Methodology

How the scores work

One axis, six pillars, cohort-conditional weighting. Everything reconstructed from public artefacts.

One axis

Every company gets one score. Built (0–100) measures internal capability: how seriously the company is operating with AI agents inside the org.

Six pillars

Cohort weighting

AI-natives and enterprises are scored on different weights — talent posture matters more for natives, organizational formation matters more for enterprises. Final scores are normalized to 0–100 within each cohort.

Single-source cap

If a company's highest-scoring pillar lacks convergent signal in any other pillar, the Built score is capped at 60. We flag this on the profile page.

Updated monthly. Published by Deepnote.