Barclays scales Claude across the bank, aiming Claude Code at half its developers

Anthropic said Thursday that Barclays is expanding its strategic collaboration to roll Claude across global operations, including a plan for Claude Code to reach 50% of Barclays developers by the end of 2026 and a majority of software engineers in 2027.

A major UK universal bank saying half its developers will be on Claude Code by the end of the year is a concrete adoption figure from inside regulated banking work. The same release ties Claude to live UK retail support, more than 16,000 colleagues, and to Global Markets email routing of about 120,000 messages a day.

On Thursday, 1 October 2026, Anthropic said Barclays is expanding a strategic collaboration so Claude can run across the bank’s global operations. The newsroom page is titled “Barclays scales Claude to upgrade operations and improve client experience.” The visible date is October 1, 2026. The page does not print an hour. Anthropic calls Barclays a British universal bank. A universal bank takes deposits and makes loans for households, and it also runs markets and investment-banking work for companies and investors. The page says the bank is integrating secure, enterprise-grade AI systems. Enterprise-grade, here, is Anthropic’s phrase for software built to a large company’s security and control rules. Those lines are Anthropic’s.

What the expansion covers, as the page states it. Barclays is extending Claude across the bank to speed software development, modernize legacy systems, and improve operational efficiency. A legacy system is an older platform the bank already runs. As part of that rollout, Barclays expects Claude Code adoption to reach 50 percent of its developer population by the end of 2026, rising to a majority of software engineers in 2027. Fifty percent is half of that developer population. A majority is more than half. The page uses “developer population” for the 2026 figure and “software engineers” for the 2027 figure. It does not print how many developers Barclays employs, so a reader cannot turn 50 percent into a headcount. Claude Code is the Anthropic product named in the target: the tool a developer uses to write and change software with Claude. The page names the product and the two targets. It does not say half the developers are on the tool today. Those lines are Anthropic’s, reporting what Barclays expects.

Anne Marie Darling, group co-chief operating officer at Barclays, is quoted on the page. “The true measure of any technology is the impact it has on customers, clients, and colleagues, and the outcomes it enables,” she said. “Expanding the use of Claude across Barclays is about more than adopting AI. It is about transforming how work gets done: reimagining processes, enhancing customer and client experiences, and building a simpler, more efficient organization within a secure and well-governed environment. By reducing time spent on routine activities and streamlining decision-making, we can enable colleagues to focus their expertise where it matters most, solving more complex problems, accelerating innovation, and delivering better outcomes for our customers, clients, and business.” That quotation is hers, in the Anthropic release. The page gives her the title group co-chief operating officer. A second person on the same page holds the same title.

Craig Bright, also group co-chief operating officer at Barclays, is quoted beside her. “Software engineering and cyber security are both being reshaped by increasingly capable AI systems,” he said. “We’re moving towards AI as an increasingly agentic capability embedded within how we build, test, secure, and operate technology. At Barclays, we’re using Claude to help modernize legacy platforms, improve software quality, and allow our technical experts to focus on the most complex challenges. The combination of skilled engineers, cyber specialists, and trusted AI capabilities has the potential to fundamentally improve how technology is developed and managed across large-scale enterprises.” Agentic, in his sentence, means the software can take a next step in the work — build, test, secure, operate — rather than only suggest the next line. “Has the potential” is his condition. It is not a measured result the page already has in hand. That quotation is his. The page spells it “cyber security,” as two words.

One system the page says is already in use is Barclays’ Colleague Knowledge Assistant. It helps Barclays UK employees find information and answers more quickly when they are supporting the bank’s more than 20 million UK retail customers. More than 20 million is the page’s count of those retail customers. It is not a count of people using the assistant. The page says the assistant has been live since 2025, and it does not print the month. It is powered by Claude through a retrieval-augmented generation architecture. Retrieval-augmented generation, often shortened to RAG, pulls passages from the bank’s own documents into an answer, instead of relying only on what the model already stored. A model, here, is the AI system that reads the question and writes the reply. More than 16,000 colleagues have adopted the assistant, and it has handled over one million searches. The page says the result is faster access to information and quicker support for customers. Those adoption figures are Anthropic’s, in this release. The page does not print how many minutes a customer call got shorter.

Inside Barclays’ Global Markets business, the page says Claude models help streamline how client inquiries are processed. Teams use the models to classify an incoming email, enrich it, and choose the processing route. Classify means sort the message by what kind of request it is. Enrich means add the context a colleague needs. The route is which desk, or which next step, should handle it. The page says the point is to prioritize requests and to make sure they contain the information operations colleagues need before they act. The platform processes approximately 120,000 emails each day. Approximately means about 120,000, not a count to the last message. The page says that volume cuts manual handling and helps colleagues find the relevant information and act on it sooner. It does not say a person no longer reads the mail, and it does not say every markets email goes through Claude.

How the page says the bank keeps control. As Barclays expands its use of AI, responsible deployment stays central to the partnership. The bank applies robust governance, security controls, and human oversight, so the technology is deployed safely and effectively across the organization. Governance, here, is the set of rules for which system runs, what it may see, and who can check the result later. Human oversight means a person stays in a position to review the work. Those lines are Anthropic’s.

Paul Smith, chief commercial officer at Anthropic, is quoted on the same page. “Our work with Barclays is a significant milestone for us in the UK. Barclays serves millions of customers and clients across the globe, so their standard for AI is high. That’s why Barclays is rolling Claude out with strong security and oversight,” he said. “Claude now helps 16,000 Barclays’ colleagues find answers for customers, sorts 120,000 emails a day, and will be in the hands of most Barclays engineers by 2027. Very few institutions hold AI to such a high standard.” That quotation is his. His “16,000” sits next to the newsroom sentence that says more than 16,000 colleagues have adopted the assistant. His “sorts 120,000 emails a day” sits next to the newsroom sentence that says the platform processes approximately 120,000 emails each day. His “most Barclays engineers by 2027” is the quotation’s wording for the newsroom’s “majority of software engineers in 2027.”

The closing section is Anthropic’s account of what the wider collaboration is for. The expanded partnership, the page says, reflects Barclays’ long-term commitment to using AI to improve productivity, strengthen resilience, and improve outcomes for customers and clients. It says the two organizations are combining Anthropic’s frontier AI with Barclays’ focus on responsible innovation, to scale practical applications across one of the world’s largest financial institutions. “One of the world’s largest” is the page’s description of Barclays. Frontier AI, here, is Anthropic’s phrase for its most capable models. The page does not print a dollar value for the collaboration, an equity stake, or a line that the arrangement is exclusive. It does not say Claude replaces every system Barclays runs.

The picture is a cream-and-navy graphic. The cream side reads Claude × Barclays, with the line “Anthropic enterprise scale.” Two white cards carry figures from the release: 16,000+ UK colleagues on the Claude assistant, and 120,000 Global Markets emails a day. A green rule splits the frame. The navy side reads Claude Code, developer adoption, and 50% of Barclays developers by the end of 2026, with a line that a majority of engineers is targeted for 2027. The graphic states those figures. It does not print a calendar date.

In plain terms, Anthropic said on Thursday that Barclays is widening a collaboration so Claude is used across the bank, for software, for older systems, and for day-to-day operations. The targets on the page are Claude Code for half of Barclays’ developer population by the end of 2026, and for a majority of software engineers in 2027. Already described as live: a UK colleague assistant, powered by Claude, adopted by more than 16,000 people, with more than one million searches since 2025, used while those colleagues support more than 20 million UK retail customers; and Global Markets email handling of about 120,000 messages a day. Darling and Bright, both group co-chief operating officers, are quoted on customers and on how the bank builds and secures software. Smith, Anthropic’s chief commercial officer, calls the work a significant UK milestone and points to security and oversight. The page does not print a price.