I asked AI to poke holes in a client pitch last month before I sent it, and it found the exact objection the prospect raised twenty minutes into the actual meeting. I’d already built the answer, so instead of getting caught flat-footed, I watched the objection come and had a real response ready. That’s the whole value of this issue in one sentence, but I’ll keep going anyway, because the why matters more than the story.
Using AI to think harder, not just type faster.
Typing faster is the obvious use case, and it’s fine, but it’s the smallest part of what these tools are actually good at. The bigger use case is pressure testing, feeding it a plan and asking what’s wrong with it, what’s missing, what a skeptical customer would say back.
Here’s the honest reason most people don’t do this, and it isn’t laziness. Asking AI to build something feels productive, you get a visible artifact, a draft you can point to, a thing that didn’t exist a minute ago. Asking AI to attack something feels like the opposite of progress. You hand over something finished and get back a list of problems. It feels like going backwards. But going backwards on purpose, before a customer does it for you, is the entire discipline. The objection you find yourself is free. The objection your customer finds costs you the sale.
There’s a deeper reason this works, and it’s worth understanding so you can push it further. You are, structurally, the worst possible critic of your own idea. You know why it’s good, you’re attached to it, you’ve already spent energy on it, and your brain is quietly motivated to protect that investment. That’s not a character flaw, it’s just how human minds work with things they’ve made. AI has none of that attachment. It didn’t make your plan, it isn’t invested in your plan, it will cheerfully tell you your plan has a hole in it without any of the ego that makes it hard for you to see the hole yourself. You’re not using it for intelligence you lack. You’re using it for detachment you can’t have about your own work.
The trick is asking in a way that gives it permission to actually disagree. “Is this good?” invites reassurance, because agreeable is its default setting. “Give me the three strongest reasons a smart, skeptical version of my customer would say no to this” invites the thing you actually need. Same tool. The question does all the work.
Takeaway: Don’t just ask AI to help you finish a thought. Ask it to interrupt one.
Try this: Before your next big send, pitch, launch email, price change, ask AI what the strongest objection to it would be, phrased as if it were coming from your actual customer, not a generic critic. Then answer that objection inside the thing itself, before anyone else raises it. You’ll ship something that already survived its own worst review.
Talk soon, Tiffany