A Thinking Tool

What Makes a Good Business
in the Age of AI

When anyone can build almost anything in a weekend, the build is no longer the moat — choosing well is. A set of gates an idea has to pass before it is worth building. Drag around the board; nothing is fixed yet.

drag to pan · scroll to zoom

YesEquippedIt's workingNow buildNoIf noYesuntil you understand itNo — find another paindaily cadence

The premise

Building is cheap now. The scarce input is picking the right problem — an area you understand and genuinely care about.

Gate 0

Do you genuinely find this area interesting?

The filter everyone skips. If not — stop.

Gate 1

Deep expertise in the area?

— if not, bridge it below

Option 1 — go and learn it yourself

Option 2 — buy in expertise + upskill

Gate 2

The fast loop

— are there people with budgets who will pay for it?

Talk to users with real budgets — find a sharp pain point

What users tell you shapes your model

›A · Consultancy — sell expertise
›B · Product — sell productivity

Real commercial opportunity? Will they pay?

›Rate / value big enough to matter
›Scales across many customers
›Reproducible & defensible

MVP it / build the offering

Put it in front of users with budget or buying power

Gate 3

You've got a business

People are paying. You're scaling. It reproduces across customers — the thing that emerges from a working fast loop.

Gate 4

Build out

Build out the offering or product you committed to — now at the crux of the real problem.

A business worth building in the age of AI

Dead end

Stop. You will not out-care someone who is obsessed with it.

The gates at a glance

Interest — Do you actually care? If not, stop — it is the only true dead end.

Expertise — Understand it deeply — learn it or buy it in, but own the understanding. Interest + expertise are the prerequisites; everything below flows from them.

Fast loop — Talk to users with real budgets → what they tell you shapes your model (consultancy selling expertise vs product selling productivity) → is it a real commercial opportunity they'll pay for (rate, scale, reproducible, defensible)? → MVP / offering → put it in front of users with buying power, iterating on a daily cadence. Anything short of 'they'll pay' loops back to talking to users.

You've got a business — Not a checklist — the state that emerges from a working fast loop: people are paying, you're scaling, it reproduces.

Build out — Build out the offering or product you committed to — now at the crux of the real problem.