Jakarta, INTI - Microsoft introduced an internal multi-agent AI system designed to proactively identify and help fix security vulnerabilities across the company's cloud infrastructure. The system was developed as part of Microsoft’s Secure Future Initiative (SFI), amid the rise of increasingly sophisticated cyber threats driven by the use of AI.
The system not only examines application code but also analyzes various aspects of the cloud infrastructure, including identity settings, network configurations, infrastructure, runtime environments, and other cloud resources.
This approach allows Microsoft to detect vulnerabilities that emerge from interactions between systems, rather than simply finding bugs in isolated software.
Microsoft Increased Operational Efficiency
The platform uses specialized agents to map the cloud architecture, validate security controls, evaluate layered protections, and recommend long-term correction steps. With the AI system, Microsoft said the inspection process, which previously took weeks, can be reduced to just a few hours.
This technology has been used to help strengthen Microsoft's production environment. Over 90% of AI-generated findings reviewed by security engineers are confirmed as real security issues, including several cross-domain risks that could potentially go undetected by traditional scanning systems.
The system is currently being used internally and is not yet offered as a product to customers. However, Microsoft plans to leverage the findings and results from this system's development to improve future security products and services.
Conclusion
Microsoft developed an internal multi-agent AI system to proactively detect and help fix security vulnerabilities across its cloud infrastructure. The system examines various aspects including application code, identity settings, and network configurations. By leveraging specialized AI agents, security review processes that previously took weeks can be completed in hours.
Read more: Dabeeo and IPB Collaborate to Identify Oil Palm Deficiencies Using AI and Satellite Data