Generic advice about AI and jobs only goes so far. You want to know what will AI automate in YOUR job, in YOUR field. Which parts are most at risk? You can answer that yourself. You just need the right method. This method works for any field. Here is how to find out what AI automates first in your own work.

The core idea: tasks, not jobs

AI does not automate jobs. It automates tasks. Every job is a bundle of tasks. Some of those tasks are at high risk. Others are almost safe. The McKinsey research has said the same thing since 2023. The unit of automation is the task, not the job. So start by breaking your job into its tasks. Then look at each one on its own.

The risk test for any task

Look at each task in your job. Ask these questions. The more "yes" answers, the higher the risk:

Now flip it. The more "yes" here, the safer the task:

Putting it to work: a few examples

Take design. Making routine layout versions is at high risk. Senior creative direction and client relationships are safer. Take law. Document review and basic research are at risk. Courtroom work and tricky talks are safer. Take medicine. Image analysis and routine notes are at risk. Hands-on care and hard diagnosis are safer. Take finance. Routine models and reports are at risk. Advice built on trust and judgment calls are safer.

The pattern holds every time. Repetitive information work is at risk. Judgment, relationship, and physical work is safer.

What to do with your results

Now you have mapped your tasks by risk. The plan is clear. Move your time and skill toward the safer tasks. Learn to use AI for the risky ones. Do not fight the automation of your routine work. Welcome it. Then spend the freed-up time on judgment, relationships, and ownership. AI cannot do that work. This is how you grow more valuable as your field changes, not less.

The hard part

For some people, this test gives a tough answer. It shows that most of their job is at high risk. That is hard to face. But it is far better to know now. You still have time to shift toward safer work or build new skills. That beats finding out when the automation shows up. The test is a gift, even when the answer hurts. It buys you time to act.

Running this analysis on a team

You can run this same test on any team's roles. It shows which parts of the work are most at risk. It also shows which parts stay valuable. Those safer parts are creative judgment, client relationships, and brand strategy. The strongest move is to push work toward the high-judgment tasks. That is where people add the most value. Let AI handle the routine, at-risk work. The test gives you the map. Acting on it with care is what counts.

The honest take

AI automates tasks, not jobs. So break your job into tasks to see your real risk. Check each one. Repetitive, rule-based information tasks go first. Judgment, relationship, physical, and ownership tasks are safer. Map your own job in an honest way. Shift toward the safer work. Use AI for the rest. People who run this test early and act on it stay ahead of the automation. People who dodge the question get caught off guard.

Sources

Related reading: Will AI Change My Job? What the Data Says About Which Roles Are Actually Exposed · Which Jobs Are Most Affected by AI? The Data on Exposure and Adaptation