The Future of Work in 10 Years: What Work Could Look Like by 2036

Ten years ago, few people expected video calls to replace most meetings or a chatbot to write first drafts of their code. Long-range forecasts about work have a poor track record.

So this look at the future of work in 10 years sticks to what can be traced: forecasts from the World Economic Forum and McKinsey, data on remote work from Stanford, and the largest four-day-week trial so far. Each section starts with what the data says. Where we go further and guess at 2036, we mark it as opinion.

What Do the Big Forecasts Say About Jobs?

They expect churn, not collapse. The World Economic Forum's Future of Jobs Report 2025, based on data from more than 1,000 companies, projects that by 2030:

• 170 million new roles will be created and 92 million displaced, a net gain of 78 million.
• Job disruption will equal 22% of today's jobs.
• Nearly 40% of the skills required on the job will change.

The fastest growth in absolute numbers is in frontline roles like farmworkers, delivery drivers and construction workers, followed by care and education. Technology roles in AI, data and cybersecurity grow fastest in percentage terms. Roles like cashiers and administrative assistants keep declining, and the WEF now adds graphic designers to that list.

Most forecasts stop at 2030. For 2036, we have to extend the lines ourselves.

How Much Work Could AI Actually Take Over?

More tasks than jobs. McKinsey Global Institute's report "Agents, robots, and us" finds that technologies already demonstrated today could, in theory, automate activities that account for about 57% of US work hours.

McKinsey is explicit that this is technical potential, not a forecast of job losses. Adoption takes time, and most skills survive: more than 70% of the skills employers seek today are used in both automatable and non-automatable work. What changes is how they're applied.

Two more numbers from the same report point at the next decade:

• Demand for AI fluency, the ability to use and manage AI tools, grew sevenfold in two years in US job postings, faster than any other skill.
• About $2.9 trillion of US economic value could be unlocked by 2030, but only if organizations redesign whole workflows around people, agents and robots, not single tasks.

Our forecast (opinion): by 2036, most knowledge workers won't do fewer things. They'll do different things, with a large part of the doing handed to agents and a large part of their day spent checking it.

Will AI Replace People or Work Alongside Them?

So far, mostly alongside. Anthropic's Economic Index (January 2026) splits Claude conversations into automation, where the user hands over a task, and augmentation, where they work through it together. In November 2025, 52% of Claude.ai conversations were augmentation and 45% automation, a shift back toward collaboration after automation had briefly pulled ahead in August 2025.

That back-and-forth is a useful signal. Delegation grows when tools get better, then people pull back to iterate when the work gets harder.

Our view (opinion): in ten years the split won't be "human jobs" and "AI jobs". It will be inside every job: the routine part runs on its own, and the judgment part stays with a person who is accountable for the result. Microsoft's 2025 Work Trend Index calls that person an "agent boss". We'd just call it a manager of work instead of a doer of it.

We wrote about what still counts as real work once execution gets cheap in what is actual work in the AI era.

Where Will People Work in 2036?

Probably where they work now: mostly hybrid. A 2025 work-from-home analysis from Stanford SIEPR, drawn from a 40-country survey, found that work from home fell from an average of 1.6 days a week in 2022 to 1.33 in 2023 and 1.27 in 2024/2025. After 2023, it stopped falling.

Photo: Rodeo Software on Pexels

The pattern isn't uniform. English-speaking countries average about 1.5 to 2 days a week at home, Europe slightly less, Asia 0.5 to 1 day. Parents are more likely to be hybrid, while people without children are more often fully in person or fully remote.

Our forecast (opinion): the "return to office" fight is mostly over. In 2036, the default for desk work will still be two or three office days a week. The bigger change will be in who you work with: more contractors, more solo operators and more agents on the team. Stack Overflow's 2026 survey already shows developers working in an organization of one rising from 4% to 10%.

Will the Workweek Get Shorter?

The evidence for trying is getting stronger. The largest four-day-week study so far, published in Nature Human Behaviour in July 2025, followed 2,896 employees at 141 companies in six countries. A six-month move to four days without lost pay reduced burnout, raised job satisfaction and improved mental and physical health.

Photo: Min An on Pexels

The design detail matters more than the headline. Before the trial, each company spent about eight weeks restructuring work to hold output at 80% of the previous hours, mostly by cutting time-wasters like unnecessary meetings. Scientific American's report on the study notes that 90% of the companies kept the arrangement afterwards.

Our forecast (opinion): a shorter week won't arrive by law in most countries within ten years. It will arrive company by company, in places that can prove where their hours actually go. You can't cut 20% of the week if you don't know which 20% is waste.

What Could a Normal Workday Look Like in 2036?

This section is a scenario, not a forecast backed by a dataset. It's our best guess at how the trends above combine for one knowledge worker. Call her a product designer at a 40-person company.

• 8:30. She reads a summary of what her agents did overnight: three layout variants, a usability test analysis, a draft spec. Two are fine. One is subtly wrong, and catching that is the most valuable thing she'll do before lunch.
• 10:00. A focused block on the part no agent handles well: deciding what the product should do next, and why.
• 13:00. An office day. She meets the team in person for the decisions that need trust and argument, not status updates.
• 15:00. She writes briefs for the next agent runs and reviews a contractor's work.
• 17:00. She checks her own time data. Six hours on judgment and direction, two on review. That ratio, not hours logged, is what her company now asks about.

Nothing in this day requires a technology breakthrough. Every part of it exists in some form today. What changes in ten years is that it becomes ordinary.

Which Skills Will Matter Most in 10 Years?

The WEF expects technology skills in AI, big data and cybersecurity to grow fastest. It also expects human skills such as creative thinking, resilience, flexibility and agility to stay in high demand, and calls the combination of both increasingly crucial. 63% of employers already name skills gaps as their main barrier to change.

Based on the data above, we'd add three skills to the list (opinion):

• Specifying work. Writing a brief clear enough that a person or an agent can finish it without you.
• Verifying work. Spotting what's wrong in output you didn't produce, fast.
• Measuring your own time. Knowing, with data, where your hours go.

How Will Work Be Measured When Output Is Cheap?

This is the open question of the next decade. When an agent can produce a report, a landing page or a pull request in minutes, counting output tells you less. Counting hours at a desk tells you even less.

Our view (opinion): measurement will move from "how much did you produce" to "where did your attention go, and did it go to the right things". That kind of measurement only works if it's private and owned by the person being measured. Tools that watch employees for managers will keep losing trust.

That's the approach we took with Didon. It takes a screenshot every 30 seconds, analyzes it on your Mac with a local model, and writes a daily journal with time per project and a short summary. Nothing leaves your machine. It's a small, present-day version of the self-measurement we expect to become normal. For the patterns it tends to surface, see hidden productivity killers.

What Can You Do Now to Prepare?

• Learn to direct AI, not just use it. AI fluency is the fastest-growing skill in US job postings.
• Keep your core skills sharp. You need them to judge what an agent hands back.
• Expect hybrid. Plan for two or three office days, not a full return or a full escape.
• Find your wasted 20%. If your company ever trials a shorter week, you'll want to know where your hours go.
• Track your own time, privately. A week of honest data beats any forecast about the average worker.

Nobody knows exactly what work looks like in 2036. But the people who see their own week clearly will adapt to it fastest.