NatWest AI Job Cuts Explained: 1,000 Roles Gone, New AI Roles Created
Key Takeaways
- NatWest Group cut about 1,000 full-time jobs in the third quarter of 2026, bringing total reductions so far in 2026 to about 1,400.
- The bank reported around £420 million in cost savings and an operating pre-tax profit of £2.2 billion for the quarter, in results published on 30 October 2026.
- Chief executive Paul Thwaite says roles are changing rather than simply disappearing: new AI-dedicated roles are being created while other roles are reshaped by AI.
- NatWest’s fraud-triage agent inside its Cora digital assistant has now handled 52,000 credit-card conversations, up from nearly 19,000 when it was announced in September.
- A Spending Insights trial lets customers discuss spending by voice or text, built on NatWest’s own in-house small language models.
- Overall headcount actually rose, because the Evelyn Partners acquisition added around 2,200 staff — so the job cuts and the headcount rise are both true at once.
- The cuts are part of a wider UK banking trend, with other major lenders also reducing roles as AI automates administrative and customer-service work.
The headline numbers
NatWest Group’s third-quarter results, published on 30 October 2026, contained two stories that sound contradictory until they are placed side by side. The first is a significant workforce reduction: about 1,000 full-time staff left in the quarter, bringing the total number of jobs lost so far in 2026 to about 1,400. Most of this year’s reductions therefore happened in a single three-month period, which is why the quarter’s figures attracted so much attention — this was an acceleration, not a steady trickle.
The second story is financial. The bank reported an operating pre-tax profit of £2.2 billion for the three months to September, about 2% higher than the same period last year, alongside net interest income of around £3.6 billion — the difference between what the bank makes from loans and what it pays out on deposits — up nearly 9% year on year. It also reported cost savings of around £420 million, which helped bring its cost-to-income ratio — the amount it spends running the business as a percentage of what it generates in income — down to 45.7% for the year to date.
One figure in coverage of the results needs a note of care. Some reporting of the same results cited a profit figure of £1.7 billion, down 3.9% year on year. That is a different profit measure from the £2.2 billion operating pre-tax profit reported by The Gazette, reflecting different treatments of items such as tax and one-off charges. Readers comparing headlines should check which measure is being quoted before concluding that the results were either strong or weak. On the operating measure, the quarter was slightly ahead of the same period last year; on the other measure, it was lower. Both can be accurate descriptions of different lines in the accounts.
The efficiency drive behind the savings is explicitly technological. NatWest said it had made progress towards becoming a simpler and more technology-driven bank, and pointed to investment in AI tools as part of how the savings were achieved. That framing — AI as efficiency, AI as customer service, AI as workforce change — is the thread that ties the whole quarter together.
What Thwaite actually said about AI and jobs
Chief executive Paul Thwaite’s comments are more nuanced than the headline “AI takes 1,000 jobs” suggests, and they are worth quoting directly because they set the terms of the bank’s own argument. He said the “underlying shape of the workforce is changing quite a bit”, adding: “It’s important we’ve got the right skills, the right individuals in the right roles.” On AI specifically, he said: “Roles are changing. We’ve got new roles that are very much dedicated on AI, and we’ve got other roles that are being changed because of AI.”
He also pushed back against a narrow reading of AI as a cost-cutting device: AI was “pervasive across the organisation and it is too often framed in quite a narrow way around efficiency”. “The way that we’re using AI is for the benefit of our customers and for the benefit of our colleagues… so from that perspective, it is pretty universal,” he said, adding that there are “big opportunities in respect of AI for the banking sector”.
Thwaite would not be drawn on whether there would be further job cuts, saying that “the first people to hear about that would be our colleagues”. That refusal matters: it means the 1,400 figure is a statement about 2026 so far, not a cap on future reductions, and anyone treating it as the final total is reading more into the results than Thwaite offered.
The honest reading of his position is that NatWest is doing two things simultaneously: removing roles that AI and automation make redundant, and creating roles that AI makes necessary. The net effect on any individual employee depends entirely on which side of that line their job falls, and on whether retraining bridges the gap. Thwaite’s emphasis on skills is an acknowledgement of that — but it is not, on its own, evidence that every displaced employee finds a new role inside the bank.
The AI tools behind the story: Cora’s fraud agent and Spending Insights
The results were not only about job numbers. They also gave an update on two AI tools that show what NatWest’s AI transformation looks like in practice, from the customer’s side.
The first is the fraud-triage agent inside Cora, NatWest’s existing digital assistant. Customers worried about a suspicious credit-card transaction can explain what happened in their own words, and the agent guides them to the most appropriate fraud, scam or dispute team, instead of requiring them to work out which category their problem fits. NatWest says the agent has now handled 52,000 conversations with credit-card customers. That figure is worth putting in context: when NatWest announced the tool on 25 September 2026, it said nearly 19,000 conversations had been handled since launch. The jump to 52,000 by the third-quarter results, roughly five weeks later, shows how quickly usage scaled — though it is important to note what the number counts. It counts conversations handled, not fraud cases resolved, money recovered or confirmed fraud prevented, and NatWest has not published those outcome figures. A routing tool that gets customers to the right team faster is genuinely useful, but it is not the same as a tool that detects fraud on its own or blocks payments, and the bank’s own description presents it as the former.
The second tool is Spending Insights, a trial that lets customers talk through their spending with the bank by voice or text. Customers can ask questions about their spending, budgeting and financial habits in a natural two-way conversation, and the tool can handle follow-up questions, changes of topic and interruptions. It presents transaction data as personalised insights and graphs, aiming to help customers understand their money and identify potential savings. NatWest built the tool on its own small language models, developed in-house by its Chief AI Research Office rather than bought from a third-party provider, and worked with Nvidia to improve speech recognition across regional accents and everyday money language. The trial began with Royal Bank of Scotland customers, and the bank has described it as the first in the industry to test a fully generative, real-time voice-to-voice conversational AI for personal finance — a “first” claim that is NatWest’s own framing and has not been independently verified.
Both tools illustrate Thwaite’s broader point: AI at NatWest is not only about removing staff, but about changing what a banking conversation is. A customer who once navigated menus and waited for an agent can now describe a problem in plain language and be routed, or discuss spending in a flowing conversation rather than staring at a static dashboard. Whether that is better depends on execution — a routing agent that misroutes is worse than a menu — but the direction is clear, and it is the same direction other banks are travelling.
The headcount paradox: cuts up, total staff up
Here is the detail that most headlines miss, and it changes the story. The recent acquisition of wealth manager Evelyn Partners brought around 2,200 staff into the group, driving up its overall headcount. The reported workforce reductions exclude any employment changes associated with that acquisition. So NatWest cut about 1,000 jobs in the quarter and about 1,400 in the year — and still ended the period with more employees than it started with, because it bought a business with 2,200 people in it.
That paradox is not a contradiction or an accounting trick. It is two different strategies running at once: shrinking the traditional banking workforce through efficiency and automation, while expanding the wealth-management workforce through acquisition. NatWest has reiterated its ambition to expand wealth management, reducing its reliance on income sensitive to interest-rate changes, and the £2.7 billion Evelyn Partners acquisition — its largest since its involvement in the ABN Amro takeover before the global financial crisis — is the vehicle for that ambition.
For anyone trying to answer the question “is NatWest shrinking or growing?”, the honest answer is: both, in different parts of the business. The traditional bank is getting leaner and more automated. The group as a whole is getting bigger in wealth management. Treating either half as the whole story misleads.
A wider UK banking trend
NatWest is not an outlier. Its announcement coincided with a broader movement across UK banking, as major lenders turn to AI to improve efficiency and reduce operating costs. Barclays recently outlined plans to cut approximately 5,000 jobs during 2026, with AI expected to play a significant role in reshaping its workforce. HSBC is also reportedly planning substantial reductions in its UK financial advisory operations as it introduces AI-driven services, while Standard Chartered has previously announced thousands of redundancies as part of its restructuring strategy.
The pattern is consistent: customer support, administration and other functions where technology can automate repetitive tasks are the first to change. That does not mean every bank job is automatable — relationship management, complex advice, fraud investigation and regulatory judgement all resist simple automation — but it does mean the centre of gravity in banking employment is shifting from processing towards expertise, oversight and technology itself.
The context around the results adds another pressure. The UK Government announced plans to increase taxation on banking profits, with the bank corporation tax surcharge set to double from 3% to 6% for a temporary 15-month period beginning in January 2027. Thwaite stopped short of directly criticising the measure but stressed the importance of maintaining the proposed time limit, arguing that banks needed sufficient capital to support households and businesses. Higher taxes do not cause AI adoption — the efficiency programme predates the Budget — but they sharpen the incentive to find savings wherever they can be found, and AI is currently the most politically presentable source of savings a bank can name.
What this means for bank workers and customers
For employees, the NatWest story is a case study in how AI-driven change actually arrives: not as a single dramatic replacement event, but as a quarter in which 1,000 roles go, some new AI roles appear, and the chief executive talks about skills and shape rather than headcount. The practical questions for staff are concrete: which roles are being reshaped, what retraining is offered, and whether the new AI-dedicated roles are open to existing employees or filled externally. The results announcement does not answer those questions in detail, and that gap is where the real human impact will be decided.
For customers, the implications are more immediately visible. A fraud-triage agent that has handled 52,000 conversations suggests that more customers are already being served partly by AI, whether they notice it or not. A spending-insights trial suggests that the next generation of banking may feel less like filling in forms and more like having a conversation. The risk, as with any AI deployment, is that convenience comes with opacity: customers may not know when they are talking to an AI, what data it uses, or how to reach a human when the AI gets it wrong. Banks that deploy these tools well will make the human fallback obvious; banks that deploy them badly will discover that an AI that misroutes a fraud victim has done reputational damage no efficiency saving can repair.
For readers interested in how AI is reshaping financial services more broadly, our guide to building fintech apps with AI for banking covers the product side of the same transformation, while our report on Westpac’s AI token migration shows a very different bank using AI to cut costs — in that case, replacing a seven-figure project with a far cheaper AI approach. NatWest’s story belongs in that same family: AI as a cost and capability lever, applied at the scale of a national bank.
The bottom line
NatWest’s third quarter of 2026 is best read as three stories in one. An efficiency story: about 1,000 jobs cut in the quarter, 1,400 in the year, around £420 million saved, and a cost-to-income ratio down to 45.7%. A technology story: a fraud agent handling 52,000 conversations and a voice-and-text spending tool built on in-house small language models. And a strategy story: a bank shrinking its traditional workforce while growing its wealth business by 2,200 people through acquisition.
None of those stories alone explains the quarter. Together, they show a large bank in transition — using AI to do more with fewer people in some areas, hiring specialists in others, and buying its way into a business less exposed to interest rates. Whether that transition ultimately serves employees and customers as well as it serves the cost-to-income ratio is the question the next few quarters will answer, and it is one worth watching closely.
