i
News
News · 2026-10-01

Barclays expands Claude across customer service and engineering

@neuronium_ai @neuronium_ai

Barclays is expanding its use of Anthropic’s Claude across customer service, markets operations and software engineering. The bank says the tools are already helping employees find answers for customers and process incoming emails; it also plans to make Claude available to most of its engineers by 2027. The announcement is less about a new model than about putting AI into the workflows of a large, regulated bank—and whether those deployments can make service faster without weakening oversight.

Cover: Barclays expands Claude across customer service and engineering

Where Claude is already in use

Barclays UK’s Colleague Knowledge Assistant helps staff find information while serving the bank’s more than 20 million retail customers in the UK. It has been running since 2025, using Claude with retrieval-augmented generation. More than 16,000 employees have used it, and it has handled over one million queries.

In Barclays Global Markets, Claude models help process about 120,000 incoming emails a day. They classify and enrich messages, route them to the appropriate teams, help set priorities and check whether requests contain enough information for operations staff to act.

16,000employees
1 millionqueries
120,000emails daily

The bank also says Claude is helping modernise legacy platforms and improve software quality. Barclays is moving towards using AI agents to build, test, secure and operate technology, with engineers and cybersecurity specialists focused on more complex work.

The harder test is the workflow

Barclays presents the expansion as a way to reduce routine work, speed up decisions and give staff more time for difficult customer problems. Its group operating chief Anne-Marie Darling says the value of the technology should be judged by its effect on customers and employees, while keeping the bank secure and properly governed.

Craig Bright, the bank’s other group operating chief, says AI is changing both software development and cybersecurity. Anthropic’s chief commercial officer, Paul Smith, says Claude will be available to most Barclays engineers by 2027, and describes the partnership as an important step for the company in the UK.

Barclays says its AI use is subject to strict governance procedures, security measures and human oversight. Those safeguards matter, but the announcement gives little detail on how the bank measures whether faster handling translates into better customer outcomes, or how it checks that automated email classification and routing work reliably.

I think the most telling figures are the usage numbers, not the broad claims about transforming banking: 16,000 employees, more than a million queries and 120,000 emails a day show tools embedded in real operations. The next test is whether wider access for engineers and more automated work can deliver measurable improvements while keeping humans accountable for the decisions that matter.

Daily AI news

Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.

Only what matters — every day

Follow on X