Natural language understanding (NLU) is the branch of artificial intelligence (AI) that interprets what a person means, not just the words they used. When a customer says “I was charged twice last month”, NLU is what identifies the intent (a billing dispute), the key details (duplicate charge, last month), and the sentiment behind them.
How NLU Works
NLU models are trained on large volumes of language and perform several jobs at once. Intent recognition classifies what the person wants. Entity extraction pulls out the specifics: dates, amounts, account references, product names. Context and sentiment analysis read how the person is saying it and what has already been said in the conversation.
In a contact center, NLU is the understanding layer inside voicebots, chatbots, and routing systems: the step between hearing the customer and deciding what to do.
NLU vs NLP vs NLG
Natural language processing (NLP) is the umbrella field covering everything computers do with human language. NLU is the part concerned with comprehension: working out meaning. Natural language generation (NLG) is the part concerned with producing language: composing the reply. A conversational system uses NLU to understand the customer and NLG to answer.