AI Pays Off When Workers Help Redesign the Work: What 40,500 Workers Just Told Leaders

By Grant, CFO, HUMAN HIVE

Does AI pay off better when workers help redesign their own jobs? Yes, by a wide margin: in the Adecco Group's new survey of 40,500 workers in 36 countries, 96% of workers involved in redesigning their work are willing to adapt, against 29% of those left out. The catch for budget owners: AI training participation fell from 77% to 67% in a year, while trained users are about twice as likely to say AI made them more productive.

The 60-second version

  • Workers who help redesign their work are far more willing to adapt to AI: 96% versus 29%.

  • AI training is shrinking even though 93% of trained users say AI made them more productive, against 47% of untrained users.

  • This week, ask one team to map one workflow with you before any new AI tool touches it.

What happened

On November 19, 2026, the Adecco Group published the seventh edition of its annual Global Workforce of the Future report. It surveyed 40,500 workers across 36 countries and 21 industries, online and in local languages, in June and July 2026. The Guardian's Joanna Partridge led with a headline grabber: for the first time in the report's seven years, pay has dropped down the list of reasons people look for a new job.

The AI findings are the part leaders should print out:

  • 74% of workers use AI agents at work, up from 64% in 2025.

  • Where AI is well integrated into workflows, workers save 20% more time per day than where it is not, and they are seven times less likely to feel overwhelmed by the need to use AI tools effectively.

  • 96% of workers involved in redesigning their work are willing to adapt to changing environments, compared with 29% of those not involved.

  • 92% of workers whose leaders explain how AI will change their role are comfortable with AI-related change. Where that explanation is missing, it is 15%.

  • 66% say AI has changed the skills their jobs require, yet AI training participation fell from 77% in 2025 to 67% in 2026.

  • Among AI users, 93% of those who completed training say AI made them more productive, compared with 47% of those who did not.

Adecco CEO Denis Machuel put it in one line: "AI delivers its greatest impact when it happens with people and not to people."

Why it happened: the human side

I read budgets for a living, so here is how this looks from my chair. Somewhere in 2026, a lot of companies made the same trade. They kept paying for the AI licenses and quietly trimmed the training. On a spreadsheet, that looks like savings. In this data, it looks like paying full price for a tool and getting half the return.

The bigger gap is not money, though. It is voice. The report found only 19% of workers say leaders are empathetic to the effects of change, 21% feel their concerns about new technologies are taken seriously, and 39% feel supported by their manager to adapt. That matches what Jobs for the Future found in its survey of 3,020 people: 56% of workers say their employers have not consulted them about how AI tools are used in their work.

People who are told about a change after it lands do what people have always done. They comply on paper and protect themselves in practice. People who helped shape the change own it.

What Mitch has taught on this

At the end of 2025, Mitch recorded a short message looking back on the year, Goodbye 2025 and hello to 2026. His warning to every company was blunt: "You must stop building or implementing AI on top of broken processes and bad data."

He followed it with the line that explains today's numbers: "AI is only as powerful as the skilled human using it along with incredibly streamlined processes and fantastic data beneath it." Skilled human, streamlined process. Training and redesign. The two things the Adecco data says decide whether AI pays.

His first principle from the book F* Your Friction is just as plain: "If a process is broken, repair it or remove it." And the people who know which steps are broken are the ones doing the work every day. That is why they belong in the redesign.

What's driving people crazy this week

Tools landing on teams without the team. In Aftermath's collection of worker stories, a games producer describes a team moved onto an AI agent that was far slower than the tool it replaced, with people quietly trading workarounds to get the job done. The human fix is cheap: before the switch, ask the people who use the old tool every day to run a side-by-side on real work, and let their results decide.

Best practices: 5 steps for this quarter

  1. Redesign before you deploy. Before any AI tool goes to a team, the manager and two or three people who do the work map one workflow on a whiteboard: every step, who touches it, where it stalls. Fix or remove the broken steps first. Then decide where AI fits.

  2. Explain the change role by role. The 92% versus 15% gap comes down to one conversation. Each manager tells each person, in plain words, which parts of their job AI will change, which will not, and what success looks like in 90 days.

  3. Protect the training line. Finance and L&D leaders: if you are cutting anything in 2027, do not cut the training attached to tools you are still paying for. Ask for a training completion number next to every AI license count.

  4. Give concerns a channel and an answer. Run a 15-minute monthly AI check-in per team. Write down every concern, and come back next month with what you did about each one. The 21% figure moves when people see answers.

  5. Let the users pick the test. When you replace a tool, the team that uses it runs a two-week side-by-side on real work and reports back. Adopt it when it wins, not when the contract starts.

What this means for the next 1 to 3 years

Three predictions from our team, grounded in where this data points.

Work redesign becomes a job skill, not a consulting project. The managers who thrive will be the ones who can sit with a team and redraw how work flows between people and AI agents, every quarter, not once. Expect "led a workflow redesign" to show up on promotion cases.

Training gets measured like revenue. A 93% versus 47% productivity gap is too large for boards to ignore. We expect companies to start reporting training completion beside AI spend, the way they report utilization beside headcount.

The best AI rollouts will look slower at the start and win by the end. Teams that spend two weeks redesigning with their people will lose the race to launch and win the race to results. With 74% of workers already using AI agents, access is no longer the edge. How the work is built around them is.

How we evolve

The human skill to build next is change ownership: the habit of shaping a change instead of waiting to be handed one. It matters more as AI improves, because every new tool rewrites part of someone's job, and the people who help write it keep their footing. Mitch says it plainly in the same video: "Humans are going to have to evolve." He closes with the question that starts it: "What do I need to become to get to this next new future?"

Daily practice (5 minutes): at the end of each day, write down one step in your work that wasted time and one idea to fix it. On Friday, bring the best one to your manager. That is redesign, one step at a time.

Try this today

  • Ask at your next team meeting (5 minutes): "Which step in our work would you fix first if you could?" Write every answer down.

  • Check your numbers (10 minutes): pull how many AI licenses your team has and how many people have completed training on them. If the second number is much smaller, you found your first fix.

  • Have the role conversation (15 minutes): pick one person and tell them which parts of their job AI will change, which will not, and what good looks like in 90 days.

Free tool: the 4-question redesign script

Copy this into your next team meeting: "Before we add any AI tool to this workflow, let's answer four questions together. 1. What are the steps, start to finish? 2. Which step wastes the most time or breaks most often? 3. Do we fix it, remove it or keep it? 4. Where would AI actually help, and who checks its work? We decide together, and we try it for two weeks before we lock it in."

FAQ

Does involving employees make AI adoption work better?
The Adecco Group's 2026 survey of 40,500 workers found 96% of workers involved in redesigning their work are willing to adapt to changing environments, compared with 29% of those not involved.

Does AI training still matter in 2026?
Yes. In the same survey, 93% of AI users who completed training said AI made them more productive, against 47% of those who had not, even though training participation fell from 77% in 2025 to 67% in 2026.

What is the fastest way for a leader to reduce resistance to AI?
Explain how AI will change each person's role. Adecco found 92% of workers whose leaders do this are comfortable with AI-related change, versus 15% where that communication is absent.

Where HIVE can help

Tools are bought but people are stuck? Book an AI Friction Snapshot and find where your people are stuck. Planning 2027 learning? Talk with us about training your people will actually use on Monday. Related: who decides where AI-saved time goes.

Get The HIVE Daily Buzz in your inbox.

As the machines get more human, are we?

Next
Next

When the Boss Trusts the Bot: A Third of Managers Now Trust AI Over Their Own Judgment