BSL Insights

The Last 10%: What AI Can’t Replace

Big Strategy Labs

In Kyoto, there is a centuries-old practice called Kintsugi — repairing broken pottery with gold rather than hiding the cracks. The piece isn’t considered finished when it’s whole. It’s considered finished when the break has been made visible, and valuable, by the care applied afterward.

That’s a useful way to think about where AI is taking knowledge work.

AI is doing the first 90%

Drafting, summarizing, searching, coding, modeling — the technical heavy lifting that used to take a team a week now takes an afternoon. That shift is real, it’s fast, and it isn’t slowing down. Treating it as a threat to resist is already the wrong frame. The more useful question is: once the first 90% is commoditized, what’s left to compete on?

The last 10% is where the value moves

Three things don’t compress, no matter how good the tooling gets:

Judgment — knowing which problem is actually worth solving, and when a technically correct answer is still the wrong call.

Empathy — reading a room, a market, or a counterparty well enough to know what isn’t being said.

Creativity — the leap from “here’s what the data shows” to “here’s what we should build because of it.”

As the first 90% gets cheaper and faster for everyone, these three stop being soft skills and start being the entire competitive advantage. They’re the gold in the crack — the part that doesn’t just repair the work, but makes it worth more than it would have been unbroken.

What this means for builders

The founders, operators, and students we work with aren’t being asked to out-build AI on execution speed. They’re being asked to get sharper on the 10% that AI is not designed to replace: the strategic judgment that decides what gets built, and the human read on the people it’s being built for.

That’s the room Big Strategy Labs is built around.

Get involved

Ready to build what’s next?

Join a community of founders, operators, and executives moving from strategy to scale.