In recent years, cybersecurity has crossed a critical threshold: we are no longer facing merely “digital” attacks, but autonomous, fast, and adaptive operations—often conducted by agentic AI systems capable of making decisions without human supervision.
Between 2021 and 2024, the mean time to data exfiltration (MTTE) after an initial intrusion dropped from 9 days to less than one hour. In some simulated scenarios, a ransomware attack carried out with AI across the entire kill chain can be completed in just 25 minutes—a speed up to 100 times faster than traditional attacks.
Agentic AI marks a further evolutionary leap: it does not simply automate individual stages, but observes the environment, evaluates options, negotiates with other systems, and autonomously decides how and when to strike. This enables attacks that can adapt to defenses in real time, change strategy during execution, and exploit even the smallest signs of weakness, including those across the digital supply chain.
At the same time, the same agentic logic is transforming defense: security systems capable of correlating events, anticipating attack patterns, orchestrating automated responses, and drastically reducing the time between detection and containment. The contest is no longer human versus machine—but AI versus AI.
The event “Cybersecurity in the Age of Agentic AI” was created to help companies, decision-makers, and security leaders understand this new landscape: not only the emerging threats, but also the models of defense, governance, and accountability required when critical decisions are made by autonomous systems.