Most People Think AI Is a Chatbot. That’s the Expensive Mistake.

Most people think AI is a chatbot. That one mistake will decide which Amazon businesses fall behind, and they won’t fall a little behind.

It’s Sunday, right after church, and my friend is worried.

He manages SQL databases for a living. He’s heard the predictions. Companies won’t need as many people to run their databases. He isn’t sure where that leaves him.

“Imagine something a hundred times smarter than ChatGPT,” he says.

I tell him a hundred is too small. In a few years it won’t be a hundred times. It will be thousands.

So I borrow an example from Jeff Booth.

“Take a sheet of paper and fold it in half. Then fold it again. If you could fold it 50 times, how tall would the stack be?”

He thinks about it. “Probably a couple of miles.”

“Halfway to the sun.”

I undersold it

Later I ran the math, and I was wrong too.

A sheet of printer paper is about a tenth of a millimeter thick. Fold it 50 times and the stack is about 113 million kilometers tall. The sun is about 150 million kilometers away.

That’s not halfway. It’s three-quarters of the way.

I’ve watched this curve for years, and my own gut still came up short by about 37 million kilometers. That’s the trouble with exponential growth. Nobody’s intuition is built for it, mine included.

AI is on a curve like that. Moore’s Law, the number of transistors on a chip doubling about every two years, was enough to rebuild the world around the internet. METR, an independent research group, tracks a different curve: how long a task an AI agent can finish on its own, measured by how long the same task takes a skilled person. Since 2023, METR’s data shows that length doubling about every four months.

Run both curves for five years. Moore’s Law gets you about 6 times. The AI curve, if it held, gets you about 16,000 times.

That’s what I meant by thousands.

“Like talking to ChatGPT”

I keep going. The doubling rates. The agents. What I’ve built with them. Fact after fact.

He stops me.

“Talking to you is like talking to ChatGPT.”

He means it as a compliment.

And that’s when it lands. When he pictures the most powerful AI he can imagine, he pictures ChatGPT, only smarter. A box you type a question into. A box that types an answer back.

He isn’t behind on the news. He’s picturing the wrong kind of AI.

The chatbot isn’t what’s on the curve

At the time of writing, most Amazon sellers and agency owners I talk to use AI the way my friend does, though that’s shifting fast. Ask it to write a bullet point. Ask it what a metric means. Paste the answer somewhere.

That’s the chatbot. It’s useful. It’s also not the thing on the curve.

The thing on the curve is the agent: AI that does the work instead of describing it.

A chatbot tells you what ACoS means. An agent connected to your account pulls the last 60 days of search terms, finds the ones spending without converting, drafts the negative keywords, and waits for you to approve them.

A chatbot writes a client update after you paste in the numbers. An agent pulls the numbers itself.

The chatbot saves you minutes. The agent changes who does the work. And the agent is the one doubling every few months.

So the risk for a seller or an agency isn’t the one my friend is worried about. It isn’t that AI replaces the person running your ads. It’s that a competitor’s person learns to run an agent, and the gap between you keeps doubling. Not left behind. Exponentially left behind.

For a seller, that gap shows up as the campaigns nobody had time to check, the reorder caught a week late, the fee nobody noticed. For an agency, it shows up in what a client gets for the same retainer: one team reviews every account every week, another reviews the three loudest.

The way out is the advice I gave my friend. Learn what AI can do and what it can’t. Not by reading about it. Hand it one real task from your Amazon business, a search term report, a reimbursement check, a client report, and watch what happens.

My friend guessed a couple of miles. I guessed halfway to the sun. We were both wrong in the same direction.

The paper doesn’t care what we guess. It keeps folding.

Amazon AI Blog post

Ready to Sell Smarter?

Your All-in-One Solution for Amazon Success

Feedback Genius, Ad Genius and Profit Genius - three powerful tools designed to work together for seamless growth.

Get the bundle now