Want to work at UBS? You'll have to prove you're comfortable using AI. The Swiss financial services giant is requiring prospective junior investment bankers to demonstrate their familiarity with the technology as a condition of their employment starting next year, the Financial Times reports, citing unnamed sources. Rival Santander has made similar moves, seeking out "advanced AI users" for some roles, but UBS is among the first major financial institutions to make the skills mandatory. It's also experimenting with in-house uses for AI, like avatars that can make client presentations.
Your next interview may ask you to prove something your CV only claims.
“I know how to use AI.”
Okay. Show me.
UBS is introducing AI-proficiency requirements for junior investment-banking candidates from 2027, according to the Financial Times.
That caught my attention because this is not a technology company hiring AI engineers.
It's traditional banking.
And it raises an important question for every profession:
What does being “AI proficient” actually mean?
Is it knowing how to write a prompt?
Using ChatGPT every day?
Or being able to take a real business problem, use AI to solve part of it, and explain where human judgement is still needed?
From a recruitment perspective, I think the third is what we should be moving toward.
Imagine two candidates with similar qualifications and experience.
One says:
“I've completed five AI courses.”
The other says:
“Here's a workflow I built. This is the problem it solves, how I tested it, and where I kept human approval.”
The second candidate has given you something much more useful to evaluate.
Not because courses don't matter.
But because demonstrated capability is different from claimed capability.
For anyone thinking about their career, the practical takeaway is simple:
Don't just learn AI as a separate subject.
Learn how to apply it to your actual profession.
Recruitment. Finance. Research. Marketing. Operations. Teaching.
Pick one repetitive problem. Build a small solution. Measure what improves. Understand its limitations.
That is the kind of learning you can take into an interview.
And it makes me wonder whether the next big shift in hiring will be from:
“Which AI tools do you know?”
to:
“Show me what you can actually do with them.”
A degree and experience still matter. But the ability to demonstrate how you work with AI may increasingly become part of the evidence employers expect.
If you were hiring today, what would convince you that a candidate is genuinely AI-proficient?