Nudge Security report: AI agents inherit employee access permissions, and their use is unsupervised
AI agents inherit employee access, and nobody’s watching how they use it
AI summary
Nudge Security reports that AI agents employees build themselves on platforms such as project tracking and CRM typically access company data using the permissions of their creator. These agents mostly run on the creator's OAuth authorization or API key, and logs show only the employee, not the agent; among organizations that have agents, the median number is 7, and about 44% have more than 10. The report recommends inventorying agents within platforms and revoking OAuth authorizations that are no longer needed.