The Junior Skills Problem AI Didn’t Mean to Create
AI has rapidly automated many of the entry‑level tasks that once shaped early careers in banking technology, and junior engineering pathways are quietly being compressed. Short‑term efficiency gains now risk creating a long‑term capability gap, and financial organisations must rethink how they develop the next generation of technical specialists in an AI‑augmented world.
Ben Dowdle
Senior Principal Consultant
ben.dowdle@caspianone.co.uk
AI adoption in financial technology is accelerating quickly, but one of its most significant side effects is discussed far less often. As AI tools become embedded across development, testing, data analysis, and operational workflows, the impact on junior specialists in AI banking is becoming increasingly difficult to ignore.
I speak regularly with senior leaders, engineers, and tech managers across investment banks. Over the past year, a consistent pattern has emerged. Entry‑level technical responsibilities are being automated faster than most organisations expected. In doing so, many of the traditional junior pathways that once defined early careers in banking technology are being quietly compressed. AI was never designed to remove the need for junior engineers, but in practice, it is reshaping where, how, and even whether those early opportunities exist.
This is not an argument against AI. The gains in efficiency and capability are real, and for many teams long overdue. But as banks move further along the AI adoption curve, it is worth examining which work is being automated first, why junior resourcing is slowing, and what this shift means for long‑term capability across financial institutions.
Is junior work disappearing before anyone notices?
Historically, junior engineers in banks learned through exposure to repetition and scale. Early careers were shaped by clearly scoped, lower‑risk tasks that nevertheless sat close to real systems and real consequences. Over time, those experiences built intuition, domain knowledge, and confidence. Today, many of those same responsibilities align almost perfectly with what modern AI tools do best.
Across development and data teams, AI is increasingly responsible for generating boilerplate code, wiring interfaces, drafting initial test cases, supporting basic quality assurance, analysing logs, assisting first‑line incident triage, cleaning and validating data sets, and producing early versions of documentation or internal reports. None of this work was glamorous, but all of it was formative. It is precisely where junior engineers learned how systems behave under stress, how data degrades in real‑world conditions, and how small technical decisions compound into operational risk.
The outcome is an uneven redistribution of value. Senior engineers immediately benefit as AI removes large volumes of repetitive work from their day. Productivity rises, delivery speeds improve, and teams appear more efficient on paper. For junior engineers, however, the learning surface shrinks. The work that once built foundational competence is now abstracted behind automation.
Why banks are seeking fewer entry‑level engineers
At a leadership level, this shift is often framed as pragmatic. Cost pressures across banking technology organisations are intense, particularly in high‑expense centres such as London and New York. AI initiatives promise measurable efficiency gains and slot neatly into broader narratives around consolidation, productivity, and return on technology investment.
In many banks, AI is positioned as a way to do more with fewer people. Resource allocation is tightly scrutinised, backfilling is slower, and graduate skillset acquisition is sometimes treated as discretionary rather than foundational. When AI absorbs a significant share of junior‑level work, the short‑term business case for maintaining large entry‑level intake programmes appears weaker.
Industry data reinforces this direction of travel. Research cited by McKinsey indicates that nearly 30 per cent of organisations expect to rely on fewer junior‑level employees as they scale AI and automation, particularly in technical and knowledge‑intensive functions. Banks are not always explicitly deciding to remove junior positions, but many are slowing entry‑level opportunities because those requirements no longer appear immediately productive in an AI‑augmented environment.
In parallel, the bar continues to move upward. Organisations increasingly look for engineers who can contribute value earlier, often blending technical skills with domain understanding, communication ability, and confidence operating alongside AI‑driven tooling.
The longer‑term capability risk no one is modelling properly
The most significant risk is not what happens this year or next. It is what happens several years from now if this pattern continues unchecked. Senior engineers do not materialise fully formed. They develop through exposure to incidents, failures, trade‑offs, regulatory constraints, and performance bottlenecks. Those experiences accumulate gradually over time. If junior engineers are systematically removed from environments where those learning moments occur, banks risk hollowing out the future of their own SME pipeline.
AI improves output today, but it reduces the number of people building foundational experience. Over time, that dynamic creates fragility at exactly the point where AI is least effective without strong human oversight. Banking systems remain complex, regulated, and deeply contextual. Accountability still resides with people, not tools.
This concern is increasingly reflected in analyst commentary. Gartner has noted that productivity gains from generative AI are currently most pronounced among senior engineers, while junior specialists face greater displacement risk unless organisations actively redesign learning and progression models. In other words, efficiency gains in the short term can translate into capability shortages in the medium term.
Why optimism about ‘new AI positions’ only solves part of the problem
It is often suggested that AI will simply create different opportunities for early‑career engineers. Conceptually, this is true. Practically, the transition is uneven. Opportunities associated with AI operations, model supervision, data observability, or prompt design often assume a baseline level of technical maturity. These positions require engineers who understand systems well enough to challenge, validate, and contextualise automated outputs. They are not, in most cases, genuine entry points.
This is where junior specialists in banking face their most structural challenge. The ladder into these new positions has fewer rungs than the ladder they replaced. Without deliberate intervention, banks risk creating environments where the traditional entry point disappears faster than alternative pathways mature.
How early‑career engineers can add value earlier
For junior engineers and graduates, the nature of value is shifting. Competing with AI on speed or volume is no longer viable. What differentiates early‑career expertise now is the ability to develop judgement, context, and accountability alongside automation.
Those progressing fastest are typically investing early in understanding the business problems behind the technology, whether that means how markets operate, how risk is managed, or how regulatory constraints shape system design. They are also able to articulate why decisions matter, not just how solutions are implemented.
AI fluency is rapidly becoming a baseline expectation rather than a differentiator. What matters more is the ability to interrogate outputs, recognise edge cases, and understand downstream consequences. In an AI‑augmented environment, accountability becomes the defining skill.
What we need to relearn about developing expertise
If we want durable engineering capability, AI cannot be treated purely as a cost‑reduction technology. Professional development now requires more intention, not less. Some institutions are beginning to respond by reshaping early‑career exposure rather than abandoning it. This includes involving juniors in production environments earlier, embedding them in AI validation and oversight work, pairing them closely with experienced engineers who model decision‑making rather than just delivery, and creating explicit learning mechanisms around failure rather than hiding it behind automation.
Crucially, this is not about preserving legacy structures for their own sake. It is about recognising that capability does not emerge by accident. Banking technology remains complex, regulated, and accountability‑heavy. AI may change how work gets done, but it does not remove the need for engineers who understand systems deeply enough to challenge outputs, manage risk, and take responsibility when things go wrong.
McKinsey’s research into AI transformation consistently shows that organisations extract the most value when AI adoption is paired with workforce redesign, rather than treated as a standalone efficiency tool. The real choice facing banks is not whether to adopt AI, that decision has already been made. The more consequential decision is whether they continue to invest in how engineers are formed, or whether they allow short‑term efficiency gains to quietly erode the foundations of long‑term capability.
The responsibility AI never asked for
AI did not intend to create a junior engineering gap in banking. But intent is less important than outcome. If we want sustainable technology organisations, we must continue to invest in how people learn, not only in how systems perform. Someone still needs to understand the foundations, manage ambiguity, and grow into the next generation of technical leaders.
For early‑career engineers, the bar has moved, but the opportunity remains. Combining AI fluency with deep contextual understanding will remain valuable, albeit hard to come by. The future of junior specialists in AI banking will look very different from the past. The real question is whether the industry chooses to shape that future deliberately, or whether it simply absorbs the consequences. .
Disclaimer: This article is based on publicly available, AI-assisted research and Caspian One’s market expertise as of the time of writing; written by humans. It is intended for informational purposes only and should not be considered formal advice or specific recommendations. Readers should independently verify information and seek appropriate professional guidance before making strategic hiring decisions. Caspian One accepts no liability for actions taken based on this content. © Caspian One, 2026. All rights reserved.
Read more or search for topics that matter most to you!
