Observability, Prevention and Remediation for AI Security

Rubrik's Devvret Rishi on Using a Three-Pillar Approach for Production Gen AI Chris Riotta (@chrisriotta) • August 28, 2025

Organizations deploying generative artificial intelligence in production environments face unprecedented security challenges that require a fundamental shift in how they protect sensitive data and infrastructure. The complexity of securing AI systems demands a strategy to address proactive protection and reactive response capabilities. This will become critical as AI agents gain more autonomy and can interact directly with production systems.

See Also: Post-Quantum Cryptography - A Fundamental Pillar in the Future of Cybersecurity [ES]

"In terms of securing applications, there's really three key things that I think organizations need to be prepared with from a tooling standpoint. The first is observability, knowing where each of these AI calls are being recorded and what types of data are going into those systems," said Devvret Rishi, general manager of AI at Rubrik. "The second ...