Barclays expands Claude across customer support, market operations and software development, turning Anthropic’s models into production systems inside one of Britain’s largest banks. The October 1 expansion includes tools already used by 16,000 employees and an email workflow processing about 120,000 messages each day.

 

The figures make this more than a pilot announcement. Barclays is reporting active use in high-volume, regulated work while setting a target for Claude Code to reach half of its developers by the end of 2026 and a majority of software engineers during 2027.

 

The Barclays Claude expansion covers three operating areas:

  • A knowledge assistant for UK retail-banking staff
  • Email classification and routing in Global Markets
  • Claude Code for software and legacy-system work

 

Related Coverage

 

Barclays Expands Claude Beyond a Bank Pilot

Anthropic said Barclays is extending Claude across its global operations to accelerate software development, modernize older systems and reduce manual work. The companies described the program as an expansion of an existing strategic collaboration, not a first trial.

 

The most established deployment is Barclays’ Colleague Knowledge Assistant. Live since 2025, the retrieval-augmented system helps staff find information while serving more than 20 million UK retail customers. More than 16,000 employees have adopted it, according to the announcement, and it has handled over one million searches.

 

Those numbers indicate breadth, but they do not establish accuracy, customer satisfaction or labor savings. Barclays and Anthropic did not publish error rates, response-time comparisons or an independent evaluation of the assistant. The available evidence shows scale of use rather than a complete measure of business performance.

 

Claude Routes 120,000 Global Markets Emails a Day

In Barclays’ Global Markets business, Claude models help process incoming client enquiries. The system classifies and enriches messages, checks whether requests contain information needed for handling, and identifies the most appropriate processing route for operations teams.

 

The platform processes approximately 120,000 emails per day. That volume is notable because routing is an operational task with immediate downstream consequences: a useful system must identify priority, context and destination without obscuring responsibility for the final action.

 

Barclays says the workflow reduces manual handling and helps employees act on relevant information more efficiently. The bank also says it applies governance, security controls and human oversight to its AI deployments, although the announcement does not specify review rates, accuracy thresholds or escalation procedures for the email system.

 

Claude Code Targets Half of Barclays Developers

Software engineering is the next major expansion point. Barclays expects Claude Code to reach 50% of its developer population by the end of 2026, followed by a majority of software engineers in 2027. Bloomberg independently reported the target and the daily email volume.

 

The bank plans to use Claude for development, testing, security work and modernization of legacy platforms. Large financial institutions carry decades of software and tightly connected systems, so coding assistants face a harder standard than producing isolated snippets: changes must fit architecture, controls and audit processes already in place.

 

Barclays has not disclosed the number of developers covered by the percentage target, the model versions involved or the commercial terms of the partnership. It also has not published measures such as accepted-code rates, deployment frequency, defect counts or time saved, leaving the productivity case to be tested as adoption grows.

 

Banking AI Competition Moves Into Daily Operations

The rollout shows how competition among enterprise AI providers is shifting from general chatbots toward deeply embedded workflows. In banking, model access alone is not enough; deployments must connect to internal knowledge, route sensitive communications and support software estates while preserving permissions, review and recordkeeping.

 

Anthropic is also selling specialized Claude tools to financial advisers, while OpenAI and Google are targeting investment banks, asset managers and other financial institutions. Barclays’ program is distinctive because it combines employee search, market operations and developer tooling under one expansion with disclosed usage figures.

 

The next evidence will come from operational results. Barclays has provided clear adoption milestones for developers, but customers and regulators will care about whether the systems improve service and resilience without creating new errors, data risks or hard-to-audit decisions. Usage at this scale makes transparent measurement increasingly consequential for the bank and its technology supplier.