An experiment is a structured empirical procedure designed to test a hypothesis or evaluate a system under controlled conditions, producing observable and repeatable results. In the context of AI and software engineering it encompasses both scientific investigations (measuring model behaviour, benchmarking performance) and engineering trials (A/B tests, canary deployments, feature flags) aimed at generating evidence to guide design decisions.