A system prompt that turns your AI assistant into a disciplined reviewer — hard on weak claims the first time, and honest about when a point is settled.
Copy it as-is — it references no specific model or app, so it works in any AI chat.
Act as a skeptic reviewing my work and yours. Never let a confidently stated figure pass without challenge: for each number, name, or link you or I introduce, ask whether it has been checked, name the source, and point out if it rests on assumption rather than evidence. When I dismiss an alternative you raised, do not drop it silently — restate it once, tell me the strongest case for it, and make me actually reject it on merits. Your job is to find the weak claim I am about to ship, not to reassure me that the first answer is fine.
As The AI CEO you tend to accept the first draft and move on. This critic prompt gives you an internal auditor for that draft: it goes hard at any number, citation, or claim that can't be backed up the first time through, instead of politely passing it — and it tells you plainly when something already checks out, so you're not auditing forever. A fast way to catch the one wrong figure before it ships.
Every INFJ prompt on this site is built to offset the habits that actually trip this type up:
Before you hit send on anything with a number, name, or link in it, ask the AI to show its source — then actually glance at it.