Published 2026-08-10 · The trust rules behind every score

What counts as AI adoption

Only direct AI signals score points: LLM APIs (OpenAI, Anthropic), agent frameworks (LangChain), vector databases (Pinecone), model hubs (HuggingFace), AI-native docs and AI hiring.

What never counts

• Generic infrastructure: Cloudflare, Next.js, SvelteKit, Astro, Vite, WordPress, nginx, AWS
• Generic chat software: Crisp, Drift, Intercom, Tawk.to — shown as context, never scored as AI

Three numbers, one profile

• AI Adoption Score (0-100) — direct AI signals only
• Evidence Confidence (0-95%) — signal count, corroboration across channels (website/docs/careers), hiring evidence
• Coverage — which channels were scanned

The core trust principle

Absence of evidence is not evidence of absence. A company with no public AI signals scores honestly at 0 — even if they use AI internally. We never inflate scores to look good.

Full details: Methodology page

← All articles · Methodology · Search the database