System prompt
Definition
A system prompt is a set of standing instructions given to an AI model before the conversation starts, telling it how to behave in every reply. It sets the model's role, tone, rules and output format, and people chatting with the assistant usually never see it.
By Ronan Liedmeier, fractional head of growth and course author · Published 17 October 2026 · Last updated 17 October 2026
Also called: system message, system instructions
In an API call it is labeled separately from the user's messages. In chat products it sits behind custom instructions or a custom assistant's setup. Anyone building an assistant or an automation writes one, and it shapes every answer that assistant gives.
How it works
The system prompt goes out with every request, ahead of the conversation. Anthropic's documentation shows it passed through a dedicated system parameter and notes that even one sentence setting a role there changes the model's focus and tone. Because the model reads it on every turn, it suits rules that must always apply.
Keep it lean and stable. Reference material that only some requests need is better loaded when it's needed, a core idea in context engineering. An unchanged system prompt also pairs well with prompt caching, since the same opening text can be reused across calls.
Example
A garden-center owner sets up an assistant to draft replies to customer emails. Its system prompt says: write as the shop, warm and short, quote prices only from the attached price list, never promise a delivery date, and pass anything about refunds to a person. Every draft follows those rules without her repeating them.
Common mistake
The kitchen-sink system prompt, with every policy and edge case stuffed in just in case. It costs tokens on every call and buries the rules that matter.
Learn it in AI Academy
- Prompt Engineering at Production QualityExpert · 8 min
- Context EngineeringExpert · 8 min
Related terms
- Prompt
- A prompt is the instruction or question you give an AI model to tell it what you want. It can be one line or a full brief with background, examples and the format you need back, and the answer is shaped by everything in it.
- Context engineering
- Context engineering is the practice of deciding what information an AI model sees each time it works: the instructions, documents, tool descriptions, memory and conversation history. It extends prompt engineering from the wording of one request to everything loaded around it.
- Prompt caching
- Prompt caching is an AI API feature that stores the processed start of a prompt you send repeatedly, so later requests beginning with exactly the same text are cheaper and faster. Only the repeated beginning is reused; the answer is still written fresh each time.
- Custom GPT
- A custom GPT is a version of ChatGPT set up for one job, with its own instructions, reference files and enabled tools, so anyone who uses it gets the same behavior without explaining the task each time.