How AI agents work in Cutover runbooks

This is a demonstration of how AI agents can operate in Cutover runbooks. This example is on a server failover, and I've pulled the relevant agentic task to the top of the runbook for ease of demo. As I click start on the task, we can see how human oversight is built into the way that Cutover interacts with agents, from the ability to set a default instruction to the ability to edit it at run time if necessary. As I click continue, the agent kicks off. Once started, the agent provides full transparency from the instruction given to it to its every thought. These thoughts also include every function used and every data point accessed. This completeness of logging is what enables us to have completely explainable outcomes. This, in turn, gives us complete confidence in the outcome that it provides.