The next decade, inside a company
The next
ten years.
Forget the endgame for a second. Here’s what the next decade of AI actually looks like inside a company.
Everyone wants to argue about where this ends up. Fair questions. But they skip the part that’s actually actionable, and it’s far simpler than either the doom or the hype: the work of the next five to ten years is your company learning to use a new and very powerful tool. That’s the whole assignment.
Stage one
First, five people become the expert class.
They learn what something like Claude Code is. They teach themselves to build, even though they’re “business people.” They stop thinking in budgets and start thinking in tokens.
Quietly, they begin doing the work of fifty.
Stage two
Then it compounds.
Each of those five stops doing the work by hand and starts running little kingdoms of agents — dozens, eventually hundreds — each doing the job of ten or fifty people. They become the champions of their departments.
And then, the way everyone slowly learned to use a computer two decades ago, the other forty-five start learning too, until running your own agents is simply how everyone works.
The quiet thing underneath
The company’s memory is what actually compounds.
As the champions build, they don’t just automate tasks. They build the company’s memory. The brand voice. The context. The decisions and the reasons behind them, the knowledge that usually dies in someone’s inbox.
Pulled into one place, it lets anyone ask the questions you can never quite ask out loud:
“How would the boss phrase this?”
“What happened with that client back in the spring, and why did we lose them?”
“What did we decide last quarter, and why?”
“What does our best work actually look like?”
It gives a leader a real view of what the company is actually working on — not what gets said in the standup. And it gives every employee, every day, the leverage of everything the company has ever learned.
The tool is already sitting on the desk.
The team teaches the AI, the AI makes the team better, the two keep compounding. None of it requires betting on what the world looks like in fifteen years. The only real question is who picks it up first.