NatWest Group chief executive Paul Thwaite has warned that artificial intelligence will take over some existing banking work, adding weight to the debate over how financial services companies redesign roles as automation enters core operations.
Speaking at The Times CEO Summit, Thwaite said the shape of NatWest’s workforce “is definitely going to change” as the bank deploys AI. The group employs around 60,000 people, with more than a quarter of its staff now working as software engineers, according to remarks from the summit.
Asked whether NatWest would have fewer employees in 10 years, Thwaite said: “The honest answer is I don’t know.” He added that some roles that currently exist would, “to all intents and purposes,” be delivered by AI.
One of Britain’s largest banks is now openly linking AI adoption to workforce redesign. Banking is particularly exposed to the technology because it combines large customer service operations, complex compliance demands, document-heavy processes, fraud monitoring, credit assessment, data analysis, software engineering, and increasingly digital customer journeys.
NatWest has already described AI, data, and digital capability as part of its technology agenda. Its workforce now includes roles such as AI ethics and AI agent orchestration, according to Thwaite’s summit remarks, showing how AI adoption can create new control and oversight functions while automating parts of existing work.
The tension for banks is that productivity gains and job redesign arrive together. AI can support customer queries, summarise information, identify anomalies, generate code, assist relationship managers, monitor risk signals, and accelerate operational workflows. Poorly controlled models, however, can create conduct, governance, data, and accountability risks.
The board-level control challenge was explored in directors put AI governance on agenda, after Institute of Directors polling showed AI entering governance planning. NatWest’s comments bring the same question into large-employer workforce strategy.
Financial services companies have spent years reducing branch dependency, digitising customer journeys, consolidating back office processes, and investing in technology platforms. AI accelerates that pattern because it affects white-collar work previously protected by complexity, judgement, or regulated process.
Workforce planning will not be as simple as counting roles removed. Banks may need fewer people in some process-heavy areas while hiring more in software, model risk, data governance, cyber security, product, ethics, and customer journey design. The distribution of work changes as much as the headcount requirement.
Middle managers will carry much of the transition. They will need to understand where AI changes process quality, customer handling, employee supervision, and risk accountability. A poorly explained shift could undermine trust internally, while a poorly controlled deployment could expose the bank externally.
Training is one of the harder constraints. A bank can deploy AI tools faster than it can redesign every role, job description, control process, and career pathway. Employees using AI informally may find productivity gains before formal policy catches up, while employees whose work is being automated will need credible routes into adjacent roles.
Reputation also weighs heavily in banking. Errors in customer communication, fraud handling, lending decisions, or complaints resolution can quickly become conduct risk. AI systems may increase speed, but regulated businesses still need auditability, explainability, and human accountability for outcomes.
Thwaite’s comments also sit against a wider growth debate. Banks are central to business investment, household confidence, payments infrastructure, savings, and lending conditions. If AI enables lower cost and better service, it could strengthen competitiveness. If adoption is treated mainly as labour substitution, employee uncertainty and public concern over service quality will rise.
The more durable model is likely to be AI-enabled banking work rather than AI-only banking. Relationship management, complex lending, vulnerable-customer support, dispute resolution, and judgement-heavy decisions still depend on trust, context, and accountability. The question is which tasks are automated, which roles are rebuilt around AI supervision, and whether the skills base can be developed quickly enough.
NatWest’s remarks make the issue concrete. AI is no longer a broad future of work theme inside banking; it is becoming a live operating model question for one of the UK’s largest employers.
You must be logged in to post a comment.