British bank Barclays announced on October 1 that it is expanding its strategic partnership with Anthropic. Under the expanded collaboration, Claude AI systems will be deployed across the bank's global operations — from software development to customer servicing. It is being described as one of the largest deployments of agentic AI in a heavily regulated industry such as financial services.

Under the plan, 50% of Barclays developers will be using Anthropic's agentic coding tool Claude Code by the end of 2026, rising to a majority of software engineers in 2027.

What the partnership covers

The partnership spans four main areas: accelerating software development, modernizing legacy systems, improving customer servicing, and increasing operational efficiency in the Global Markets business.

According to Craig Bright, Group Co-Chief Operating Officer at Barclays, the bank sees AI as "an increasingly agentic capability embedded within how we build, test, secure, and operate technology." Claude helps modernize legacy platforms and improve software quality, while technical specialists focus on the most complex challenges.

Systems already in production

The partnership is not on paper — several systems are already live:

  • Colleague Knowledge Assistant — a Claude- and RAG (retrieval-augmented generation)-based knowledge assistant live since 2025. It helps more than 16,000 staff serving over 20 million UK retail customers find information faster and has handled over one million searches.
  • Global Markets — Claude models are used to classify, enrich and determine the optimal processing route for incoming client emails. The platform processes approximately 120,000 emails each day, reducing manual handling.

"Responsible scaling": governance first

The bank stresses that as it expands AI, it maintains robust governance, security controls and human oversight. According to Ann Marie Darling, Group Co-Chief Operating Officer at Barclays:

"The true measure of any technology is the impact it has on customers and colleagues, and the outcomes it enables. Expanding the use of Claude across Barclays is not just about adopting AI — it is about transforming how work gets done."

Anthropic's Chief Commercial Officer Paul Smith called it a significant milestone for the company in the UK: "Claude now helps 16,000 Barclays colleagues find answers for customers, sorts 120,000 emails a day, and will be in the hands of most Barclays engineers by 2027. Very few institutions hold AI to such a high standard."

Why it matters

The deal matters greatly to both sides. For Anthropic, preparing for an IPO, growing its enterprise customer base is important; in financial services, competition with OpenAI and Google is intensifying. For Barclays — under a plan announced in February, the bank aims to save £2 billion (~$2.7 billion), including through AI; Morgan Stanley estimates AI could enable European banks to significantly reduce headcount.

More broadly, this is further proof that AI is moving beyond pilot projects into the core operations of large regulated institutions.

Uzbekistan context

Uzbekistan's banking sector is also undergoing digital transformation: a number of banks are rolling out mobile apps, remote identification and automated scoring. The Barclays example could serve as a practical roadmap for local banks: first direct AI at internal efficiency — staff knowledge bases, document processing, classifying customer inquiries — then move to customer-facing services.

Such projects in Uzbekistan carry additional considerations: quality AI models in the Uzbek language, local data-storage requirements and cybersecurity. Barclays' "governance first" approach — starting an AI project with strong controls and human review — is the right model, especially for financial services. For IT Park residents and local fintech startups, it is a signal that partnership opportunities with banks on AI integration are opening up.