Find the Leaky Prompt Wall
Identify where unlabeled prompt material can cause the model to confuse instructions, examples, and data.
A product team war room with a launch planner standing at a wall of blank sticky-note clusters and a laptop showing an empty document layout, cool blue-gray room, realistic coworkers, clear central wall for overlay hotspots, no text, words, letters, numerals, logos. A launch planner faces three blended clusters: current memo notes, last quarter's example checklist, and output-format notes. The issue is the unlabeled boundary between the old example checklist and current requirements. The AI keeps importing old launch tasks. Which part of the scene is most likely causing the model to treat reference examples as live instructions? Old example…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in