Data Security Governance—The First Step to Safe AI Adoption
Krishnan has spent his career in security at companies large and small, including Juniper Networks, PGP Corporation, Symantec, and HPE, where he ran the security portfolio after the company acquired a behavioral analytics startup he helped found. Over those years, he watched organizations struggle to put data security into practice. Tools built on regexes and rigid rules lacked any real understanding of context. Customers spent far more on the people needed to tune and manage them than on the tools themselves, and the results were still unreliable.
Concentric AI started from a different premise. Long before ChatGPT, Krishnan’s co-founder proposed using large language models to read data the way a person would: first understanding what a document is, then judging whether it is sensitive. The result is a platform that classifies data with human-like accuracy at petabyte scale and surfaces the risks around it. The biggest of those risks is oversharing: Krishnan estimates that roughly 40% of an enterprise’s sensitive data is accessible to people who shouldn’t have it.
AI makes that exposure far more dangerous. Information that once required knowing where to look can now be surfaced by an AI assistant with a single prompt, and data pasted into public AI tools can end up training models used by billions. Concentric AI addresses this by protecting data at rest, in motion, and across the AI tools employees use. The goal is that people, and the agents acting on their behalf, see only what they’re entitled to.
The Innovator Series is meant to get inside the mind of the founder, and learn what inspired and motivated them to find success in the cybersecurity industry. Learn more about Karthik Krishnan and Concentric AI in our 14-minute conversation on YouTube.
