Barclays Pushes Claude Across Global Operations, Targets Half of Developers by Year-End
Barclays is scaling its use of Anthropic's Claude models across its global operations, moving the technology well beyond the pilot stage and into core banking workflows. The expansion, announced by Anthropic and confirmed by Barclays leadership, targets 50% of the bank's developers using Claude Code by the end of 2026, with a majority of its software engineering workforce expected to be on the tool by 2027.
The rollout builds on Barclays' existing Colleague Knowledge Assistant, a Claude-powered internal search tool that has been live since 2025 and is now used by more than 16,000 employees, who have run over 1 million queries through it. In the bank's Global Markets division, Claude models are already classifying and routing roughly 120,000 client emails a day, a volume previously handled through manual triage.
From chat assistant to agentic engineering
What distinguishes this phase from Barclays' earlier generative AI work is the shift toward agentic use in software development itself. The bank, which serves more than 20 million UK retail customers and is running a multi-year efficiency program aimed at roughly £2 billion in savings, is positioning Claude as a tool for modernizing legacy systems and accelerating code testing and security review, not just answering employee questions.
We're moving towards AI as an increasingly agentic capability embedded within how we build, test, secure, and operate technology. — Craig Bright, Co-Chief Operating Officer, Barclays
Part of Anthropic's broader enterprise financial push
The Barclays deal lands alongside a wider Anthropic effort to lock in large regulated-industry customers, following reported engagements with hedge funds and asset managers including ExodusPoint, Millennium Management, and Man Group, and a separate $100 million commitment this month to train 10,000 'Frontier Deployed Engineers' with consulting partners to help enterprises actually operationalize AI deployments rather than stall at the pilot stage.
For a 330-year-old institution operating under heavy regulatory scrutiny, the emphasis on risk management is notable. Barclays executives framed the expansion carefully around governance rather than speed.
The true measure of any technology is the impact it has on customers, clients, and colleagues. — Anne Marie Darling, Co-Chief Operating Officer, Barclays
The announcement is a useful data point for machine.net's Entity Signal Scores and AI Momentum Index trackers: a top-tier bank committing to majority developer adoption on a specific calendar timeline is a concrete, falsifiable adoption claim in a sector where most enterprise AI announcements remain vaguer pilots. Whether Barclays hits the 50% developer-adoption mark by December, and whether that translates into measurable productivity or cost gains rather than just usage numbers, will be worth tracking into 2027.