OpenText recently surveyed 1,900 executives and found a worrying trend: 52% had used AI solutions, but only one in five saw real results This article explores enterprise cybersecurity opentext. . 59% said that privacy and regulatory compliance issues were getting in the way of AI projects growing.

OpenText emphasizes how important it is to have ongoing, context-aware controls. For example, sensitive data should be de-identified before it goes into AI pipelines, agent behavior should be controlled by intent-based guardrails, and agents should be scanned for vulnerabilities before they are put into use. Greg Clark, Director of Product Management and Strategy for Enterprise Cybersecurity at OpenText, said, "It has to be ongoing and relevant." He said, "If not, chaos reigns."

The challenge gets harder when there are more non-human identities than human ones, with a ratio of 20 to 1 or 50 to 1.