Barclays is expanding how deeply it uses Anthropic's AI model Claude across the bank, with a specific target: half of its software developers will use the coding tool Claude Code by the end of 2026, and most of them by 2027. For a bank with tens of thousands of technology staff, that is a full change in how software gets built and maintained, not a side experiment.
This is not Barclays' first step with Claude. The bank has already had an internal tool called the Colleague Knowledge Assistant running since 2025, which helps staff find answers while supporting the bank's 20 million UK retail customers. More than 16,000 employees use it, and it has handled over one million searches. Separately, in the bank's trading and markets division, Claude now sorts around 120,000 incoming client emails every day, figuring out what each one is about and sending it to the right team.
The timing is not a coincidence. Anthropic launched the deal on the same day it announced Claude for Financial Services, a new product line built around ten ready-made AI agents for banking tasks: screening new clients, reviewing earnings, building financial models, and closing monthly books. Barclays is the headline example of that launch, which fits a wider pattern. Anthropic has been racing to sign up Wall Street names including JPMorgan, Goldman Sachs, and Citi, and is reportedly preparing for a stock market listing that could value the company near two trillion dollars. Landing a marquee client like Barclays, in one of the most tightly regulated industries, is exactly the kind of proof point that supports that valuation story.
There is a harder edge to this too. Barclays has already been cutting jobs and moving work offshore to India as it leans more on AI tools for everyday tasks. Across the banking industry, firms including Goldman Sachs, JPMorgan, and Citi have been shrinking the size of their junior analyst classes, the entry-level jobs that used to be the training ground for future bankers. AI is not just speeding up existing staff, it is changing how many new staff a bank needs to hire in the first place.
For businesses outside banking, the lesson is less about the technology and more about what it signals. Barclays operates under some of the strictest data and security rules of any industry, and it is still choosing to put AI in front of customer data, trading operations, and its own codebase. That removes a common excuse heard in other regulated or risk averse sectors: insurance, healthcare administration, and logistics firms often say their work is too sensitive for AI tools. A bank managing money for 20 million customers has decided the risk is manageable with the right oversight.
The honest takeaway is that the gap between early adopters and everyone else is widening faster than most non-tech leaders expect. Barclays did not get here overnight, it built up from a single internal tool in 2025 to a bank wide rollout within about a year. Companies that wait for AI tools to feel fully proven risk watching competitors build that same year of practical experience first.