Audit Your Role: Which of Your Tasks AI Can Automate

Stop guessing whether AI affects your job. Build a clear, honest map of your own tasks in one sitting.

In Part 1, we learned that AI replaces tasks, not whole jobs. That idea is comforting. But it stays abstract until you make it personal.

So let's make it personal. Today you build a map of your own role.

This is not a test. It's a flashlight. You're going to shine it on your week and see exactly where AI fits and where it doesn't. By the end, vague worry becomes a clear list you can act on.

Why an audit beats worrying

Anxiety loves the unknown. When you can't name the threat, your brain imagines the worst.

An audit removes the fog. You stop asking "will AI take my job?" and start asking "which of my tasks can AI touch, and what do I do about it?" That second question has answers.

Here's the reframe. Every task you list is a task you now control. You can keep it, speed it up, or hand it off. That choice is yours.

You don't need special software. A notebook, a spreadsheet, or a blank document all work. The thinking matters more than the tool.

Step one: capture every recurring task

Most of us underestimate how much we do. The work blurs together. So we'll slow down and write it out.

Track your tasks for one normal week. Each time you switch to something new, jot it down. Don't judge it yet. Just capture.

Write tasks as verbs, not job titles. "Write the weekly status email" is a task. "Communications" is not. The more specific you get, the more useful the audit.

Include the small stuff. Scheduling, formatting, chasing approvals, and copying data between systems all count. These hidden tasks are often where AI helps most.

Step two: sort each task into three buckets

Now you have a raw list. Time to label it. We'll use three buckets that are easy to remember.

The buckets are Assist, Automate, and Anchor.

• Assist: AI helps, but you stay in charge. You draft, it speeds you up, you decide. Most tasks land here.
• Automate: A tool could do nearly the whole thing with little input from you. Often repetitive, rule-based work.
• Anchor: Tasks that rely on trust, judgment, relationships, or real-world presence. AI barely touches these. They anchor your value.

Go down your list and mark each task with one label. Trust your gut on the first pass. You can adjust later.

If a task feels like it sits between two buckets, that's normal. Pick the closest fit. The goal is a useful picture, not a perfect one.

Step three: pressure-test your "Anchor" tasks

We all want to believe our work is safe. That hope can make us label too many tasks as Anchor. So let's test them.

For each Anchor task, ask one honest question. "If I gave a smart assistant good instructions, how much of this could they do?"

If the answer is "quite a lot," it's probably an Assist task, not an Anchor. Move it.

True Anchor tasks share a few traits. They need accountability, a relationship, taste built over years, or a decision someone must own. A machine can suggest. It can't be responsible.

Here's a quick gut check you can run on any task. Notice how the same skill repeats across answers.

Pressure-test prompt for one task:

For the task "[describe your task]", answer briefly:
1. Is the goal clear and repeatable, or does it change each time?
2. Does it require judgment or accountability a person must own?
3. Could a careful assistant do it with written instructions?
4. Suggested bucket: Assist, Automate, or Anchor — and why.

Drop that into the AI prompt generator or any chat tool. Run it on your trickiest tasks. The answers often surprise people.

Step four: spot the "Assist" wins worth claiming

Your Assist column is gold. These are tasks where AI makes you faster without removing you.

Scan that column for two things. First, tasks you do often. Second, tasks you find boring or slow. Where those overlap, you have a fast win.

A win means you keep ownership but cut the time in half. You're still the author. The tool is the helper.

You stay the editor and the sender. That distinction matters. In Part 4 we'll go deep on directing AI output like a manager. For now, just flag the wins.

Want a head start? Browse ready-made expert templates for common office tasks. Reusable, well-built prompts beat starting from a blank box every time.

Step five: turn the map into a one-page plan

A list is good. A plan is better. Let's compress your audit into something you'll actually use.

Make a simple one-pager with three short sections.

That's it. Three moves. Notice the shape of it. You offload the boring, speed up the routine, and pour the saved time into work only you can do.

This is the quiet secret of career-proofing. You don't fight automation. You ride it toward higher-value work.

What your audit is telling you

Step back and read your buckets like a chart.

Lots of Automate tasks? You have time to reclaim, and a reason to learn the tools fast. That's opportunity, not threat.

Lots of Assist tasks? You're in the sweet spot. AI makes you stronger while keeping you in the driver's seat.

Lots of Anchor tasks? Your role leans on trust and judgment. Protect that, and use AI to clear the busywork around it.

Most people find a mix of all three. That mix is healthy. It means you have levers to pull in every direction.

The big shift is psychological. You walked in worried about a force you couldn't see. You're walking out with a labeled map and a short plan. That's agency.

Your self-assessment

Before you move on, do this one thing. Count your tasks in each bucket.

Write down the totals: how many Assist, how many Automate, how many Anchor. Then circle the single task that bothers you most.

That circled task is your starting point. In the next parts, you'll learn the human skills that make you the one who runs these tools well, not the one replaced by them.

You've turned fear into an inventory. That's the hardest step, and you just took it.

Keep going

Next → Part 3: The Skills That Compound — Judgment, Taste, Knowing What to Ask

Or see the full Career-Proof: Staying Valuable as AI Reshapes Your Job series.