Speakers: Tim Gürtler and Markus Alzer

Generative AI answers questions. AI agents go a step further: They access data and applications, evaluate information, and trigger actions on their own.

But AI agents must not become a black box. Companies must clearly define what an agent is allowed to see, decide, and change—and who is responsible for it.

Using a specific information security process as an example, this presentation demonstrates how AI agents can be identified, evaluated, and controlled as assets relevant to governance and integrated into business processes—with a clear mandate, limited permissions, a designated owner, and clear approval workflows.

An integrated management system links requirements, processes, risks, responsibilities, and evidence.

S ...