Barclays scales Claude to upgrade operations and improve client experience
Anthropic Anthropic
Barclays is rolling out Claude across its bank to speed up coding and customer support. It wants most engineers using it by 2027, which is a big test for AI inside a heavily regulated bank.
Based on reporting by Anthropic, Anthropic — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Barclays is deepening its work with Anthropic and pushing Claude further into day-to-day banking operations. The British universal bank says the goal is to speed up software development, modernize older systems, and make operations run more smoothly across its global business.
The biggest numbers are in engineering. Barclays expects Claude Code to reach half of its developer population by the end of 2026, then a majority of software engineers in 2027. That is a meaningful bet for a bank that has to balance speed with control. It is not just using AI for experiments or side projects; it is aiming to embed it into how software gets built and maintained.
One place where the partnership is already visible is the Colleague Knowledge Assistant for Barclays UK employees. The tool helps staff find information faster when they are supporting more than 20 million UK retail customers. It has been live since 2025, has been adopted by more than 16,000 colleagues, and has handled over one million searches. The promise is simple: quicker access to answers, faster customer support.
Claude is also helping in Global Markets, where Barclays uses the models to sort and enrich incoming emails and decide where they should go next. The platform handles about 120,000 emails a day, cutting down manual work while helping colleagues find the right information sooner. That is the kind of unglamorous automation banks actually care about.
Barclays keeps stressing the guardrails. The bank says deployment is backed by governance, security controls, and human oversight, and Anthropic is leaning hard on that message too. The pitch here is not just that AI can help a bank move faster. It is that a large, regulated institution can adopt it without pretending the risks do not exist.
My take — AI-written commentary, not fact-checked reporting
This is the right way to sell AI in banking: show the boring workflows, not the demo theater. Half the industry still treats model choice like a religion, but the real contest is whether a tool can survive compliance, oversight, and actual employees using it all day. Closed models will keep winning these deals until open systems can offer the same control without the hand-holding.
Read more about this at: Anthropic
