Ask DECIMAG

DECIMAG is a runtime architecture that adds a decision layer before execution. It evaluates whether the action itself should happen in the current context.

It is relevant to cybersecurity, but it is not a classic cybersecurity product. It is a decision layer that can use information collected by security systems.

Between authorization and execution. After existing systems finish their checks, and before the action actually happens.

IAM helps prove who the actor is and what they can access. DECIMAG asks a different question: should this action happen under the current conditions?

Rules are useful for defined cases. DECIMAG is designed for a broader runtime evaluation where policy is one input, not the entire picture.

The heavy work is prepared offline. At runtime, the system evaluates against an existing model, with the goal of adding minimal latency. Performance must be validated in pilots.

The model can reason, analyze or suggest. The execution decision should be handled by a runtime layer that is controllable, explainable and auditable.

Prompts influence model behavior. They do not create an independent decision layer before execution.

No. DECIMAG does not initiate actions or define business goals. It evaluates whether a requested action should be executed.

Yes. A pilot can start with one workflow, one system or one specific high-impact action.

Yes. The goal is to store each decision with the version of reality that produced it, so the decision can be reviewed later.

Fail-open or fail-closed behavior depends on the organization's risk model and the action type. Sensitive actions may require fail-closed behavior.