Work,
Augmented.

Your job is not disappearing.
It is being rewritten.

Work rarely changes as cleanly as a job title disappearing on Friday and a new one appearing on Monday. It changes task by task.

A report that took a morning can become a reviewable first draft in a fraction of the time. A first pass through a hundred pages can happen before the meeting starts. A blank presentation can arrive with a structure, supporting material, and a list of the questions it still cannot answer.

Some of those tasks will be automated. Many will be shared between a person and a machine. The rest will become more important because everything around them has become cheaper.

This is already visible in the data. Anthropic's research finds AI use split between automation and augmentation, with augmentation slightly ahead in its consumer product. The World Economic Forum expects nearly 40% of job skills to change by 2030 while human skills such as analytical thinking, leadership, creativity, and collaboration remain critical. The transition is real. It is also messier than the headlines suggest.

So no, we do not know exactly which jobs will exist in five years. Anyone who claims to know is selling certainty they do not have.

We do know the work inside those jobs is changing now.

Access is not the problem anymore.

Most leaders we talk to have tried ChatGPT, Claude, or Copilot. They ask for a summary, an email, or a strategy. The answer is technically competent and practically useless. It sounds like something. It does not know enough to mean anything.

So they conclude that AI is overhyped.

Sometimes it is. But often the problem is simpler: they handed a capable assistant a one-sentence request with no history, no examples, no constraints, and no definition of good. Then they judged it for guessing badly.

Generic context produces generic work.

The next step is not a cleverer prompt. It is a better working relationship.

The new job is to direct the work.

When execution becomes cheaper, direction becomes more valuable.

That does not mean sitting above the work and issuing commands. It means doing the part most people skip because it is hard: deciding what should happen, explaining why, defining what good looks like, and checking whether the result actually meets the bar.

This is what strong managers already do with people. AI makes the practice explicit.

You need to know:

• What outcome matters.
• What context changes the answer.
• What can be delegated and what cannot.
• What a good result looks like.
• How the work will be checked.
• Who remains accountable when it is wrong.

The answer to the last question is still a person.

Microsoft's 2026 Work Trend Index describes its most advanced AI users in similar terms: they delegate routine execution, research, and synthesis while people set direction and take responsibility for how the output is used. The tool may do more of the execution. It does not inherit your accountability.

THE PRACTICE

Six principles for augmented work

The catch
(because there is always a catch)

Augmentation takes work before it saves work.

You have to document things that have lived in your head for years. You have to confront unclear goals and inconsistent processes. You have to decide what information can enter which system. You have to review output even when it looks convincing. The first version of a workflow will probably disappoint you.

And some work should not use AI at all.

If the risk cannot be controlled, the information should not be shared, or the task depends on a relationship that automation would cheapen, leave it alone. A spreadsheet, a checklist, or an honest conversation may be the better tool.

This is not a race to put AI everywhere. It is a practice of putting it somewhere useful and learning from what happens.

What this looks like in an actual week

An augmented leader does not spend the day talking to a chatbot.

Their system prepares a morning brief from active projects and open commitments. It turns approved numbers into a first report, then flags the claims that need a source. It builds a meeting pack from the history of an account. It drafts the difficult email in the leader's own language and explains which assumptions shaped the draft. It compares a new request with the goals the leader already committed to.

The leader reads, questions, corrects, decides, and communicates.

The machine handles more of the preparation. The person brings more of the judgment.

That is the promise of augmented work. Not more output for its own sake. More capacity for the work that moves something forward.