Enterprise security teams are increasingly turning to inference security platforms to combat a new wave of AI-driven runtime attacks, according to a recent report. These attacks exploit vulnerabilities in AI agents operating in production environments, where traditional security measures struggle to provide adequate protection.

The shift is driven by the speed and sophistication of modern attacks. CrowdStrike's 2025 Global Threat Report revealed that breakout times, the time it takes for an attacker to move from initial access to lateral movement within a network, can be as short as 51 seconds. This rapid pace often allows attackers to compromise systems before security teams can even react. The same report indicated that 79% of detected attacks were malware-free, relying instead on "hands-on keyboard" techniques that bypass traditional endpoint defenses.

Mike Riemer, field CISO at Ivanti, highlighted the accelerating pace of weaponization. "Threat actors are reverse engineering patches within 72 hours," Riemer told VentureBeat. "If a customer doesn't patch within 72 hours of release, they're open to exploit. The speed has been enhanced greatly by AI." This shrinking window of opportunity puts immense pressure on security teams, many of whom still rely on manual patching processes that can take weeks or months to complete.

Inference security platforms are designed to address these challenges by providing real-time visibility and control over AI agents at runtime. These platforms typically employ techniques such as behavioral analysis and anomaly detection to identify and mitigate malicious activity. By monitoring the behavior of AI models and their interactions with other systems, inference security platforms can detect attacks that would otherwise go unnoticed by traditional security tools.

The rise of AI-driven attacks represents a significant paradigm shift for enterprise security. Traditional security models, which focus on preventing initial access and detecting known malware signatures, are no longer sufficient to protect against these sophisticated threats. Organizations must adopt a more proactive and adaptive approach to security, one that leverages AI to defend against AI. The deployment of inference security platforms is emerging as a key strategy in this evolving landscape, enabling security teams to regain control and protect their critical assets from the growing threat of runtime attacks.