Declaring what the AI cannot do or must avoid before the real request gives those constraints cognitive weight from the start. Models handle limitations better when they appear early rather than as afterthoughts.
Prompt anchoring is the practice of stating your most critical requirements, limitations, or non-negotiable conditions at the very beginning of a prompt so the AI treats them as fixed parameters throughout its entire response.
Anchoring prevents AI from drifting toward generic or off-target answers, and it is especially useful for tasks with strict word counts, legal boundaries, audience restrictions, or formatting requirements that cannot be ignored.
Peri can explain this concept, give practical examples, help you decide whether it applies to your situation, or recommend a journey if appropriate.
Explore related journeys or tell Peri what you're working through.