A large language model (LLM) is a type of artificial intelligence trained on very large amounts of text so it can understand questions and generate human-like language. It underpins modern AI assistants and chatbots.
The model learns patterns in language and predicts what text should come next. When given a prompt, it generates a reply word by word. Businesses can guide it with instructions, supply it with their own documents, or connect it to tools so it can act, such as looking up an order.
Retailers use LLMs for customer support chat, product descriptions, summarising feedback and helping staff find information. When combined with business data they can answer questions specific to the company.
LLMs can produce confident but wrong answers, so important replies need guardrails, approved sources and a clear route to a human. Privacy of customer data also needs consideration when choosing a provider.