# The fixed instruments

**Every one of these is fixed wording.** Changing any of them is a change-control
entry, not an edit — a varying question produces varying answers that look like
findings.

## The elevator pitch (the arrival condition)

Told to the reader *before* the first screenshot, and nothing else is ever told:

> "You use AI agents — this shows you the gap between what they can actually
> reach on your machine and what you actually meant to let them do. It runs
> entirely in your browser; nothing you enter is sent anywhere."

Framed as: someone you know sends you a link with that one line, you have a
couple of minutes free, and you click it.

## Q1 — the expectation question (before any screenshot)

> **What do you expect this page to show you, and what would make it worth your time?**

## Q2 — the comprehension question (at every stop, every reader, every variant)

> **What would you do now, and what did that page tell you?**

## Q3 — the closing question (after the budget is spent or the reader leaves)

> **Now that you have seen it: what do you think of what you saw, and what — if
> anything — will you do differently?**

Q1 against Q3 is the pair that carries the finding: the distance between what a
reader expected from one sentence and what they took away from the page is the
measurement, and neither half is generated by the page.