Barclays’ headquarters at One Churchill Place, Canary Wharf, London, in 2005. Image: Bajajvikram / Wikimedia Commons, Public domain, cropped
Barclays is rolling Anthropic’s Claude out across the bank, and expects half of its software developers to be using Claude Code by the end of 2026, Anthropic announced on Thursday. By 2027, most of the bank’s engineers should have it.
The British bank says it will use Claude to speed up software development, modernise its ageing legacy systems and cut the time staff spend on routine work. Anthropic called it “a significant milestone for us in the UK”.
What Claude already does at Barclays
The expansion builds on tools that are already running. A Colleague Knowledge Assistant, live since 2025 and powered by Claude, helps Barclays UK staff find answers when they serve the bank’s more than 20 million UK retail customers. More than 16,000 employees have adopted it, and it has handled over one million searches, according to the announcement.
In Barclays’ Global Markets business, Claude sorts incoming client emails: it classifies each one, adds missing details and decides where it should go, so operations staff can act on it. The system handles about 120,000 emails a day. These usage figures come from the two companies and haven’t been independently verified.
Craig Bright, one of Barclays’ two group chief operating officers, said the bank is heading towards AI as “an increasingly agentic capability embedded within how we build, test, secure, and operate technology”:
Software engineering and cyber security are both being reshaped by increasingly capable AI systems.
Craig Bright, Group Co-Chief Operating Officer, Barclays
His co-COO, Anne Marie Darling, said the goal was “a simpler, more efficient organization”, with staff freed from routine tasks to focus on harder problems. Neither company said how many jobs the changes might affect.
A day after the Bank of England’s warning
The timing stands out. On Wednesday, the Bank of England’s Financial Policy Committee warned about the risks AI poses to the financial system, from a sharp fall in AI stocks to incidents in tests of frontier models and the cyber threat from powerful AI tools. Governor Andrew Bailey repeated several of those concerns in a BBC interview on Thursday morning.
Barclays says “robust governance, security controls, and human oversight” sit around every AI use, and Anthropic’s chief commercial officer, Paul Smith, said the bank was “rolling Claude out with strong security and oversight”. Banks are among the most tightly supervised businesses in the world, so how Barclays fares will be watched by regulators and rivals alike.
The deal also adds a big, well-known customer to Anthropic’s list as it prepares for what could be the largest stock market debut ever, where enterprise demand for tools like Claude Code is central to its pitch.
Why it matters
When one of Europe’s biggest banks plans to put an AI coding agent in front of most of its engineers within a year, AI in banking is no longer a pilot project. It also puts Anthropic’s safety claims to a practical test in one of the most regulated industries there is, at the moment regulators are starting to sound the alarm.
Sources: Anthropic; Bank of England.
Latest Anthropic news
