Skip to content

Agents

When software acts, the question is who may act and what stops it.

Prompt injection: a CISO playbookInjection is the method. The tool budget is the cause. Six controls you can show, and a test that measures tool calls — not manners.
Who is liable when an AI agent acts?An agent has no legal personality. Provider, deployer, product liability, and ordinary agency decide who pays — and a jailbreak does not change that.

Put AI to work. Keep control of what matters.

Five engagements. Start with the problem in front of you: the architecture, one department, the whole organization, the stack you already run, or the work after launch.

Architecture ReviewSee what is running, where control is thin, and what has to change before you invest further.

Department DeploymentGoverned AI for one department and one real use case, on infrastructure you control.

Enterprise RolloutOne governed stack across the organization, on infrastructure you control.

Governance OverlayClearer controls and a verifiable record on the stack you already run.

Managed OperationsHardening, training, and engineering after launch, scoped to the engagement.