Anything you do repeatedly can potentially be automated.
Think about the work you’d gladly hand off tomorrow. An accountant might pick the monthly invoice check. A purchasing manager might pick the first pass through supplier quotes. A customer support rep probably has a particular email in mind. Once AI can handle that task, a harder question comes up: What will your employer count on you for next?
Start with one assignment you finished last month. What took the most time? Where did you need to make a judgment call? The examples below come from a team that supports online stores, but you can ask the same questions about a purchase order, a customer complaint, or a monthly report.
Look for tasks you repeated last month
Become the person who knows your part of the business best
Find out why the work is done this way
Agree with your manager on how to use the time AI saves
Look for other roles that need your experience
Save examples you can talk about in an interview
Summary
Open your list of completed assignments. If you don’t keep one, check your calendar and sent emails. Pick a few items and retrace what you actually did.
Take a support ticket labeled “Fix promo code.” Those three words can hide half a day’s work. A customer reported that their discount disappeared. The developer tried the code, and it worked. Only after talking with customer support did the developer learn the missing detail: The customer applied the code and then added another item to the cart. With those steps, the developer could reproduce the problem, fix the calculation, and test it.
Now look at two or three more assignments. Maybe you spent half an hour preparing similar test data each time. That’s a good task to try with AI. Or maybe most of the time went into chasing down details missing from the original request. In that case, adding one question to the support form might save more time: “What did you do just before the discount disappeared?”
Tasks with different names can contain the same repetitive step. Look for that step before deciding what to automate.
Suppose a store offers a discount on orders of $100 or more. An AI tool writes two tests. A $90 order gets no discount. A $110 order gets the discount. Both tests pass. Would you sign off on the promotion?
If you know the offer, your first question is probably, “What happens when the cart total is exactly $100?” That’s where the difference between “$100 or more” and “more than $100” shows up. The developer might catch it. So might the account manager who worked out the terms with the client.
Try AI first on work you know well enough to check yourself. Which rule applies? What exception would a newcomer miss? When could an answer look right and still be wrong? If you can show a colleague the exact case that would expose a mistake, you’re in a good position to decide which part of the job to hand off.
A shopper receives two confirmation emails for one order. An AI tool reviews the code, spots a second attempt to send the message, and suggests removing it. The team places a test order. One email arrives. So far, so good.
But why was that second attempt there in the first place? An old support ticket tells the story. During an earlier outage, the store kept accepting orders and charging customers, but the email service temporarily stopped accepting messages. Customers started calling to ask, “Did my order actually go through?” The team added the second attempt after that outage.
Now there are two things to check. If the email service rejects the first attempt, the store should try again. If the service has already accepted the message, the customer should receive just one copy.
Most workplaces have a step that looks unnecessary to someone new: a second review, an extra field on a request, or approval before an order ships. Before you automate it away, ask someone who remembers what happened before it existed. Their answer will help you decide whether that step still earns its place.
First, count the time you actually saved. If AI produces a draft in one minute and you spend another 40 fixing it, the task took 41 minutes. Suppose you’ve tracked several assignments from start to finish and found that you freed up two hours this week.
What has the team been putting off? Maybe support has asked for a shipping address correction three times this month. A customer enters a new address, but the old one sometimes ends up on the order. Each correction takes only a few minutes, so nobody has investigated the cause.
That gives you something specific to discuss with your manager: “I saved two hours on testing this week. I can take the next ticket, or I can review those three address mix-ups with support. Which would help the team more?”
If a client launches tomorrow, the next ticket might take priority. Or your manager might be hearing about the address problem for the first time. If you investigate it, check back with support after the fix to see whether new cases are still coming in. That tells you both what the team got out of those two hours.
Imagine you’re a designer at a tech company. Fewer requests for standard banners are coming in because clients now create some versions themselves. But after years of designing online stores, you’ve also learned to spot where shoppers get stuck. You may already have a good idea why people abandon checkout on their phones just as they’re asked to choose a delivery option.
Where else could you use that skill? Open a few job postings you’d genuinely consider. Read the day-to-day responsibilities. When you recognize part of the work, talk to someone who does it. The designer in our example could join the team testing the store’s mobile checkout and see whether investigating shopper behavior is something they enjoy.
Look for a small assignment in your own department or on a neighboring team. Trying the work will tell you which skills carry over and what you’d need to learn before taking it on independently.
Even a skilled employee can lose a job. A company might lose a major client, close a business unit, or decide to cut staff. You may have little influence over that decision. You can, however, keep a record of work you’ll want to discuss if you start looking for another role.
After a difficult assignment, jot down a few sentences. What problem were you asked to solve? What did you personally do? How did you check the result? Use your own words and leave out customer data and internal documents.
Six months later, “helped automate business processes” is easy to put on a résumé. This gives an interviewer more to work with: “Support had corrected shipping addresses manually three times. I reviewed those cases, found the step that caused the old address to reappear, and tested the fix with a colleague.” Now the interviewer knows how you approach a problem and can ask about a decision you actually made.
Pick one task from last month that you remember well. Which part could you hand to AI next time? Which part would you feel comfortable taking responsibility for yourself? Your answers give you something concrete to discuss with your manager now and, if you ever need it, in your next interview.
