Junior Bankers at UBS Must Now Prove They Can Use AI…

UBS has added a new condition to getting hired as a junior banker, and it is one very few employers have written down this plainly. According to reporting from the Financial Times, graduates and interns applying to the Swiss bank's global banking and markets division must now show they can already use AI to improve outcomes and efficiency, making UBS one of the first major financial institutions to require AI literacy explicitly rather than treat it as a bonus. At Metaintro, we track hiring criteria as they change because a requirement written into a graduate scheme this year usually becomes an industry standard within two. Here is what the bank is asking for, what it already runs internally, and what it means if you are trying to get in.

What Exactly Is UBS Asking Junior Candidates to Prove?

The requirement is narrower and more practical than the headlines suggest. UBS is not asking graduates to build models or write production code. It is asking them to demonstrate that they can apply AI tools to the actual work of a junior banker, which means research, analysis and client preparation, and to do it in a way that improves the outcome rather than just the speed. That distinction matters when you are preparing for the interview, because a candidate who says they use AI to write first drafts is answering a different question from one who can explain what they checked before they trusted the output.

The condition applies to people joining the global banking and markets division from 2027, which gives current students and interns a full cycle to prepare. That is unusually generous notice for a hiring change of this kind. Most employers adjust their screening criteria quietly and candidates only discover the new bar when they fail to clear it. Publishing the requirement ahead of the intake turns it from a filter into a study guide, and the students who treat it that way will have a measurable advantage over the ones who assume their degree still covers it.

It also sits alongside the criteria that already existed rather than replacing them. Academic record and demonstrated interest in finance still matter. What has changed is that a third column has been added to the scorecard, and it is one that most finance degrees do not teach directly. AI literacy at entry level is a lower bar than most candidates assume and a considerably more specific one.

Why Would a Bank Screen for AI Before It Trains You?

Because the bank is not short of training. It is short of time. UBS states on its early careers pages that it is investing in AI upskilling from day one, and that new joiners get access to hands-on learning, internal AI certifications and global knowledge-sharing communities. An employer that already offers all of that and still screens for the skill on entry is telling you something specific: the training is for going deeper, not for starting from zero.

There is a second reason, and it is the more interesting one for anyone trying to read where entry-level work is heading. The traditional junior banking apprenticeship ran on volume. Building comparison tables, checking data, formatting decks and grinding through market research gave analysts pattern recognition long before they had judgment. When software absorbs a large share of that work, the bank still needs the judgment, but the old route to acquiring it gets shorter. Screening for AI capability on the way in is one answer to that problem, because it lets the first year be spent on the parts of the job the tools cannot do.

Whether that answer works is genuinely unresolved across the industry. AI is already reshaping analyst hiring on Wall Street, and in London the same pressure has been quietly erasing junior coder, lawyer and analyst roles. Against that backdrop, a bank that responds by raising the entry bar rather than shrinking the intake is taking a different bet from most of its peers, and it is a bet that favours prepared candidates.

How Much AI Training Does UBS Already Run Internally?

The internal numbers are unusually public, and they give the hiring change context that the announcement alone does not. On its own technology pages, UBS reports that more than 38,000 employees have completed AI learning journeys and that more than 49,000 have enrolled in AI-related online courses since 2024. By July 2026, more than 21,000 employees had earned its internal AI Citizen badge, a figure the bank puts at 28 percent of its workforce.

The programme runs upward as well as outward. UBS says more than 96 percent of managers had completed the initial stages of its AI Enabler track, and that roughly 1,300 senior leaders, including the Group Executive Board, had completed an AI Senior Leadership Journey. Its 2026 Global Learning Weeks included 56 AI sessions with more than 45,000 enrollments, and 12,000 employees took part in role-specific AI training.

Read those numbers next to the hiring requirement and the logic becomes clear. A bank where the executive board, most managers and more than a quarter of staff have been through structured AI training is not going to accept a graduate intake that arrives without any. The requirement is less a prediction about the future of work than an attempt to stop the newest employees being the least equipped ones in the room. That is a very different situation from most workplaces, where only a fraction of junior staff report having the AI tools they need.

What Does This Change About Breaking Into Banking?

It raises the floor without raising the ceiling, and that is better news for candidates than it sounds. AI proficiency is one of the few hiring criteria that does not correlate with which university you attended or who your parents know. A student at a non-target school who has genuinely learned to use these tools on real work can now demonstrate something a better-connected candidate may not have. In a field where entry has historically been gated by network and pedigree, a skills test that anyone can prepare for is a rare opening.

The wider graduate market makes that opening worth taking seriously. The Federal Reserve Bank of New York reported that labour market conditions remained challenging for recent college graduates through the second quarter of 2026, with an unemployment rate of around 5.6 percent and an underemployment rate of 42 percent. Almost half of recent graduates who are working are in jobs that do not require their degree. Against that, a named, published, learnable requirement at a major bank is a clearer target than most of the market offers.

Finance itself is not shrinking as fast as the discourse suggests. The US Bureau of Labor Statistics projects employment of financial analysts to grow 7 percent between 2025 and 2035, which it classifies as much faster than average, adding about 32,000 jobs to a base of 443,100. Median pay was $103,570 in 2025 and the typical entry requirement remains a bachelor's degree. The jobs are still there. The description of what a first-year does inside them is what is changing.

Is the Junior Analyst Job Being Rewritten or Removed?

Both, depending on the employer, and UBS has picked the first. The bank's public framing is that AI takes on work analysts currently spend a great deal of time on and frees them for higher-value tasks. That is the optimistic reading, and it is not obviously wrong, but it leaves a real question open. If the boring work was also the teaching mechanism, removing it does not automatically produce better analysts. It produces analysts who reach judgment-heavy work sooner with less pattern recognition behind them.

The honest answer is that nobody in the industry knows yet which model wins. Some employers are cutting intake and expecting fewer, more senior people to do more with better tools. Others, including UBS here, are keeping intake and changing what they expect the intake to arrive knowing. The second approach is better for anyone currently trying to enter the field, because it preserves the door. It also puts more responsibility on the candidate, since the bank is explicitly no longer promising to teach the basics from scratch.

What is not in dispute is the direction. AI skills now appear in a large and rising share of job postings, and the entry-level squeeze has been raising the productivity bar for new hires across sectors well beyond banking. UBS is not creating that trend. It is writing it into a job description, which makes it visible and therefore actionable.

What Does This Mean for Your Career Right Now?

If you are targeting a graduate scheme in finance, treat this as a syllabus rather than a threat. You have until the 2027 intake, and the requirement is specific enough to prepare for. Learn the tools on tasks that resemble the job: summarising a set of filings, building a first-pass comparison of similar companies, drafting a client-ready brief and then checking every number in it. The checking is the part most candidates skip, and in an interview it is the part that separates a real answer from a rehearsed one.

If you are already in an analyst seat somewhere else, the transferable move is to document what you have automated and what it saved. A specific example carries weight in a way that a certificate does not. Being able to say you cut a recurring research task from six hours to ninety minutes and describe the errors you caught while doing it is stronger than any badge, and it works in interviews at employers who have not published a requirement like this one.

If you are switching into finance from another field, this change works in your favour more than it works against you. The finance-specific knowledge can be learned and demonstrated, and the AI capability you may already have from another sector is now an explicitly valued credential rather than an incidental one. Building AI skills around a full-time job in ninety days is a realistic route for exactly that situation, and there are free and low-cost options employers trust if a paid programme is out of reach.

Which Proof of AI Skill Actually Counts?

Not a certificate on its own. What UBS has described asking for is evidence of application, which is a different artefact from a completion badge. The strongest version is a short, concrete account of a real task: what you were trying to produce, which tool you used and why, what it got wrong, and how you verified the result before anyone relied on it. That structure works whether the task came from an internship, a university project or your own research, and it survives follow-up questions in a way that a course name does not.

Internal certifications matter more once you are inside. UBS runs its own, and the same is increasingly true elsewhere, which means the badge that gets you hired and the badge that gets you promoted may be different things. Before paying for an external programme, it is worth checking whether the employers you are targeting recognise it at all, because plenty of paid courses are not the certifications employers actually recognise. Many of the best-regarded internal programmes are only available after you join.

The last piece is the least technical and the most durable. Every AI-fluent candidate can produce output quickly. Far fewer can explain what they would not delegate to a model, and why. In a first-year banking role where a wrong number reaches a client, that judgment is the whole job. Practise articulating it, because it is the answer that a bank screening for AI literacy is actually listening for.