AI for students.

Do this

Use an assistant to get a fast first-pass framework or find a comparable case before class — it compresses the warm-up, not the thinking. Then interrogate what it gave you: what's assumed versus shown, what breaks the framework, what the case's actual numbers say that a generic template wouldn't know.

Not this

Walking into class with an AI-generated answer you haven't pressure-tested yourself. In a cold call, you defend your reasoning out loud — an assistant can't do that part for you, and it shows immediately when someone hasn't done the interrogation.

Do this

Use AI to mock a guesstimate or case interview, draft a first pass at CV bullet points, or rehearse an answer out loud against a simulated interviewer. It's a genuinely good practice partner precisely because it's tireless and available at 2 a.m. before a deadline.

Not this

Treating a rehearsed AI answer as the actual interview. A recruiter is testing your live reasoning under pressure, not your ability to recite something an assistant helped you prepare. See the full placement guidance for the rest of the process — this is only the AI-specific piece of it.

Do this

Use AI to summarise a long reading fast, organise scattered notes, or restructure a messy first draft. That's real time saved on the parts of writing that were never the actual assignment.

Not this

Letting AI write the actual argument, or submitting a fact, statistic or citation you haven't personally verified. Know your programme's disclosure policy and follow it — an unverified AI citation in a submitted assignment is a preventable, embarrassing failure mode, not a shortcut.

Do this

The fastest way to actually understand what these tools can and can't do is to try building something with them — a small prototype, not a polished product. That's the entire premise of the AI Mini Hackathon.

Not this

Staying purely on the consumption side — using AI only to answer questions faster — without ever testing what happens when you try to make it do something new. The gap between using a tool and building with one is where the real learning is.