AI Agents Are Identities: Survey Finds 82% Spotted Unmanaged Ones

As AI agents increasingly become "identities" that authenticate to enterprise systems, access data and take actions, Token Security argues most organizations are failing to govern them as such. In a survey commissioned by the company, 82% of organizations discovered at least one AI agent over the past year that their security, IT or governance teams were unaware of, and 41% experienced this multiple times.