Barclays Deepens Its Claude Deployment Across Global Operations
Barclays is widening its use of Anthropic's Claude, with specific targets for developer adoption and two live use cases already running inside the bank: a knowledge assistant used by thousands of employees and an email classification system handling a high daily volume in Global Markets.
Barclays is widening its use of Anthropic's Claude, with specific targets for developer adoption and two live use cases already running inside the bank: a knowledge assistant used by thousands of employees and an email classification system handling a high daily volume in Global Markets.
A Bank Sets Numbers, Not Just Intentions, on AI Adoption
Barclays and Anthropic are extending their collaboration, and the bank has attached concrete figures to that extension rather than leaving it as a general statement of intent. By the end of 2026, Barclays expects Claude Code to reach 50% of its developer population. By 2027, it expects the tool to reach a majority of its software engineers. These are adoption targets, not performance claims — Barclays has stated how many people it expects to use the tool, not what results it expects from that use. The distinction matters: a rollout schedule tells you about organizational commitment and planning, while a results claim would tell you about outcomes, and only the former is on the record here.
What Is Already Running Inside the Bank
Two concrete deployments exist alongside the developer rollout plan. The first is the Colleague Knowledge Assistant, live since 2025, which more than 16,000 Barclays employees have adopted and which has handled more than one million searches. Those numbers describe scale of adoption and volume of use — they establish that a large number of people have used the tool a large number of times, nothing more or less. The second deployment sits inside Global Markets, where Claude models help classify and route incoming emails on a platform that processes approximately 120,000 messages per day. That volume gives a sense of the scale at which classification is operating, though no figures on accuracy, speed, or downstream effects on trading or client response have been disclosed.
Oversight Attached to Every Use Case
Barclays has stated that it applies governance, security controls, and human oversight across its AI use cases. This is presented as a standing practice tied to how the bank runs these systems, not as a response to a specific incident or risk calculation disclosed publicly. Combined with the staged adoption targets for Claude Code and the already-measured scale of the Colleague Knowledge Assistant and email classification system, it paints a picture of a large financial institution moving deliberately: setting milestones, tracking usage figures, and keeping oversight structures in place as the footprint of the technology grows across both internal knowledge work and operational email processing.
Specs
- Claude Code adoption target by end of 2026: 50% of developer population
- Claude Code adoption target in 2027: Majority of software engineers
- Colleague Knowledge Assistant launch: 2025
- Colleague Knowledge Assistant adoption: More than 16,000 colleagues
- Colleague Knowledge Assistant searches: More than 1 million
- Global Markets email processing: Approximately 120,000 emails per day
Sources
- Anthropic — Anthropic