Capital One Says Agentic AI Deployment Hinges on Data Foundation and Platform-First Approach

Capital One Says Agentic AI Deployment Hinges on Data Foundation and Platform-First Approach

Rashmi Shetty, Vice President of Enterprise AI at Capital One, said the bank's scaled deployment of agentic AI depends on a solid data foundation and a platform-first approach; agent systems can already take real actions and must be treated as end-to-end systems. She leads the bank's enterprise platform engineering team, bringing generative AI and agentic AI applications into business lines including retail banking, risk, legal and compliance.

According to Shetty, Capital One runs both internal and external agentic AI use cases. On the external side, Chat Concierge is one of the bank's earliest multi-agent applications and has been in production for more than two years, helping customers through the car-buying process, including connecting with dealers, scheduling test drives and advancing transactions online; internally, it is used to improve customer service workflows. Shetty attributes the current progress to long-term investment in the data foundation, including well-governed data pipelines and data lineage.

The bank's CEO, Richard Fairbank, said on the second-quarter 2026 earnings call in July that over the past 14 years the bank has rebuilt itself from the bottom of the technology stack upward and has continued to invest in foundational capabilities and AI infrastructure. Shetty said the next key step is platform-first: enterprise platforms need built-in codified policy and strategy validation, runtime controls, compliance and mandatory cybersecurity guardrails to support agentic use cases at scale.

Source: https://www.bankingdive.com/news/capital-one-agentic-ai-strategy/831461/