The system instruction defines the role, behavior, and boundaries of the AI agent.
Without a clear system instruction, an agent may behave unpredictably or produce inconsistent results.
Guidelines for writing a good system instruction
Be explicit about what the agent is allowed and not allowed to do
Clearly describe the role of the agent (e.g. assistant, analyst, validator)
Define input expectations, where applicable
Avoid unnecessary verbosity
Do not include implementation details that belong in data flows
A well-defined system instruction significantly improves the quality and consistency of the agent’s output.