AI won't take your job. Someone using AI might.

It's the sentence everyone has heard in some version, and it's both true and unhelpful. It scares people without telling them what to do.

So here's a more practical way to look at it.

Jobs are bundles of tasks
Almost no role is one thing. It's a mix of tasks: some creative, some repetitive, some relational. AI is very good at certain tasks and poor at others. What changes is the bundle, not the existence of the job.

The tasks that go first
The ones you'd happily hand off if you could: summarizing, drafting, sorting, reporting, first-pass research. For most people, the reaction isn't loss. It's relief.

What becomes more valuable
Judgment, taste, trust and context. Knowing what to ask, what to ignore, and when the output is wrong. These skills grow in value because AI makes everything else cheaper.

Much of this is covered in my upcoming book, When AI Takes Your Job – And You Love It. It looks at where our work sits between what AI can do and what we're allowed to do, and how to find the work that's still yours.

What to do on Monday

• Pick one task you do every week that you dislike.
• Try to hand it to an AI. Expect a mediocre first result.
• Refine it three times. Notice how quickly it improves.
• Repeat with the next task.

Within a few months, you'll have shifted your own bundle toward the work only you can do.

For leaders
Don't ask "which roles can we cut?" Ask "which tasks can we free our people from, and what could they do with that time?" The companies that get this right will grow, not shrink.

The risk isn't AI. It's waiting while others learn to use it.

The book launches at the end of January 2027. Join the waitlist at marcustornell.com.