Skip to content
// The method

Two habits, and the whole thing works better.

Getting a reliable answer out of AI isn’t about a magic prompt or a better model. It comes down to two habits: asking well, and checking what you get back. This site is built on both. Here they are.

01
How to ask

The Prompt Stack

Four steps that turn a vague question into one an AI can’t bluff its way through: SCOPE, FILTER, RISK, VERDICT. The prompt I run before I trust anything it tells me.

See the Prompt Stack →
02
How to check

The Guardrails

Seven checks you run on the answer you got back, before you act on it. No enterprise tooling, no dev setup — the checks a normal person can run for free, proven on real receipts.

See the Guardrails →

The one-page version you can paste straight into any chat is the Bluff Filter. And the proof that this matters — where the big assistants actually break — is the Scoreboard.