Summary

• Starling is cutting around 130 roles as it restructures banking and technology operations.
• The bank’s latest annual results show £887 million revenue, £217 million pre-tax profit, and £70 million committed ARR for Engine.
• The move connects AI adoption to workforce redesign, cost control, product delivery, and regulated financial services operations.

Starling Bank is cutting around 130 roles as the UK digital bank restructures parts of its banking and technology operations while continuing to invest in artificial intelligence.

The reductions amount to roughly 3% of Starling’s workforce and follow a period in which the bank has been simplifying operations, removing duplicated roles, and trying to accelerate product delivery. The bank is still hiring technology and AI engineers, which makes the restructuring a shift in the shape of work rather than a straightforward pullback from investment.

Starling’s latest annual results give the move its commercial setting. The bank reported £887 million in revenue for the year to 31 March 2026, down from £940 million, and profit before tax of £217 million, compared with £223 million the previous year. Customer platform accounts rose to 6.2 million, while deposits increased to £12.7 billion.

Engine by Starling, the group’s banking software arm, is now central to the growth plan. Starling said Engine had £70 million of committed annual recurring revenue and had opened offices in New York, Toronto, Dubai, and London. Engine allows the bank to sell its proprietary platform to other financial institutions, offering a route to international growth that does not require Starling to become a licensed retail bank in every market.

AI runs across both sides of that model. Starling has launched Starling Assistant, an agentic AI financial assistant built using Google Gemini and Google Cloud technologies, while also using AI in areas such as spending insights and scam detection. The bank has framed these tools around customer control and fraud protection, but the operational question is how automation changes service delivery, product work, and internal staffing.

Banks are not ordinary software companies. AI tools used in financial services must sit within data protection, consumer duty, financial crime controls, accessibility, resilience, and human review requirements. Starling’s own AI privacy materials state that customers can request human review where automated decision making has a legal or similarly significant effect, showing how automation has to be built into regulatory processes rather than bolted on later.

The workforce consequences are therefore more complex than a headline redundancy number. Automation can reduce routine work and duplicated roles while increasing demand for AI engineering, data governance, product risk, model oversight, cyber resilience, and compliance skills. The bank that emerges from AI adoption may employ fewer people in some functions but require deeper technical and operational capability in others.

Market pressure gives the restructuring a sharper edge. Falling interest rates have weakened income tailwinds for banks that benefited from higher rate environments, while UK fintechs face more scrutiny over profitability, controls, and credible international expansion. Starling’s previous regulatory problems around financial crime controls also mean it cannot treat automation as a substitute for robust oversight.

The bank’s task is to show that AI can make operations more useful and resilient as well as cheaper. Faster product delivery and lower costs will appeal to investors, but regulated banking punishes weak controls and poor customer outcomes. Automated support, fraud detection, and financial assistants need to perform reliably when customers are confused, vulnerable, or facing losses.

Starling’s restructuring is therefore part of a broader change in digital banking. The early challenger bank contest was about app design, customer acquisition, and escaping branch networks. The next phase is more operational: which banks can turn software, AI, compliance, and platform licensing into durable economics without weakening trust.