In this insightful discussion, Eon co-founders Ofir Ehrlich and Gonen Stein explore the evolving landscape of data infrastructure in the AI era. They highlight how traditional concerns around data security, primarily focused on human threats like ransomware, are now compounded by non-human actors—AI agents with legitimate access that can rapidly cause disruptions. This shift necessitates new methodologies for detecting and mitigating risks, emphasizing the need for granular, real-time protection and recovery mechanisms within enterprise environments.
Eon’s core offering is a cloud-based data foundation that maps, classifies, and ingests data from diverse sources across multiple cloud providers. This platform not only safeguards data through efficient backup and disaster recovery but also unlocks its value by enabling easy querying and AI model integration. The founders stress that while many organizations already possess vast amounts of data, much of it remains locked, fragmented, or underutilized. Eon’s technology addresses these challenges by providing a unified, cost-effective, and secure way to access and leverage historical and current data for AI-driven applications.
The conversation also delves into the emerging trend of acquiring legacy enterprise data, exemplified by Google’s purchase of Spirit Airlines’ data post-bankruptcy. Such data sets are invaluable for training AI models, especially for building realistic agents and applications that require authentic, high-quality data reflecting real-world operations. The founders note that as AI adoption accelerates, companies increasingly recognize data as their most critical asset, often more valuable than models or compute resources, and are exploring innovative ways to monetize or repurpose it.
Security concerns in the AI era are multifaceted, with enterprises facing threats not only from external attackers but also from internal AI agents operating with legitimate permissions. This complexity is heightened by the proliferation of non-technical users creating AI agents without full awareness of security or compliance implications. Ehrlich and Stein emphasize the importance of assuming breach scenarios and implementing robust monitoring, classification, and access controls to maintain data integrity and prevent unauthorized exposure or misuse.
Finally, the founders compare the current AI-driven transformation to the earlier cloud migration wave, noting that AI adoption is occurring at a much faster pace and with greater urgency. They discuss how legacy data infrastructure and processes must evolve to support dynamic, agent-driven interactions with data, requiring new tools that enable seamless data discovery, ingestion, and governance. Despite the challenges, they remain optimistic that these changes will democratize AI capabilities across organizations, fostering innovation and efficiency while reshaping how enterprises manage and utilize their most valuable resource: data.
Useful Links
