The most expensive real estate in the world isn’t oceanfront - it’s the default button.

Google reportedly paid Apple $18–$20 billion to be the default search on Safari - not the best search, the one that shows up without a thought. That price isn’t about features. It’s about gravity: the path everything takes when no one is pushing.

This matters because most teams still try to win by arguing, branding, or persuading. Meanwhile, the winners are quietly grading the slope so the flow moves their way. Control the slope and you don’t have to shout; you just have to be there when the decision makes itself.

We’ve seen this before. A tiny policy tweak that changes nothing but the starting point can change everything about the outcome. Automatic enrollment boosts 401(k) participation by entire workforces - not by inspiring them with retirement sermons, but by making saving the path of least resistance. When countries switch organ donation to opt-out instead of opt-in, consent rates leap to above 90%. When Apple turned off third-party tracking by default, ad platforms didn’t adjust their pitches; they bled. Different domains, same pattern: defaults quietly bend behavior.

We keep treating the world like a debate club. It’s closer to a river.

Rivers don’t negotiate with rocks; they follow the smallest gradient and start carving. The Grand Canyon didn’t appear because the Colorado River was persuasive. Water followed the easiest route, and the route compounded itself: flow deepened the channel, a deeper channel increased speed, and speed accelerated erosion. The flow creates the canyon that then dictates the flow.

Products, policies, and markets work the same way. The riverbed is the default. The flow is human behavior. Every click you remove, every field you pre-fill, every setting you make the starting point is a millimeter off the riverbank - but across years, it’s a canyon.

Here’s the important nuance: the riverbed doesn’t need to be perfect; it only needs to be preferred. Users are satisficers. They don’t climb hills for tiny gains; they follow the slope that’s already downhill. A merely OK experience on a well-graded slope beats a great experience you have to hike to.

This is why the AI wars won’t be won by the smartest model. They’ll be won by whoever becomes the default layer between you and everything else - and can actually deliver.

Microsoft Copilot looks like it should have already won. It’s embedded in Office, connected to your SharePoint, reading your emails, summarizing your Teams meetings. It’s the perfect default—pre-installed, pre-integrated, pre-authorized. The riverbed couldn’t be better graded.

But defaults only work if the water actually flows. Copilot shows how even unmatched distribution can backfire if the product misses the minimum bar of usefulness. If every summary misses the point, if every SharePoint search returns nonsense, if the AI can’t actually help with the work… people will climb out of the canyon. They’ll copy-paste into ChatGPT. They’ll try Claude. They’ll find their own rivers.

That’s the paradox: the default position is priceless, but only if it’s good enough to keep people in the channel. Google Search wasn’t perfect; it was just good enough that climbing out felt pointless. Copilot risks teaching millions of enterprise users the opposite lesson: that the default can be worse than nothing.

Microsoft owns the most valuable real estate in enterprise AI: every Office toolbar, every Teams window—but they’re fumbling the handoff. The channel is perfect, but the water won’t flow.

Which means the throne is still empty. The next trillion-dollar company won’t just become the AI default - they’ll be the first one good enough to keep it.

The riverbed is ready. We’re just waiting for water worth flowing.