There are three ways to respond when a user asks for something. They look similar from the outside. They are very different jobs.
Level 1: Do what they asked
A user says: “I want a dashboard with all 30 numbers on one screen.” You build that dashboard.
You were responsive. But you handed the design thinking to the user. Users are experts in their own pain. They are rarely experts in interface design. They describe solutions based on software they have already seen.
Level 2: Find the need behind the request
You dig deeper. You talk to them. You find out they do not want a wall of numbers. They are scared of missing an urgent alert on an account.
So instead of a crowded dashboard, you design a notification that tells them when something needs attention. That takes real skill. You listened past the words.
Level 3: Ask why the problem exists
Now you go one step further. Instead of designing a better alert, you ask:
- Why is this person watching these numbers by hand at all?
- What is broken upstream that keeps creating these emergencies?
- Can we remove this task so they never have to worry about it?
Sometimes the best interface is no interface.
Where AI fits
- Level 1 can be automated. Turning a request into screens is something AI does quickly.
- Level 2 can be assisted. Give AI lots of real feedback and it can help you spot patterns and suggest what people might really need.
- Level 3 cannot be automated. AI was never in the room.
Empathy is not a maths problem. It comes from being there and noticing things. A user who pauses before clicking. A notebook kept next to the computer because they do not trust the software. The quiet sigh before they say “it’s fine”.
None of that shows up in a transcript in a way a model can weigh properly. And none of it can be found by a tool that only sees what you paste in.
Why this matters for your career
If your work is mostly level 1, you are doing the part that is easiest to hand to a machine.
Level 2 and level 3 are where your value lives. They are also the parts that are hardest to show in a portfolio, so they often go unnoticed. Start naming them. When you change a project because you asked why, say so.
The one line to remember
AI works on the problem you hand it. It will never ask whether that problem should exist in the first place.
Asking that question, out loud, in a room where it is uncomfortable, is the real job.
So what should you hand over, and what should you keep? That is next: the fast intern rule.