Barclays (NYSE: BCS) is widening its partnership with Anthropic, bringing the Claude family of models deeper into day-to-day work across the British universal bank. The October 1, 2026 announcement frames the move as an effort to place secure, enterprise-grade systems throughout global operations, with particular emphasis on software engineering, aging technology stacks, and everyday efficiency.
Under the expanded arrangement, Barclays expects Claude Code to be in use by half of its developer population by the close of 2026, and by a majority of software engineers during 2027.
The bank is treating the coding tool as part of a broader shift toward agentic systems that sit inside how technology is built, tested, secured, and run.
Group co-chief operating officer Craig Bright said software engineering and cybersecurity are both being reshaped by more capable models, and that Claude is being used to help refresh legacy platforms, raise software quality, and free technical specialists for harder problems.
He argued that pairing skilled engineers and cyber specialists with trusted AI could change how large enterprises develop and manage technology.
The rollout is also framed as a customer and colleague story rather than a pure technology upgrade.
Anne Marie Darling, also group co-chief operating officer, said the real test of any tool is the effect it has on customers, clients, and staff.
Expanding Claude, she said, is meant to change how work is done: redesign processes, improve experiences, and produce a simpler organization inside a controlled environment.
Cutting time spent on routine tasks and tightening decision paths, she added, should let people concentrate on complex problems, speed innovation, and improve results for the business and those it serves.
One deployment already live is the Colleague Knowledge Assistant used by Barclays UK staff who support more than 20 million retail customers.
Running since 2025 on Claude through a retrieval-augmented generation setup, the assistant has been taken up by more than 16,000 colleagues and has handled over one million searches, giving staff faster access to information and shortening the path to customer support.
In Global Markets, models classify, enrich, and choose a processing path for incoming client emails so that requests reach operations teams with the details needed to act.
The platform handles roughly 120,000 emails a day, cutting manual handling and helping staff spot and act on relevant information more quickly.
The bank says governance, security controls, and human oversight stay central as use grows, so systems are introduced safely.
Anthropic chief commercial officer Paul Smith called the work a significant UK milestone, noting that Barclays serves millions of customers and clients worldwide and therefore applies a high bar, which is why the rollout includes strong security and oversight.
He summarized the current footprint as helping 16,000 colleagues answer customer questions, sorting 120,000 emails a day, and reaching most engineers by 2027, and said few institutions hold AI to that standard.
Both sides describe the deeper collaboration as a long-term bet on productivity, resilience, and better customer and client outcomes, pairing Anthropic’s frontier models with Barclays’ emphasis on responsible innovation inside one of the world’s largest financial institutions.