I run a startup accelerator, which means I read a genuinely heroic number of pitch decks. And lately, almost every single one has found a way to staple the letters A-I to the front of whatever the company actually does. A bakery becomes an AI-powered bakery. A tutoring service becomes an AI-native learning platform. I understand the reflex, because that’s where the money and the attention are pointing right now. But I want to take founders off a particular hook, because I think it’s the wrong one to be squirming on.
You don’t have to be an AI company. You have to build a good company and use AI well while you do it. Those are two very different projects, and mixing them up is sending a lot of sharp founders in the wrong direction.
Here’s what I actually love about this moment if you’re small. AI has made it possible for a team of five to do what used to take fifty. A founder today can stand up a first version, handle early support, draft the marketing, and run the numbers with a fraction of the people it once required. If you’re starting from scratch, that’s a gift the big incumbents would kill for. They have to unlearn how they work. You get to design it on purpose from day one.
The trap is thinking leverage means going lean to the point of hollow, or aiming AI at the parts of your company that were never supposed to be automated. I’ve watched founders proudly automate the exact thing that made customers fall in love with them, the care, the taste, the human touch, and then wonder why growth went flat.
So the founders I see getting this right make two decisions early and on purpose.
Point AI hard at the work that drains a small team. The repetitive, the administrative, the stuff no one dreamed of doing. That’s how five people keep moving like fifty without burning out.
Fiercely protect the few things that are actually your edge. The product taste, the customer relationships, the judgment calls. Staff those with real humans, and grow those humans deliberately.
Your moat was never going to be the AI. Everybody has the same AI. Your moat is the specific, slightly weird, deeply human thing your team is great at. The tools just free you to spend far more of your time there.
I’ve spent years now sitting with early teams, and the ones that last have something in common, and it isn’t the most impressive tech stack. They know exactly what they’re for, who they serve, and what only they can do, and they use every tool available to guard their time for that. You really can build fast and human at the same time. Starting small might be the best possible position to build from.
If you’re early and trying to figure out where AI belongs and where your people should stay front and center, that’s a lot of what I do with founders.
Build the thing only you can build. Let AI carry the rest.