Definition
Large language model (LLM)
A large language model (LLM) is the type of AI, trained on vast amounts of text, that powers tools like ChatGPT by predicting and generating human-like language.
A large language model is a system trained on enormous quantities of written text to predict what words come next, which turns out to be enough to let it answer questions, summarize, draft, translate, classify, and hold a conversation. It is the engine underneath ChatGPT, Google's Gemini, and the AI features increasingly built into restaurant software. When people say "AI" in a marketing or operations context today, an LLM is usually what they mean.
For a restaurant, LLMs show up in two ways that matter. First, they power tools you use: writing menu descriptions and social posts, drafting replies to reviews, answering guest questions by phone (see Voice AI), or reading your sales history and telling you something useful. One independent operator, Cali BBQ, used Toast's AI to analyze historical sales and decide to stay closed until 11:00 AM, $4,000 saved, a decision the AI produced in under a minute. Second, LLMs power the tools your guests use to find you, which is why Generative engine optimization (GEO) now matters.
The honest caveat: LLMs generate plausible language, not guaranteed truth. They can be confidently wrong, so they are best pointed at tasks where a human reviews the output or the cost of an error is low. That is why adoption skews to lower-risk uses: where operators use AI, 19%/15% marketing, 10% admin, 6% orders. Used with that judgment, an LLM is a genuine force multiplier for a small team.
See it in action
Your next service / Ready when you are
Less busywork. A better next shift.
Turn what you have learned into a focused brief. Start with the job that matters most, and keep the tools that already work for your restaurant.
Describe it. Review it. Keep improving it. Check connections and costs before going live.