Inference

Inference is the process of running input data through a trained AI model to produce an output — a prediction, classification, or generated text. In agentic systems, every inference call has cost, latency, and governance implications.