Also known as: NLP
In plain English
NLP is what lets computers work with words: translating a sentence, detecting whether a review is positive or negative, or answering a question you type.
In practice
NLP used to be a set of separate tools for tasks like sentiment analysis, entity extraction and translation. Large language models now handle most of these in one model, but smaller specialised NLP models are still cheaper and faster for high-volume tasks.
Under the hood
Classic NLP pipelines used tokenisation, part-of-speech tagging, parsing and statistical models. Modern NLP relies on pre-trained transformers that learn contextual representations of text and are adapted to tasks by fine-tuning or prompting.
Example
"We use NLP to pull invoice numbers and amounts out of supplier emails."