NLP you already use
If you've ever used email spam filtering, autocomplete, voice dictation, or Google Translate, you've used NLP. It's the technology that lets machines make sense of messy, ambiguous human language — understanding not just words, but meaning, intent, and context.
NLP tasks for business
Sentiment analysis: Automatically gauge whether customer feedback, reviews, or survey responses are positive, negative, or neutral. Useful for monitoring brand perception at scale.
Named entity recognition: Extract specific information from unstructured text — company names from emails, dates from contracts, product codes from invoices. This feeds into data pipelines that eliminate manual data entry.
Text classification: Route support tickets to the right team, categorise inbound enquiries by urgency, or tag documents by type. Rules-based classification breaks with edge cases; NLP handles the nuance.
Summarisation: Condense long documents, meeting transcripts, or email threads into key points. Especially valuable when LLMs do the summarising — they can follow specific instructions about what to include.
NLP and LLMs
Modern large language models have subsumed most traditional NLP tasks. Where you once needed separate models for sentiment, extraction, and classification, a single LLM can handle all three with appropriate prompting. This dramatically lowers the barrier to entry for SMBs wanting to process text at scale.