NLG is the process of turning structured data or internal system state into clear, human-readable language. A reporting system that reads sales figures from a database and writes a plain-language summary of performance — rather than presenting raw numbers — is using NLG.
It is the output side of natural language processing. Any time a system needs to communicate information to a human in a natural, readable form — rather than requiring the human to interpret raw data — NLG is at work. This applies to reporting, notifications, recommendations, customer communications, and anywhere that translating data into language is currently done manually.
