Workforce Insights Survey, fielded by YouGov. The US wave ran in May and June 2026 with 10,353 adults and a stated margin of error of plus or minus 1%, according to [the methodology page](https://hiringlab.indeed.com/2026/11/18/indeed-glassdoors-2027-u-s-hiring-and-worklife-trends-report-consolidated-methodology/). The AI figures cover employed respondents only; the page does not say how many that is. Hours saved are the respondent's own estimate. "Did not agree" counts everyone who did not answer "agree" or "strongly agree", so it includes people who were neutral.

The review figures cover current US employees on Glassdoor. A review counts if it uses "AI", "artificial intelligence", "ChatGPT", "GPT" or "LLM"; it is positive when AI appears only under pros and critical when it appears only under cons.

The other numbers

  • Use: 42% of employed respondents use AI at work at least weekly, up from 36% a year earlier.
  • Employers: 31% say their employer has low-to-zero AI adoption, meaning AI is not encouraged or is discouraged; 25% report high adoption.
  • Hiring: the share of US job postings on Indeed that mention AI terms doubled from 3.6% in September 2025 to 7.2% in September 2026. Nearly two-thirds of job titles that include AI are now outside traditional tech occupations.
  • Morale: reviewers critical of AI were 3.5 times as likely to mention burnout, 3.8 times as likely to mention job insecurity and 7.1 times as likely to mention layoffs as reviewers overall. Of them, 37% approve of their chief executive, against 89% of reviewers who are positive about AI.
  • Jobs: the authors find "little evidence of widespread job losses driven solely by AI". US labour productivity has grown 2.1% a year since late 2019, in line with its long-run average, on Bureau of Labor Statistics data.

The people side and the catch

Fear of job loss is not the only complaint. In Glassdoor comments from June 2025 to May 2026, workers also said AI had made the experience worse for them and for customers; one reviewer described a "big push for use of AI tools that have provided little value". The authors say a weak labour market has let employers ignore these concerns "largely without consequence", and that those employers risk losing staff when hiring picks up.

The limits: hours saved are self-reported and no output was measured. The report gives no margin of error for the group of regular users, so the fall from 2.3 to 2.1 hours may be small. Reviews are written by people who chose to post on one site. The link between AI criticism and layoffs is an association: critical reviewers may work at companies that were already cutting, though the authors say the gaps held after accounting for each review's star rating. Nothing is broken out by team or company size, and the data come from Indeed's and Glassdoor's own platforms.

What to do if you run HR or the AI programme

  • AI programme leads: before next year's budget, replace self-reported hours with one measured output per team, such as tickets closed per agent, invoices matched per clerk or days to close the books, taken before and after the tool.
  • HR: ask staff the survey's own question, whether AI makes or could make them more efficient, before the next rollout and three months after it. Nationally, 44% agreed (100% minus the 56% who did not).
  • HR: read your company's Glassdoor reviews for AI under cons, and attach a plain statement about jobs to any AI announcement, since layoffs are what AI-critical reviewers raise most.
Was this useful for your work?

Take this to work

Forward this

US workers who use AI at work weekly say it saves them 2.1 hours a day, down from 2.3 a year ago, and 56% of employed respondents did not agree that AI makes them more efficient, according to an Indeed and Glassdoor report published on 18 November 2026. Any saving in next year's AI business case should rest on an output counted before and after, not on hours staff report.

Questions to ask

  1. Which of our AI savings figures come from staff estimates, and which from an output we counted before and after?
  2. What share of our staff would agree that AI makes them more efficient, and when did we last ask?
  3. Which tools did we push that staff say add little value, and what would it cost to switch them off?
  4. What have we told staff about jobs alongside each AI rollout, and who owns that message?

The arithmetic

employees × share who use AI at work weekly × self-reported hours saved per user per day = claimed hours a day

At the survey's US rates, per 100 employees: 100 × 42% × 2.1 = 88 claimed hours a day in 2026, against 100 × 36% × 2.3 = 83 a year earlier. More users, slightly less claimed by each. The report does not say how many of those hours became output, and gives no cost.

Questions this article answers

How much time do workers say AI saves them?

US workers who use AI at work at least weekly estimate it saves them 2.1 hours a day, down from 2.3 hours a year earlier, according to Indeed's Workforce Insights Survey, fielded by YouGov in May and June 2026. The figure is self-reported; no output was measured.

Are employees positive about AI at work?

Less than they were. Glassdoor says 45% of AI mentions in employee reviews in 2026 are positive, down from 82% in 2019, and 51% of reviews that mention AI do so only under cons.

Is AI causing widespread job losses?

Indeed Hiring Lab and Glassdoor say there is little evidence of widespread job losses driven solely by AI as of November 2026. They report that the share of US job postings on Indeed mentioning AI doubled to 7.2% in the year to September 2026.

How many US workers use AI at work every week?

42% of employed US respondents use AI at work at least weekly, up from 36% a year earlier, and 31% say their employer has low-to-zero AI adoption, according to Indeed's 2026 survey.

Send this to someone who runs a teamLinkedInXBlueskyHNEmail