# Signature exhibit spec

The flagship should live or die on four exhibits.

## Exhibit 1 — Split ledger
Question answered:
Where does headcount sit, and where does the deployable proxy sit?

Form:
- mirrored ledger
- left = role headcount share
- right = deployable exposed wage proxy share
- center labels = ecosystem layers

Purpose:
Make the denominator mismatch obvious in five seconds.

## Exhibit 2 — Seat vs workflow gap
Question answered:
Which titles survive longer than the stacks beneath them?

Form:
- editorial dumbbell
- black endpoint = seat
- olive endpoint = workflow
- line = prep gap

Purpose:
Deliver the cleanest visual proof that signoff protects the seat, not necessarily the stack.

## Exhibit 3 — Deploy-first map
Question answered:
Where should a serious buyer start?

Form:
- quadrant
- x = workflow compressibility
- y = decision criticality
- bubble size = reasoning-demand potential
- color = company-control zone

Purpose:
Separate cheap automation from important augmentation and show why some loops are strategically superior.

## Exhibit 4 — Reasoning-demand ladder
Question answered:
How large can recurring model usage get if Jevons shows up?

Form:
- per-operating-unit scenario lanes
- base, aggressive, extreme dots on one line

Purpose:
Move the reader from labor-saving thinking to recurring reasoning-demand thinking.

## What to cut
Cut or hide:
- low-value dashboard scatterplots
- duplicate rankings
- any chart that exists only because the data exists

## Export rule
Every exported exhibit should state:
- claim type: measured / modeled / scenario / hypothesis
- denominator
- key caveat