Definition
Natural language processing (NLP) is the branch of artificial intelligence concerned with enabling software to read, understand, interpret and generate human language. It covers tasks such as classifying emails, extracting names and dates from contracts, translating text, detecting sentiment in reviews, transcribing speech and answering questions. Large language models are the most powerful form of natural language processing today, but the field is decades older and includes many smaller, faster techniques.
In a company, natural language processing is everywhere text is: routing support tickets to the right team, pulling key terms out of contracts, monitoring brand sentiment, translating product content, summarising meetings and searching internal documents by meaning. Before language models, each task needed a dedicated model and labelled data. Now a general model handles most of them with an instruction, while specialised NLP models remain useful where volume is enormous and cost per item matters.
NLP as a separate discipline has largely merged into the world of large language models, and the classic pipeline of tokenising, tagging and parsing is mostly hidden inside them. Specialised models still win on speed and cost for narrow, high-volume tasks such as classification. The misconception is that natural language processing means chatbots. Most of its business value is silent: extraction, routing and search running in the background.
In practice
A telecom operator uses natural language processing to read 20,000 customer emails a day, detect intent and urgency, extract account details and route each message to the right queue before a human opens it.
Why it matters
Language is the format of most business information: emails, contracts, calls, reviews. NLP is what makes that information searchable, measurable and automatable.
Frequently asked questions
- What is the difference between NLP and LLM?
- Natural language processing is the broad field of making software work with human language. A large language model is one specific and currently dominant technology within that field. LLMs perform most NLP tasks, but NLP also includes smaller, faster models for classification, extraction and search.
- What are common NLP applications in business?
- Ticket and email routing, contract and invoice data extraction, sentiment analysis of reviews and surveys, translation of product content, call transcription and summarisation, chatbots and semantic search across internal documents.