BBVA has announced plans to work with Nvidia to explore new approaches to AI agent security, governance and oversight as it looks to expand the use of autonomous AI across the bank.

The collaboration will focus on Nvidia's OpenShell Reference Architecture, which the technology company unveiled at its GTC Berlin conference. The framework aims to help organisations monitor AI agents and intervene when they exceed predefined security boundaries.

BBVA said the initiative forms part of its strategy to scale AI agents across the organisation while maintaining oversight of their actions.

The Spanish bank said AI agents are taking on increasingly complex tasks, making it more important for organisations to establish clear limits on what they can do. In financial services, where traceability and risk management are key requirements, BBVA believes stronger governance mechanisms will become increasingly important as AI adoption grows.

As part of the collaboration, BBVA will explore ways to embed security and governance into AI agents during development, while also monitoring their decisions and actions once they are running.

Nvidia's OpenShell Reference Architecture combines infrastructure for deploying AI agents with independent security systems that monitor their interactions with external tools and detect activity that falls outside defined boundaries. Nvidia has released the OpenShell secure runtime software as open source.

Valentín Sánchez, global chief strategy officer at BBVA, said: “We are expanding the use of AI agents across our organisation, and ensuring their security is a priority. As these systems become more autonomous, we need to strengthen the mechanisms that allow us to monitor their activities and maintain control.

“Working with the technology industry and partners such as Nvidia gives us access to leading-edge approaches to AI agent security and governance, helping us continue to make progress in this area.”

Earlier this week, Santander announced that it had expanded its partnership with Nvidia after developing banking-specific AI models.

The bank said its Santander AI Lab has developed new capabilities to evaluate AI models and AI agents for banking applications, which can help identify the most suitable model for specific tasks while improving efficiency. The tools have significantly reduced inference costs while also improving customer service performance in experimental deployments, it added.

---