LLM (large language model)

An LLM (large language model) is an AI model trained on enormous text volumes to predict the most plausible continuation of an utterance — a capability that, at scale, produces understanding and writing. It's the technology behind modern conversational agents, written and voice alike: the LLM is the brain, speech recognition the ear, speech synthesis the mouth.

Its limit for a business: a general LLM knows language, not your business. It can write a polite reply; it knows neither your prices, nor your areas, nor your slots. So any serious application frames it: a system prompt setting its role and rules, a closed knowledge base supplying your information, confidence thresholds triggering human escalation. The raw LLM is an engine; the product is the chassis around it.