Starling Bank to cut 130 jobs and boost investment in AI to reduce costs

Starling Bank has told staff it will cut 130 jobs while increasing investment in artificial intelligence, a combination that has become a familiar corporate refrain but lands with particular force in a business built on the promise of doing banking differently. The redundancies, representing about 3% of the workforce, are being presented as part of a restructuring across banking and technology operations intended to remove “duplicate” roles and speed up delivery. In the same breath, the lender is signalling that it will keep hiring in the areas it considers foundational to its next phase, specifically technology and AI engineering.

For a digital-only bank that grew by defining itself against the habits and overheads of traditional lenders, the need to simplify its own structure carries an implicit admission: scale creates its own bureaucracy, even in organisations that pride themselves on escaping it. Starling employs more than 4,000 people, and in that context 130 posts is not a slash-and-burn retrenchment. It is, rather, a surgical correction to a model that has matured beyond start-up improvisation and now faces the slower, less glamorous work of ensuring that systems, teams and processes line up with the business it has become, not the one it set out to be.

The bank’s explanation rests on a claim that will resonate with investors and unsettle employees in equal measure. Starling describes “agility” as a competitive advantage over legacy banks, pointing to an ability to “rest, launch, learn and reorganise at pace”. The sentence is revealing. Challenger banks sold the public a vision of sleek apps and fewer fees, but the deeper pitch to the market has always been organisational: fewer layers, quicker decisions, and technology that does not have to be bolted on to decades-old infrastructure. If that is the edge, then internal duplication is not merely a cost problem. It is an identity problem.

The timing is not accidental. Starling has reported a 6% fall in revenue to £887m for the year ending in March, while pre-tax profit slipped 3% to £217m. Those are not catastrophic numbers, but they sharpen the choices available to management. When revenue growth is less certain, the argument for tighter control of costs becomes more urgent; when profit falls, spending that once looked like confident investment can start to look like drift unless it is tied to clear outcomes.

Starling has said the fall in pre-tax profit was partly due to investment in its digital banking software business, Engine. The insistence on “partly” matters. Engine has been positioned as a way to turn Starling’s internal capabilities into an exportable product, a route to diversification that might loosen dependence on the narrow economics of retail banking. Building such a platform is expensive, and it is also the kind of undertaking that can create precisely the overlaps Starling now wants to remove: product teams, operational teams and compliance functions often multiply as new lines of business are added, each with its own demands and reporting. A company can be both investing for the future and discovering, mid-build, that its own scaffolding has grown too heavy.

Starling’s origins are relatively recent but already have the contours of an era. Founded in 2014 by Anne Boden, a former Royal Bank of Scotland executive, it emerged alongside Monzo and Revolut during the mid-2010s wave of “neo-banks” that promised to disrupt the established order. The pitch was part technology and part cultural critique: traditional banks were slow, expensive and often indifferent to customers, while challengers would be transparent, responsive and mobile-first. That narrative helped unlock regulatory support for new entrants and attracted a generation of customers more willing to manage their financial lives on a phone.

The bank now counts 6.2 million customers, the majority in the UK. That scale is substantial enough to make Starling a fixture rather than a novelty, and it brings with it the responsibilities that disruptors sometimes prefer to treat as legacy problems. Banking is not simply software, and the strength of a digital interface does not reduce the seriousness of what sits behind it: controls, governance, and the constant requirement to know who is using the system and why.

It is here that Starling’s story has been bruised. In 2021, the UK’s financial watchdog placed restrictions on the bank after findings about weak financial crime controls, preventing it from opening new accounts for high-risk customers. In 2024, the Financial Conduct Authority escalated the critique, describing Starling’s controls as “shockingly lax” and arguing that the deficiencies left the financial system exposed to criminals and to those under sanctions. The regulator fined the bank £29m. Those episodes do more than generate unpleasant headlines and costly penalties. They shape the way the market interprets any future promise of “moving fast”, and they colour how staff and customers hear management talk about efficiency.

Against that backdrop, Starling’s assertion that it will invest more heavily in AI takes on a dual meaning. In the optimistic framing, AI is a tool for driving down costs and improving productivity. In a more cautious reading, it is also a bet that technology can help enforce the discipline that rapid growth strained: better monitoring, more consistent decision-making, fewer manual checks that depend on overloaded teams, and quicker identification of suspicious patterns. The public conversation about AI often focuses on customer-facing chatbots or sleek personal finance features. For banks, the more consequential uses may be less visible: automation in compliance, controls, reporting and internal audit, areas where errors are expensive and regulators have long memories.

Starling has been explicit that it will continue hiring tech and AI engineers while changing the structure of parts of its banking team. That is a delicate message to deliver, because it implies a reallocation of opportunity from one group of employees to another, not simply an overall tightening. It also suggests a view about where competitive advantage will sit. If Starling believes it can use AI to handle tasks that were previously labour-intensive, it must be confident not only in the tools but in the quality of the data and the governance around those systems. A bank that has already been criticised for lax controls will need to show that automation is strengthening oversight rather than thinning it out.

There is also the human question of what “duplicate” means inside a bank that spans retail accounts, business banking, payments infrastructure and now a software platform. Duplication can arise from inefficient structures, but it can also be a by-product of sensible caution: parallel checks, separate oversight and a degree of redundancy can be part of risk management. Removing overlaps is easy to describe as simplification; it is harder to execute without creating gaps that only become evident when a process fails or a regulator asks uncomfortable questions.

Yet the financial pressures cannot be waved away. Revenue falling 6% in a year is a warning sign in any growth-focused business, particularly one operating in a fiercely competitive market where customers can switch accounts with relative ease. Starling’s peers, including Monzo and Revolut, have jostled for the same audiences and expanded into overlapping products, while incumbent banks have invested heavily in their own apps and digital channels. The first era of challenger banking benefited from the contrast with clunky legacy systems. The next era is more brutal: the baseline of customer experience has improved across the industry, which means differentiation is harder and profitability matters more.

Starling has also encountered limits to its international ambitions. Like several fintech peers, it has struggled to expand abroad and, in 2022, abandoned an attempt to secure a European banking licence. That decision is a reminder that banking remains national in ways that technology companies are not. Licences, regulation, consumer protections and political scrutiny vary across borders, and the cost of compliance multiplies quickly. For a bank already managing the demands of UK regulators, the appeal of spreading into multiple jurisdictions can fade when weighed against the risk, expense and potential distraction.

The spectre of a stock market listing hovers over the current restructuring. Starling has long been linked to flotation speculation, and in January its chief executive, Raman Bhatia, told the Sunday Times that while there were no firm plans, he could see the business as a plc in a “near-term window”. An IPO is not just a fundraising event. It is a change of rhythm: quarterly scrutiny, sharper benchmarks and less tolerance for narratives that are not backed by hard numbers. If Starling is indeed moving closer to a public listing, then reducing duplication and demonstrating cost discipline are not simply operational choices; they are signals to future investors that the bank is ready to be judged like an established institution rather than a perpetual disruptor.

In that light, the decision to push further into AI can be read as part of the same preparation. Public markets reward efficiency when growth is uncertain, and they often prefer technology-led stories when those stories are tied to measurable improvements. But they also punish governance lapses, and the FCA’s language in 2024 was unusually severe. Any narrative about moving faster through AI will have to coexist with credible assurances that the bank’s controls are not merely adequate, but robust enough to withstand the harsher scrutiny that comes with size, ambition and, potentially, life as a listed company.

For employees, the consultation period will bring the usual mixture of anxiety and calculation: whether roles can be redeployed, whether skills can be repurposed, and whether the new structure creates clearer paths for those who remain. For customers, the changes may be invisible unless they affect service quality, product innovation or trust. The more intriguing question is how Starling balances its self-image as a nimble challenger with the realities of being a sizeable financial institution that has already been warned, publicly and expensively, about the consequences of moving too quickly without adequate controls.

The bank’s leadership is asking to be judged on its ability to reorganise at pace. That claim will now be tested in the most consequential way: by whether Starling can cut roles without weakening its foundations, invest in AI without repeating past mistakes, and translate a simplifying restructure into a business that can defend its margins in a market where digital banking is no longer a novelty but an expectation.