UBS Wants Proof You Can Use AI Before It Will Hire You
The Swiss bank has turned AI fluency into a screening criterion for its 2027 junior intake. The rest of European banking is watching.
If this caught your attention, that’s not accidental.
The best editorial systems don’t happen by accident. Outlever builds them.

Landing a junior job in investment banking used to mean a good degree, a decent internship and the stamina to get through several rounds of interviews. At UBS, it now also means showing you know your way around AI.
The Swiss bank has made AI proficiency a condition of employment for graduates and interns joining its global banking and markets division in 2027, the Financial Times reported, citing people familiar with the plans. Applicants will have to point to specific ways they have used the technology to make their work better or faster. Questions about AI are being added to recruitment interviews, sitting alongside the usual ones about deal structures and why you want the job.
Plenty of banks have said warm things about AI skills. UBS is one of the first to write the requirement into the hiring criteria and then say so out loud.
What counts as proof
The bar is not "I have a ChatGPT account." According to TechRadar's read of the FT report, candidates need to show they can experiment with the tools responsibly and turn that into a better business outcome. Curiosity and a willingness to keep learning appear to matter as much as any particular tool.
Everything else stays where it was. A 2:1 degree is still the baseline, and UBS has said AI literacy sits next to analytical and interpersonal ability rather than displacing it.
What happens after the offer letter matters just as much. Graduates entering the bank's Graduate Talent Program go through something UBS calls the AI Fluency Pathway, which covers real banking use cases, responsible application and the judgment to know when an output is wrong. That is a useful tell. A bank expecting finished expertise on day one would not need to build a training track for it.
The internal groundwork has been going on for a while. UBS's own technology pages say more than 38,000 employees have completed AI learning journeys, more than 49,000 have enrolled in AI courses since 2024, and its 2026 Global Learning Weeks ran 56 AI sessions with over 45,000 enrollments. Staff can drop into a weekly "AI power hour."
Some of this has already reached the pay conversation. German outlet ad-hoc-news reported that AI capability now forms part of individual performance assessments at the bank, citing group HR head Stefan Seiler, with usage tracked at an aggregated level rather than per employee.
The avatars
UBS has been experimenting with more visible uses of the technology too. Since early 2025 it has been building AI video avatars of its research analysts, using models from OpenAI and video tools from Synthesia. Analysts sit for a studio session that captures their likeness and voice. A model then reads their published notes, drafts a script, and turns the approved version into a short video for clients.
Roughly 35 of the bank's 720 analysts had digital versions when Finextra covered the rollout last year. Scott Solomon, who runs global research technology at UBS's investment bank, told the FT the project had two drivers: clients asking for more video, and the time it frees up for analysts. Studio capacity had capped output at about 1,000 videos a year. The target with avatars is 5,000. Participation is voluntary.
Not just UBS
Santander has been looking for what the FT described as "advanced AI users" on some of its trainee programs, which suggests the practice is spreading rather than starting. Neither bank is doing this in a vacuum.
The backdrop is a squeeze on entry-level banking work that has been building for two years. Debasish Patnaik, who leads McKinsey's QuantumBlack AI practice, told Fortune that banks are cutting junior analyst classes by as much as two-thirds while drawing roughly 62% of their AI talent from those same cohorts. "Banking is an apprenticeship business," he said, warning that senior judgment cannot simply be hired in later.
The projections above that layer are larger. Morgan Stanley analysts estimated that more than 200,000 European banking jobs could go by 2030, about 10% of the workforce across 35 major lenders, with back-office, risk and compliance roles taking the brunt. Bloomberg later reported the same team putting the range at 10% to 20% over five years.
Read against those numbers, UBS's move looks less like a headcount story and more like a specification change. The graduate job still exists. The price of getting one went up.
The part nobody has solved
There is an obvious objection, and senior bankers have been making it themselves.
Nestor Paz-Galindo, UBS's head of global banking for EMEA, gave the optimistic version to Financial News back in 2024. AI would absorb a lot of what analysts spend their time on and free them for more bespoke work, he said, "but not replace them." Executives at JPMorgan have been more cautious in public, warning that the industry needs to be careful about juniors losing their grip on the fundamentals.
The research is not reassuring. A Nature Medicine paper published in May coined the term "never-skilling" for what happens when a novice leans on AI before developing the underlying reasoning, distinguishing it from the deskilling that affects experts who stop practicing. The paper is about medical training, but the mechanism travels. A randomized trial by Anthropic researchers published in January found junior engineers who used an AI assistant to learn an unfamiliar Python library scored 50% on a comprehension quiz, against 67% for those who wrote the code themselves.
Screening for AI proficiency at the interview stage does nothing about that problem. It might even select for it.
Building comps, checking data, formatting decks and grinding through market research were never valuable in themselves. They were how young bankers built pattern recognition before they had judgment. Hand that work to software and the industry still owes an answer on where the judgment comes from.
UBS is betting the Fluency Pathway can do both jobs at once: screen for the skill that replaced the apprenticeship, and rebuild the apprenticeship around it. Nobody finds out whether that worked until the class of 2027 is running deals, which is the better part of a decade away.
If this caught your attention, that’s not accidental.
The best editorial systems don’t happen by accident. Outlever builds them.


Get the latest AI insights first.
Sign up for updates, interviews, and fresh analysis on how AI is reshaping business, brands, and technology.





