With AI weaving itself into the fabric of the global economy, executives are turning to creative solutions to power this new workforce. At the forefront of these conversations are an increasing need for security to mitigate enterprise risk.
One company balancing the ROI of rapid AI expansion with guardrails around an expectation gap and vulnerability exposure is Rubrik.
“The number one thing stopping AI acceleration actually isn’t that it’s difficult to build models or agents anymore,” Dev Rishi, GM of AI at the company tells GeekWire Studios during a conversation at AWS re:Invent 2025. “It’s really security and risk.”
To illustrate these challenges, Rishi shares three critical issues that organizations must solve to unlock AI’s value:
- Visibility and Discovery: Increased ease brings a management problem. As building AI agents becomes more accessible, maintaining a comprehensive view of deployed agents becomes critical to managing risk.
- Real-Time Governance: Gone are the days of playbooks of governance frameworks and time-consuming human consensus-based regulations. New agentic solutions require real-time policy enforcement.
- Rapid Remediation: Evolving traditional cybersecurity enforcement models needs to adapt to the accelerated pace and scale of potential damage from AI systems and threats.
“With AI agents, you have the potential for 10x the damage in one-tenth the time,” Rishi warns.
One solution to tackle these challenges is Rubrik Agent Cloud, a comprehensive platform that addresses all three risk dimensions.
The platform starts with a single pane of glass across various AI platforms, tracking what agents are running, their potential impact, and the type, and scope of, data they can access. Integrating natively, and through gateway-based traffic observation, these tools work with multiple platforms including those available on AWS.
During re:Invent, Rubrik announced a key integration with AWS Bedrock Agent Core, enabling seamless observability for agents running on one of the industry’s most popular AI platforms. “You need to have native hooks into the most popular areas so you can have a turnkey experience for observability,” Rishi explains.
Monitoring the scope and potential impact of AI systems is the first step, but the ability to undo destructive action by AI Agents is where Rubrik’s AI Agent Rewind steps in. Using Rubrik’s protected backup data, the ability to revert changes creates what is essentially a time machine for AI operations.
While Rubrik has long been known for data resilience, (first for disaster recovery, then for ransomware protection), the evolution of their core mission is clear.
“Over the next 3 to 5 years, it’s going to be about being resilient to the next generation of the AI workforce,” Rishi predicts. “Today, all employees go through background checks, they’re manual, and they operate at human speed. Increasingly, organizations are going to be operating at AI speed, and they’re going to need an infrastructure system that can support that.”
Weaving these intelligent observability tools, governance systems, and risk management models into the core of your AI-powered agents will provide the resilience necessary to unlock the transformative power of AI investments. Rubrik provides timely solutions to meet the challenges of the coming AI workforce.
To learn more about Rubrik, go to https://www.rubrik.com/.
