The moat is no longer the app.

AI will commoditise code faster than it commoditises context, workflow and proprietary data. That is where enduring enterprise value will sit, in the data, the workflow and the institutional logic underneath the software, not in the software itself.

How we work →

The flywheel →

The doctrine

Four lines we build the company on.

Not slogans. The structural beliefs that decide what we build, where we enter, and how we capture value.

The moat is no longer the app.

It’s the data, workflow and institutional logic underneath it. The part AI doesn’t commoditise.

It’s the data, workflow and institutional logic underneath it — the part AI doesn’t commoditise.

First do the work, then own the workflow.

You earn the right to codify a process by carrying its risk, not by advising from outside it.

If the founder is still in the room, the multiple is a mirage.

A business that depends on a person is a job. One that runs on a captured system is an asset.

Messy industries are the opportunity.

Friction creates workflow, workflow creates data and data creates enduring enterprise value.

WHAT WINNING ACTUALLY IS

The category leader is just the best-run business in the room.

Strip away the brand and the founder myth. The company that takes a service industry takes it because its machine is better — it converts more, runs leaner, knows more, and compounds faster than anyone it competes with. Every one of those advantages is now buildable. Most of your competitors are leaving them on the table.

SAME MARKET. SAME CUSTOMERS. THE GAP IS OPERATIONAL — WHICH MEANS IT IS BUILDABLE.

Where value sits now

Coding isn’t software. Software isn’t workflow. Workflow isn’t enterprise value.

For thirty years the moat was the application. When anyone can generate the application in an afternoon, the moat moves down, to the workflow and the proprietary data underneath it.

AI doesn’t eliminate that value. It changes where new value is created: not in writing the code, but in encoding how an organisation actually functions: the records, the history, the hand-offs, the approvals, the operating memory that makes the business work.

That is why we don’t start with a feature set. We start with the work: real operational pain, a clear cost line, a workflow people are already paying to execute.

We build defensibility through that workflow, the process layer where data compounds, trust builds and switching costs emerge on their own. The software is operating leverage, not the product.

Own the work. Then the workflow. Then the data. Then the economics.

Every engagement begins in the same place, inside the work, and climbs the same ladder. Each rung is harder to take and harder to copy than the last, which is exactly why the value compounds at the top.

Why N+1 is different