The moat, measured. Not a deck — a timestamp.

Collected weekly. Anonymized. Auditable.

How each number is measured.

M-1 · AI citations

A locked weekly set of real user questions is sent to ChatGPT, Perplexity, Claude, and Gemini. Each response is scanned for Fonteum URLs. The headline is the number of citations across all queries that week.

M-3 · Snapshot history depth

For each federal source family we hold, we count distinct dated snapshots. The headline sums those across families — a proxy for how much history exists that a fresh competitor cannot backfill.

M-4 · FHIR API requests

Requests to the live FHIR R4 endpoints are tallied weekly and classified by consumer (AI agents, developers, browsers). The headline is total weekly requests.

M-6 · Press mentions

Distinct news/media URLs mentioning Fonteum, discovered weekly. The headline is the count of unique mentions that week.

M-19 · Build velocity

Pull requests merged to the main branch in the trailing 7 days, read directly from version control — a founder-execution signal.

Metrics are append-only and timestamped at collection. Each tile carries its 12-month target and an honest state: live once a reading exists, pending — first run when a self-contained metric is awaiting its next scheduled collection, or pending — key when a metric is gated on an external credential or upstream feed still to be provisioned. No placeholder zeros are ever shown. Full collection details, alert thresholds, and the underlying schema are documented in the methodology index.