Hacks: The One Prompt That Makes AI Explain Anything Like You're Twelve

A friend of mine — sharp, successful, runs a small logistics company — called me last week because he'd been reading about changes to the pass-through tax deduction and couldn't figure out whether they applied to him. He'd tried Googling. He'd tried the IRS website. He'd tried asking ChatGPT, which gave him a wall of text that used the word "pursuant" three times.

"I feel stupid," he said.

He isn't stupid. He's a guy who can negotiate a freight contract in his sleep but doesn't have a tax vocabulary. And that gap — between being smart and knowing the jargon — is where most people get stuck with AI. They ask a question. The AI answers like a textbook. They nod politely and close the tab.

Here's the fix. One sentence, twelve words, and it changes every interaction you have with an AI tool from now on:

"Explain this to me like I'm a smart twelve-year-old."

Not "explain like I'm five" — that's the Reddit meme, and it produces baby talk. Not "simplify this" — that tells the AI to strip out important details. "Smart twelve-year-old" hits a specific register: clear enough that jargon gets translated, complex enough that the real ideas survive.

Why twelve?

A smart twelve-year-old can follow a logical argument. They can handle cause and effect. They know what money is, what rules are, what fairness means. They do not know what "amortization" means, what a "fiduciary" does, or why anyone would voluntarily read a lease agreement. They will call you out if you're being confusing — which is exactly the pressure you want on the AI.

When you tell Claude or ChatGPT to explain something to a smart twelve-year-old, you're not asking it to be less accurate. You're asking it to do the harder thing: be accurate and clear. That's a constraint, and constraints produce better output.

The before and after

I ran the same question through Claude twice. First, straight:

Prompt: "What is the qualified business income deduction?"

Response: A technically correct paragraph that included the phrases "specified service trade or business," "taxable income thresholds," and "combined qualified REIT dividends and publicly traded partnership income." All accurate. All useless to my friend.

Then with the twelve-word addition:

Prompt: "What is the qualified business income deduction? Explain this to me like I'm a smart twelve-year-old."

Response: "Imagine you run a lemonade stand, but it's a real business — you file taxes on it. The government says: because you're running this yourself and not through a big corporation, we'll let you skip paying taxes on up to 20% of what you earned. There are rules about who qualifies and limits if you make a lot of money, but the basic idea is: small business owners get a break that big corporations already have in other ways."

Same information. One version is a reference document. The other is understanding.

Where it works best

This isn't just for taxes. The twelve-word prompt works on anything where the gap between "expert language" and "human understanding" is the problem:

Medical results. "My bloodwork says my A1C is 6.1. Explain this to me like I'm a smart twelve-year-old." You'll get something like: "A1C measures your average blood sugar over the last three months. Think of it like a report card for how your body handles sugar. 6.1 is in the 'watch this' zone — not diabetic, but your body is working harder than it should."

Legal documents. "My lease says the landlord can enter with 24 hours' notice for 'reasonable inspections.' Explain what this means like I'm a smart twelve-year-old." The AI will strip the legalese and tell you what it means in practice, including the parts the landlord is hoping you don't ask about.

Financial planning. "What's the difference between a Roth IRA and a traditional IRA? Smart twelve-year-old." You'll get the tax-now-vs-tax-later distinction without a single mention of "marginal tax brackets."

Tech concepts. "What does end-to-end encryption actually mean? Smart twelve-year-old." Instead of a whitepaper, you get the locked-box metaphor — and it's usually a good one.

Insurance. This is the big one. Insurance policies are written to be incomprehensible. "My health insurance says I have a $3,000 deductible with 80/20 coinsurance after. Explain this like I'm a smart twelve-year-old." The response will walk you through a specific example — a $5,000 medical bill — dollar by dollar.

The advanced move

Once you've got the explanation, follow up with: "Now what are the three things about this that most people get wrong?"

This is the combination that changes everything. The first prompt gives you understanding. The second gives you judgment. Together, they take you from "I read about this" to "I actually know what to watch out for."

My logistics friend used both prompts on his tax question. The twelve-year-old explanation took thirty seconds to read. The "what do people get wrong" follow-up saved him from making a mistake that would have cost him four thousand dollars — he'd been planning to claim the deduction on income that qualified as a "specified service trade," which has lower thresholds. He didn't know that. He does now.

When it doesn't work

Two cases where you should skip this prompt:

1. When you actually need the jargon. If you're preparing to talk to a specialist — a doctor, a lawyer, an accountant — you want the expert vocabulary so you can speak their language. In that case, try: "Explain this to me like I'm a smart twelve-year-old, then tell me the actual terms I should use when I talk to my doctor about it."

1. When the topic is emotionally loaded. "Explain my mother's terminal diagnosis like I'm a smart twelve-year-old" is a bad prompt. Not because the AI can't do it, but because the clinical clarity might land harder than you're ready for. For medical results you're anxious about, try 🩺The Night Shift Nurse — she's built for exactly that conversation.

The real point

AI tools are not hard to use. The prompts are not the problem. The problem is that most people ask AI questions the same way they'd type them into Google — and Google has trained us to accept whatever comes back, whether we understand it or not.

The twelve-word prompt breaks that pattern. It tells the AI: I'm smart, I don't know this particular subject, and I need you to close the gap without talking down to me. That's not dumbing anything down. That's the opposite — it's demanding that the AI be good enough to be clear.

Twelve words. Try it tonight on whatever's been sitting in your "I should really understand this" pile. The tax thing. The insurance thing. The lease clause. The medical report. The investment your brother-in-law keeps telling you about.

You're not stupid. You just haven't had it explained right yet.