Inference

Inference is a single run of a trained AI model on an input to produce an output. It is the unit of consumption in production AI: cost, latency, and governance are all measured per inference, which is why routing decisions and access controls are applied at this level rather than per application.

Training is a capital event; inference is an operating cost that recurs with every use. Systems that are cheap to build can be expensive to run, and the difference usually comes down to how many inferences a workflow triggers and which models serve them.

Because inference is the unit, it is also the right place to attach controls. Authentication, authorization, filtering, and logging applied per inference cannot be bypassed by reaching the system through a different interface.