Large Language Model (LLM)
A large language model (LLM) is an AI model trained on vast text to predict and generate language, used in chat assistants, copilots, search answers, and agents.
Most AI UX patterns you ship today (streaming replies, citations, tool use) exist because LLMs are open-ended and conversational, not because users fill a fixed form.
What it means
An LLM reads a sequence of tokens (words and symbols) and generates the next tokens, producing paragraphs, code, JSON, or tool requests from natural-language instructions.
Why designers should care
LLM interfaces need scaffolds: prompts, examples, constraints, and post-processing, because the same UI control can produce wildly different outputs from small input changes.
Example
A PRD assistant uses an LLM behind a template with required sections. Users edit fields; the model fills gaps, but every section shows “Review before export” because tone and facts can drift.
Common mistakes
- • Letting users send unstructured walls of text with no guidance or output format.
- • Assuming the LLM “understands” your product without system prompts and context.
- • Omitting stop, regenerate, and edit affordances on long generations.