Cybersecurity and agentic AI: key insights from RSA Conference 2026 for an automated, integrated AI-Native security model

April 9, 2026

The 2026 edition of the RSA Conference, held in San Francisco, has once again positioned Cybersecurity at the centre of digital transformation. The conversations we have taken part in, along with exchanges with clients and partners, confirm a trend we have long been driving: the shift towards more automated, integrated security models, ready to support the secure and governed adoption of AI in organisations.

This shift entails both a technological transformation and an evolution in business processes and operations, as organisations need to strengthen business protection and risk management in distributed, regulated and complex environments, where AI is becoming a decisive part of operations.

Security for AI and agentic AI

Through our participation in RSAC 2026, we have observed the evolution of AI security, particularly in agentic AI environments, which represent a new level of operational and security complexity. In this context, the focus is shifting towards AI governance (including models and agents), data protection, visibility over its adoption within organisations, and oversight of interactions between models, tools and corporate systems.

We are already seeing concrete use cases for agentic AI, such as developers using local AI agents (shadow developers), whose activity cannot be secured solely at the network level and introduces direct implications for identity. This means that protecting AI—both models and agents—requires combining multiple security layers (endpoint, network, gateway and identity) within an integrated protection approach.

Security is evolving towards comprehensive governance and oversight of AI and its agents.

At the same time, risk vectors linked mainly to the use of generative AI models are intensifying, such as prompt injection and jailbreaks. These are forcing organisations to strengthen their monitoring and control capabilities and improve visibility over the unsupervised use of AI tools within the organisation.

At Telefónica Tech, we are advancing this evolution through our integrated Secure Journey to AI approach, helping organisations address challenges related to AI governance, visibility, data protection and secure operations from the design phase and throughout the entire AI lifecycle, integrating capabilities across different layers.

Protecting AI requires a multi-layered approach combining identity, data and visibility.

Data, regulation and digital sovereignty

In our discussions with clients and partners, we have also reinforced the importance of data in Cybersecurity, particularly in the European context. Data sovereignty and data residency, together with traceability and operational control, are becoming key priorities for organisations in a context of growing regulatory pressure in Europe, with new requirements around operational resilience, technology risk management, data protection and responsible AI use.

Data is becoming a critical pillar of Cybersecurity and European regulation.

At the same time, we are progressing in our preparation for post-quantum cryptography, anticipating medium- and long-term risks while strengthening information protection in current and future scenarios.

Sovereignty, traceability and control are key in an increasingly regulated environment.

Towards an AI-Native security operations model

During RSAC 2026, we have seen a market trend consolidate: the evolution of security operations towards more automated, AI-driven models, with greater integration and response capabilities.

In this context, we are moving towards security operations with enhanced context, analysis and response capabilities, where AI agents and expert teams work in a coordinated way. This AI-Native SOC model strengthens detection, response and cyber resilience in organisations with hybrid and multicloud infrastructures.

The combination of AI and expert talent is driving cyber resilience.

At Telefónica Tech, we combine these advanced automation capabilities with the expertise of our teams, integrating data protection, identity security, and advanced detection, analysis and response capabilities, aligned with regulatory requirements.

What we have seen at the RSA Conference 2026 confirms that the path we are following in the evolution of Cybersecurity is the right one to support the accelerated adoption of AI, highlighting our advanced capabilities in automation, integration and operations to support our clients in this new landscape.