Before you send an AI-generated email, copy a ChatGPT answer into a document, or paste a Copilot summary into a meeting note, run it past three questions. In order. Takes thirty seconds.

1. Is anything in here factually wrong?

Read for confident-sounding statements. Names, numbers, dates, claims. If the AI wrote it like a fact, check it like one.

AI tools are fluent, not accurate. They produce text that reads well regardless of whether the content is correct. A Copilot summary might say "the client confirmed the March deadline" when the client actually asked for an extension. A ChatGPT answer might cite a statistic that sounds plausible but does not exist.

The more confident the tone, the more carefully you should check. AI does not hedge when it is unsure. It states things plainly, whether they are right or wrong.

2. Would I have phrased it this way?

Not "is it grammatically correct," but "does this sound like me."

AI has a voice. It tends towards the generic, the formal, the slightly over-polished. If it has a phrase you would never use in real life, swap it. If it starts every paragraph with "Furthermore" or "It is worth noting that", rewrite those openings.

This matters because your colleagues, clients and customers know what you sound like. If your email suddenly reads like a corporate press release, people notice. The trust cost is real, even if nobody says anything.

3. What is the one thing it missed?

AI gives you the average answer. Yours is supposed to be better than average. The thing it missed is usually the thing your reader actually needed.

Maybe the Copilot summary captured the action items from the meeting but missed the tension between two teams about the timeline. Maybe the ChatGPT draft covered the key points of the proposal but left out the one thing that makes your offer different from the competition.

The missing piece is almost always context that only you have. AI cannot access your judgement, your relationships, or your knowledge of what matters most to the person reading.

Make it a habit

Three questions, thirty seconds, every time. The output is no longer "what AI said" — it is "what you sent." That distinction is the whole game.

The people who get caught out by AI are not the ones who use it. They are the ones who use it without checking. The three-check habit is the difference between a tool that helps you and a tool that embarrasses you.