They promised you time back. That was the pitch, said out loud and in writing: AI would handle the dull parts, shrink the admin, give you back the hours you were losing to the work no one wants to do. The four-day week came up. The language was intoxicating.

Which raises the question a lot of people are asking this week. If AI is doing this much of the work, why does your inbox feel heavier than it did a year ago?

You are not imagining it. The data caught up with the feeling this week, and it says the same thing you have been saying to a colleague in the kitchen, or to no one in particular at the end of another long day. The tool that was sold to you as a time-saver has, for most people, become a second job.

The Week the Promise Broke

A global workforce report landed this week from Korn Ferry, one of the largest consultancies in the world, based on a survey of more than 16,000 professionals across eleven countries. The headline figure should stop anyone who has been repeating the productivity story to themselves. Two-thirds of employees — 62% — report heavier workloads over the past two years. Close to half, 45%, say they are too busy to do work that matters.

Then the figure that matters most for this conversation. More than half of the workers surveyed — 52% — said the same thing, in plain language: using AI has increased their workload, not shrunk it. Not kept it the same. Increased it.

The report gave the feeling a name. They called it the "two-job job." More than three in five employees — 61% — say they are doing the responsibilities of more than one role. Layer AI on top of that, and the report's authors put it in plain terms: it can feel like doing two jobs instead of one.

Read that twice. The people who sold the tool told you it would take work off your plate. The people who bought the tool, your employers, told themselves the same story. And the workers using it — the ones whose weeks are made of this — report the opposite.

Two Floors, Two Realities

The same report holds the explanation, and it is worth sitting with. About 79% of chief executives say AI has improved efficiency in their organisation. Ask the individual contributors — the people doing the work the tool was meant to help with — and half say the same.

Same tool. Same company. Same quarter. Two different answers. The boardroom sees a dashboard that says things are moving faster. The desk sees a queue that is longer than it was last month. Both are telling the truth. The gap between them is the whole story.

Forbes, writing about the same data this week, called it "super-burden" — a play on the "superpowered" promise. The piece laid out the mechanism in plain terms. Organisations flattened their structures and cut managers. The managers left behind picked up the work of the ones who left, plus the new job of managing the AI itself. Workers kept after layoffs used the new tools to produce more, picked up the slack, and in many cases were not paid for the extra scope. The tool made people faster. The speed went to more work, not back to you as time.

That is the part the productivity story missed. Efficiency does not equal less work. It equals more expected. The capacity AI creates does not come back to you as a shorter day. More work fills it — your role expanding, the colleague who left not replaced, the tasks that fell to you without anyone asking.

This Is Not Perception. They Measured It.

This is where it gets harder to wave away as perception. A workplace analytics firm spent the better part of three years watching how work unfolds — not survey answers, observed behaviour, across more than 443 million hours of activity in over a thousand organisations.

Their finding, in one line: after people adopted AI, time spent in every single measured category of work went up. Not some. All of them. Email activity rose by over 100%. Chat and messaging by close to 150%. Time in business management tools by close to 95%. No activity category decreased. The workday got shorter, and the work inside it got denser.

Focus time — the uninterrupted stretches where thinking happens — fell to a three-year low. The average focused session lasts about thirteen minutes. Collaboration surged. Multitasking climbed. Weekend work went up by more than 40%.

Their phrasing was the cleanest version of the paradox anyone has managed. AI, they said, is functioning as an additional productivity layer — not a substitute for existing work. Not a replacement. An addition. You kept the old work. You gained a new layer on top of it. And the week did not get longer to hold it.

When the survey says 52% of workers feel AI increased their workload, that is not a feeling. It is a measured pattern. The inbox is heavier because it is heavier. The chat is louder because it is louder. You are not failing to adapt. You are adapting, in real time, to a week that has expanded to hold more than it was built to.

The Twist: The Work AI Added Is Not the Work That Matters

This is where the story turns, and the turn is the part worth your attention.

The same week the workload data landed, a separate analysis of 1.4 million job postings — across the US and the UK, every industry — asked a simpler question. What skills are employers asking for, in the ads?

Management appeared in 44% of postings. Public speaking in 40%. Problem-solving in 12%. AI tools, named in the ad? 0.2%. Not two percent. Zero point two.

Sit with the gap. The work AI multiplied in your week — the email, the chat, the task churn, the constant context-switching — is not the work employers are hiring for. The skills that show up in job ads, the ones that move careers and pay cheques, are the ones no tool licence gives you. Judgement. Leadership. The ability to stand up and make a case. The capacity to think through a problem that does not have a prompt for it.

The paradox has a second layer, and it is sharper than the first. AI gave you more than extra work. It gave you the wrong kind of work. The busyness it multiplied is the busyness that does not appear on a single job description worth having. And the skills that do appear — the ones that compound into a career — are the ones the thirteen-minute focus session and the 145% chat surge are eroding.

This is the uncomfortable arithmetic of the moment. You are busier than you were two years ago. The data says as much, in your own behaviour, measured across hundreds of millions of hours. And the work that made you busier is not the work that makes you valuable.

What the People Pulling Ahead Have

It is worth being precise about what separates the people navigating this shift from the ones swept along by it. It is not more tools. Everyone has the same tools — the licence arrived in your inbox, and in theirs. It is not more hours. The data says the hours are full.

What they have is harder to buy and harder to fake. They can look at a task and know whether AI belongs in it — and, more than people admit, whether it does not. They can tell a confident wrong answer from a reliable one, without a specialist on call. They have held on to the capacity for focused work in a week that is fragmenting it. And they have the human skills — the ones in 44% of job ads and 0.2% of tool licences — that no one is going to hand them a dashboard for.

None of that comes with a software rollout. You build it, over time, in the gaps the tools do not fill. And the gaps are widening. The workers who treat the busyness as the work will find, at the next restructure or the next interview, that they kept up with the queue and lost ground on everything else.

The Question Is Not Whether AI Adds Work. It Does.

The data has settled that. The question is what you do with the finding. You can keep adapting to a week that keeps expanding — faster email, more chat, shorter focus, more roles, the two-job job. Or you can step back and ask the question the reports are pointing at: which of the skills in your week are the ones that matter, and which are the ones AI multiplied while you were not looking?

That is not a question the tool will answer for you. It is not a question your employer is paid to ask. It is the question that separates the people being swept along by this shift from the ones staying ahead of it.

We mapped it. The Skills Economy 2026 is our free report, drawing on sixteen sources across five regions: the skills employers are hiring for, the ones disappearing from job ads, and the merge between technical and human skills that is redefining what "qualified" means in a year when 52% of workers got busier and 0.2% of job ads mentioned the tool that did it.

Download The Skills Economy 2026 report — free. Read it before your next performance review does the maths for you.

And If You Want the Structured Path

The report shows you the landscape. If you want to build the capability itself — the human skills the job ads are full of and the tool licences are not — our course range is where that happens. A short lesson in your inbox each day, built for working professionals, no technical background assumed:

• Async-First Productivity — tame the email and chat surge the data describes, and take back the focus time the tools have been eating.
• Emotional Intelligence at Work — the human skill that appears in job ads and in no software licence.
• Leading Without Authority — management is the most in-demand skill in 2026. This is how you build it without waiting for the title.

AI gave you more work. The question is whether you keep absorbing it, or start building the skills that make the work matter. Start with the report.