The moat strategy is failing everywhere AI can reach.

Why moats are eroding

Moats — the structural defenses a business builds against competitors — used to compound. Better technology. Better data. Better integrations. Better network effects. Each year you stayed ahead, the lead got harder for anyone else to close.

AI just collapsed several of those advantages at once.

The team of fifty engineers building the proprietary recommendation engine? Their work is now reproducible by a small team with a good API. The custom data pipeline? Half of it is commodity. The category-defining UI? AI is shifting the interface itself, and the assumption that users come to your interface at all.

Whole categories are watching their technological edges go flat at the same time. The companies that built around the moat are now exposed. Anyone who only had the moat is in real trouble.

What replaces moats

The opposite of a moat is gravity. Instead of keeping competitors out, you pull users in. Instead of structural defense, you build experiential depth.

The Apple version: an iPhone is a good phone. iCloud is decent storage. AppleID is convenient login. None of those individually is a moat. But the compounding benefit of having all of them in one connected world is significant. People don't stay because they can't leave. They stay because everything just works.

That's gravity. Compounding benefits the longer you engage. Depth that rewards commitment. A world someone willingly falls into.

The six gravity wells

A high-gravity world calls these emotional wells:

Data — what the user gains by using your product. Not what you collect about them; what they accumulate through use that they'd lose if they left.

Personalization — alignment with who the user is becoming. Adapting to where they are in their life arc.

Service — removal of friction. The feeling of "of course this was already handled."

Loyalty — status. Tiers. Recognition. The compounding feeling of being inside.

Connecting — the cultural tribe the product creates. Belonging. The thing users recruit for.

Transactions — bundled value. Paying once and getting many things. Exit friction by way of enrichment, not lock-in.

A great experience touches several of these. A great world touches all six.

Why this matters more in an AI-mediated world

AI matchmaking rewards clarity and depth. A product with a sharp world — clear identity, distinct tribe, recognizable feel — gives AI a model to recommend.

A product hiding behind a moat with no internal world is hard for AI to describe. It can't articulate why someone would want it.

Gravity-built products are recommendable. Moat-built products were never recommendable in the first place — they were unavoidable. That's a different thing.

Avoidance is going away. Recommendation is replacing it. Build for that.