Giving a team AI tools is the easy part. Whether those tools save anyone real time depends on something much less exciting: whether the people who do the work had a hand in changing how it is done.
That is the clearest finding in the Adecco Group's Global Workforce of the Future report, published on 19 November. The staffing group surveyed 40,500 workers across 36 countries and 21 industries in June and July 2026, and the numbers point the same way again and again. AI pays off when it is built into the work and the workers are part of the rebuild. Dropped on top of old habits, it mostly adds to the pile.
What the survey found
Start with how common AI has become. According to Adecco, 74% of workers now use AI agents at work, up from 64% in 2025. Use is no longer the interesting question. Value is.
On value, the gap between teams is wide. Workers who say AI is well integrated into their existing workflows report saving 20% more time per day than workers who describe weaker integration. They are also more likely to put that time into higher-value work, and seven times less likely to feel overwhelmed by the need to use AI tools well. The report notes that while 70% of workers say AI has changed their work activities, 62% say it is well integrated, so a sizeable group has had its work changed without the change being finished.
The sharpest numbers are about involvement. Among workers whose employers involve them in redesigning work in response to AI, 96% say they are willing to adapt to changing environments, against 29% of workers who are not involved. Communication shows the same split: 92% of workers whose leaders explain how AI will change their role are comfortable with the change, against 15% where nobody explains it.
Training is moving the wrong way. Two-thirds of workers (66%) say AI has changed the skills their job needs, yet participation in AI training fell from 77% in 2025 to 67% in 2026. Among AI users, 93% of those who completed training say AI has made them more productive, against 47% of those who did not.
And the report finds a leadership gap. Only 19% of workers say their leaders are empathetic to the effects of change, 21% feel their concerns about new technology are taken seriously, and 39% feel their manager supports them to adapt. On psychological safety, 49% feel safe to speak up, 47% feel safe to experiment, and only 23% see mistakes treated as chances to learn.
One more result sits slightly apart from AI but matters to anyone hiring. For the first time in the series, work-life balance is the top reason people would consider leaving their job. Salary has dropped to fourth, behind flexible schedules and remote work.
How much weight the numbers can carry
I'd read this report as a strong signal, not a proof, and a few things are worth keeping in mind.
It is a survey of what workers say, collected online and self-administered over about 25 minutes. "Saves 20% more time per day" is the respondents' own estimate, compared between two groups, so it tells you that people with well-integrated AI feel they save more time. It does not measure output.
The comparisons are correlations. Workers who are involved in redesign are far more willing to adapt, but the arrow could run both ways: companies that involve their staff may also be better run in many other respects, and people who already like change may be the ones who get invited into redesign. The survey cannot separate those.
The source has a point of view. Adecco sells staffing, consulting and workforce services, and "AI works best with people involved" fits its business well. That does not make the findings wrong, and the sample is large (quotas were set to be nationally representative by age and gender in each country), but it is a reason to look for independent support.
The bigger picture from other data
That support exists, from a different angle. Talent-assessment firm SHL found that 19.1% of American respondents said they had been replaced by AI and later asked to come back, a pattern it calls "AI rebounding". A Robert Half survey from July, also reported by Fortune, found that 32% of US hiring managers whose organisations cut roles because of AI later had to reinstate the same or similar roles, most often because the work needed institutional knowledge, relationships or more human oversight than expected.
Put next to Adecco's numbers, the story is consistent. Removing people without understanding the work tends to fail, and adding tools without changing the work tends to disappoint. The teams that gain are the ones that redraw the job with the people who know it.
What to do with it
If you run a team, or a small business with a few people using AI, these are the steps I'd take from this report.
Pick one workflow and redesign it with the people who do it. Sit down with them, map the steps, and decide together which steps the AI does, which a person checks, and which stay fully human. Write it down.
Say plainly what changes for each role. The gap between 92% and 15% comfort is about explanation, which costs almost nothing. Tell people what their job looks like after the change, and what it no longer includes.
Train on the actual work, not the tool. A short session on "how we now draft client proposals with AI" will do more than a general course, and the survey suggests trained users are about twice as likely to say AI made them more productive.
Make it safe to report what does not work. With only 23% seeing mistakes treated as learning, most problems with a new AI workflow will go unsaid. Ask for them directly in the first few weeks.
Watch the pressure, not only the output. The report found 39% of high performers report more stress, against 17% of lower performers. If AI raises expectations faster than it removes work, people leave, and work-life balance is now the first reason they give.
If you build AI products, the lesson is the same seen from the other side. Your product's value depends on how well it fits into the customer's workflow, so onboarding that helps a team redesign one real process will likely matter more than another feature.
I think the most useful number here is the smallest one: the 29% of workers who are not involved in redesign but are still willing to adapt. That is the default most companies are working with, and the cheapest way to move it is to ask people how their own work should change.

