What Is Named Entity Recognition?

Named entity recognition (NER) is a natural-language processing task that finds and classifies named things in text, such as people, organizations, places, dates, and amounts. It scans a passage and tags each entity with its type, turning unstructured writing into labeled information a system can use.

How named entity recognition works

Named entity recognition works by reading through text and marking spans that name a specific thing, then labeling each with a category. In the sentence "Acme hired Dana Lee in Denver in March," it would tag Acme as an organization, Dana Lee as a person, Denver as a place, and March as a date.

Modern NER is usually done by models trained on labeled examples, or by large language models prompted to extract entities. The output is structured: a list of the entities found and their types, ready to feed into a database, a search index, or the next step in a workflow.

Why named entity recognition matters for AI

Named entity recognition matters because it pulls structured facts out of free text, which is where much business information lives. Extracting the names, dates, and amounts from contracts, emails, or reports lets that content be searched, sorted, and acted on, instead of sitting as unreadable paragraphs. It is a common building block in document processing and search. At Custom AI Studio, entity extraction is one way we turn a client's text-heavy documents into data their systems can use.