In November 2025, the McKinsey Global Institute estimated that technologies already demonstrated could, in theory, automate activities accounting for 57% of hours worked in the United States, 13% of them through robots. Robots and jobs: should you be worried? The short answer: robots first take over repetitive physical tasks in predictable environments; judgement, relationships, adapting to the unexpected and now training machines remain human. For a business leader, the question is not whether it will happen, but which roles are getting ready for it.
Key takeaways
- According to McKinsey, robots could technically take on 13% of hours worked in the United States. That is a potential, not a forecast of job losses.
- According to the World Economic Forum, 58% of employers expect robots and autonomous systems to transform their business by 2030.
- Manual dexterity, endurance and precision are the skills whose demand is falling most: 24% of employers expect their importance to decline.
- In the field, the most versatile robots are still far from there: the humanoid pioneers have not yet found profitability.
What is happening
What the studies measure. The McKinsey report, published on 25 November 2025, distinguishes two families of machines: AI agents, which could take on tasks accounting for 44% of hours worked in the United States, and robots, 13%. The authors insist: this figure describes what the technology allows, not what will happen, and adoption will take time. They also stress that tasks relying on social and emotional skills remain the hardest to automate, and that people will continue to supervise, validate and guide the machines.
The World Economic Forum's Future of Jobs Report 2025, based on more than 1,000 large employers, ranks robots and autonomous systems second among the technologies that will transform businesses by 2030, with 58% of respondents, behind AI (86%). Manual dexterity, endurance and precision appear as the skills in clear decline. The report also links automation to demographics: employers affected by population ageing are more likely to want to accelerate the automation of their processes (79% versus 73% on average).
What the field shows. The International Federation of Robotics counts 5 million industrial robots in operation worldwide, but points out that most humanoids remain prototypes or pilots. In August 2026, the director of a Chinese robot training centre estimated that fewer than 5% of the 23,000 humanoids shipped in the first half of the year were capable of useful factory work, for lack of quality training data. And the chief executive of the logistics group GXO, which has been using humanoids since late 2023, acknowledged the same year that it was not yet getting a return on investment from them.
Robotics researchers know this gap well: what is easy for us, grasping an unknown object, walking through a cluttered workshop, feeling that a screw is binding, is what is hardest for a machine. That is why robots advance quickly where the world is tidy, and slowly where it is not.
Robots and jobs: what it means for you
In a workshop or a small business, think in terms of tasks, not jobs. The first tasks affected are those repeated identically in a predictable space: tending a machine, palletising, sorting, screwdriving, visual inspection. Then, more slowly, come the variable movements: a different part every time, a changing environment, a customer on site.
What remains human is not only "people skills". It is also diagnosing a fault, fine adjustment, finishing, adapting to the special case, deciding when something is wrong. And there is a new role: demonstrating, supervising and correcting machines. I devote the article on humans as robot trainers to it.
Context matters too. In many manual trades, the problem is not too many staff but the difficulty of hiring and the wave of retirements. Robots can then relieve arduous jobs rather than replace people, provided the transition is organised. For the AI side of the question, read AI and jobs: prepare your team rather than replace it.
What I have learned
In my workshops in Switzerland and Mexico, laser cutting, CNC and 3D printing have long done what used to be done by hand: cutting, drilling, shaping. They did not eliminate the work: they moved it towards design, file preparation, setup, assembly and finishing. On a watch model, the machine mills the part; it is a person who decides whether it is right.
As a Master Coach ICI (Institut de Coaching International), I also support business leaders in decisions of this kind. My conviction: a team accepts a machine when it sees what it gains from it: less arduous work, new skills, a role that evolves. It fights it when the machine is imposed as a threat.
What to do in practice
1. Sort your teams' physical tasks
For each role, distinguish three families: repetitive movements in a stable setting, variable movements, and tasks involving relationships or judgement. The first family is where a robot can already help.
2. Start with the arduous tasks
Pick heavy, repetitive or dangerous tasks first. That is where the gain is most visible to the team, and where acceptance is easiest.
3. Prepare the new roles
Running a cell, first-level maintenance, simple programming, demonstrating movements: identify who in the team wants to learn these jobs, and train them before the machine arrives.
4. Save the know-how before people leave
Your veterans' skill is the best training data there is. Document it now, whether it later serves to train a person or a machine.
Under the hood
For your technician, the criteria that make a task easy or hard to robotise:
- Variability of parts and their position: an identical, well-placed part is simple; a bin of mixed parts requires vision and much more tuning.
- Tolerances, forces and soft materials: delicate contacts, cables and fabrics remain difficult.
- Environment: a dedicated workstation with constant lighting is better than a shared, changing space.
- Maturity of the building blocks: according to the IFR, machine vision for detection and guidance is mature; general-purpose physical AI is not yet.
Frequently asked questions
Will robots eliminate manual jobs?
They will transform many of them, starting with repetitive tasks. The studies cited describe a technical potential, not a timetable of job cuts, and the pace depends on costs and organisation.
Which trades are least exposed?
Based on the criteria above, those that combine variable movements, an unpredictable environment and a relationship with people: for example maintenance, custom craftsmanship, care work or advisory roles.
Should you wait for humanoids before preparing?
No. Cobots and industrial arms are available today, and the skills to develop are the same: organising tasks, supervising, maintaining and demonstrating.
Going further
This article is part of a four-part series: where humanoid robots and cobots really stand, humans as robot trainers, and where a small business should start. If you want to prepare your organisation and your teams for automation, I can support you: see Strategy & advisory. The simplest way to discuss it: a 20-minute call.
Sources
- Agents, robots, and us: Skill partnerships in the age of AI, McKinsey Global Institute, 25 November 2025.
- McKinsey warns AI and robots could automate 40 percent of US jobs by 2030, Robotics & Automation News, 26 November 2025.
- Future of Jobs Report 2025, World Economic Forum, January 2025 (PDF).
- Executive Summary World Robotics 2026 – Industrial Robots, International Federation of Robotics (IFR), September 2026 (PDF).
- China's embodied AI boom fuels demand for human-collected training data, The Standard, 19 August 2026.
- Agility and GXO Logistics plan expanded deployment of Digit humanoid, The Robot Report, 26 October 2026.
- Photo: Pavel Danilyuk / Pexels, Pexels licence (pexels.com).
Tags: robots and jobs, future of work, automation, skills, organisation, small business, leadership support