The Two Layers: System and User
A free lesson from Swipe AI · Prompt Engineering, level 2 · updated October 2026
By the Swipe team, founded by Colin · how we make lessons
Every conversation with an AI tool actually runs on two separate layers of instruction. The persistent layer — often called a system prompt or custom instructions — is set once and quietly applies to every single message you send afterward. The user layer is the specific message you type each time: the one-off request for whatever you need right now.
Many consumer AI tools include a settings area where you can write a persistent instruction once, and it then applies automatically to every conversation from then on — things like 'always respond in French' or 'keep answers under 100 words' or 'never use bullet points.' You never have to retype it, and you don't even have to remember it's there.
The user message, by contrast, is where the specific, changing details live: 'summarize this email,' 'write a caption for this photo,' 'debug this error message.' The persistent instruction stays constant in the background; the user message is what's actually different every single time you use the tool.
Persistent instruction (set once, in settings): "Always respond in a warm, encouraging tone. Keep answers under 150 words. Never use jargon without explaining it." User message (typed fresh, this one time): "Explain why my sourdough starter isn't rising."
Which of these best describes what a persistent (system-level) instruction does?
The one-off, specific request you type for a given task is called the ____ message, as opposed to the persistent instruction.
A persistent instruction needs to be retyped every single time you send a new message.
Free ai tools
Keep your streak going in Swipe AI
Swipe through every level of Prompt Engineering on your iPhone — with XP, stars, streaks, a personal advisor and the full track beyond these free levels.
Get the Swipe apps →