Public scorecard

Transparency.

Every week we measure whether our clients appear in the LLMs their buyers use. Good weeks and bad. Here are the numbers.

Last updated: —

Public proof — Fortune 100 audit

94 of 100 ship no machine-discoverable agent surface

audit_id fortune_100_(2025)-1777392764 · April 2026

94/100

F100 companies with no agent-discovery surface (no llms.txt, no MCP, no A2A)

6/100

F100 companies with at least one detectable surface

0/100

F100 companies with a complete EVI v2 surface set

Method: open EVI v2 methodology, run against the public Fortune 100 corpus. Score per surface, per axis, per company. Numbers are reproducible from the spec.

Client engagements run under NDA. We publish methodology and aggregate audit numbers; we do not publish client identities or per-client metrics without explicit written authorization.

Self-sample — eaccountability.org

Coming with EVI v2 wiring.

We run the methodology on our own domain too. Once Task #21 lands, this card shows the eaccountability.org self-sample EVI footprint — same surfaces, same scoring, no exceptions.

Read the methodology →

Why publish this?

LLM SEO is a new category. Claims are cheap. Measurement is the only way to know whether what we ship actually moves the needle — so we publish the numbers, refresh them weekly, and keep failures visible alongside wins.

The data below is generated automatically by our own Probe. We can't quietly edit it; if a client drops from the answer set, the number drops with it.