ONS Consulting Group, LLC

Have you ever tried to teach a child common sense?

I mean really tried. You point at the stove. You say the words slowly, like that will help. “Don’t touch it. It’s hot.” The child looks at you, looks at the stove, and touches it. And you stand there thinking, I just told you. I used my clearest voice. And none of it mattered, because common sense isn’t something you can hand to someone. It’s something they have to go get for themselves, usually the hard way.

That’s the thing about common sense. It can’t be taught. It can only be learned. And learning happens through experience.

Which is exactly why I keep coming back to the way we talk about artificial intelligence. AI, as we know it, is book smart. It has read everything. It knows everything. Ask it anything and it will give you the answer with total confidence, the way a straight-A student recites the chapter.

AGI, artificial general intelligence, is what I’d call street smart.

Picture a crosswalk. The signal changes. The little walking person lights up. Book-smart AI looks at the signal and says, cross, it’s your turn. Street-smart AGI looks at the signal, then looks both ways anyway, because somebody might be taking a right on red while checking their phone, and the signal can’t see that car coming. Both of them read the same sign. Only one of them survives the intersection.

That’s the gap, and it’s the whole game. Even when everything says it’s safe to go, you still validate for the exceptions. And the exceptions aren’t in the manual. They’re learned through experience, one scraped knee at a time.

Now, there is an exception to the exception, because there’s always one. We don’t let people learn cybersecurity the hard way. We drill them on phishing emails relentlessly, the same scenarios over and over, because a bad actor in the network is a lesson nobody can afford. That’s targeted training, and it works beautifully for the things we can predict. But you can’t target-train the entire world. For everything else, you need a human in the loop who has the experience, guiding the machine.

Designer Kate Canales gave a TED talk that nails this, and I think about it more than I’d like to admit. Her point: people design perfect systems. A point of sale. An elevator. A door. Then the door gets handles on both sides, and nobody knows whether to push or pull, so somebody tapes up a little sign. A restroom never gets labeled properly, so a human being makes their own sign and tapes it to the door. The design was fine. It was just missing a couple of things, and humans filled the gaps with handwritten notes.

Your business runs on those notes. I promise you it does. Somewhere in your company there’s a legacy system with three different ways to run the same transaction, and only one of them actually works the way it was intended, so somebody left a note that says don’t use this one, use this one. Multiply that by thousands. How do you log into this site. Which report is the real one. Who actually approves the thing. That’s tribal knowledge, the human-to-human exchange that happens when a new person gets onboarded and someone takes them aside and tells them how it really works. It lives in your employees’ heads, and it walks out the door with them every night.

AI is going to clear a lot of small frictions, and we should let it. Email replaced the three-day ritual of the envelope and the stamp. E-signature killed the print-sign-mail loop. Genuine wins. But every shortcut invents a new hiding place. We used to have a file cabinet full of contracts. Now the contracts are sitting in somebody’s inbox and nobody knows whose. The challenges don’t magically disappear. They just move somewhere harder to find.

So keep the human in the loop. Not as a cost to trim, but as the teacher. Let the experience in your people’s heads go back and forth with the machine, every day, getting a little better. Start small. Stay in the loop.

People. Process. Technology. In that order.

Scott Hagberg is Chief Technology Officer at United Capital Financial Advisors, Field CTO at Creative Planning, and Founder and Principal Consultant at ONS Consulting Group.