// Press Kit
Press kit
Everything needed to cite the State of the Agent-Readable Web report. All assets and figures are CC BY 4.0 — use freely with attribution to grow.contact.
Media contact
Author: Gudmundur Eyberg Kristjansson
Email: hello@grow.contact
Response time: within 24 hours, weekdays
Topics: GEO, agent-native web, AI citation rates, llms.txt, AI crawlers
Pull quotes
- "63% of the top 390 AI companies still ship no usable llms.txt — the cheapest GEO win on the agent-native web."
- "27% of AI companies score below the threshold AI engines will cite by name."
- "Only 13% of AI companies clear the agent-native bar — meaning the rest are effectively invisible to ChatGPT, Perplexity, and Claude."
- "When the AI industry's own marketing sites can't be cited by AI, the upside for the first competitor to fix it is asymmetric."
Attribution: grow.contact, State of the Agent-Readable Web Q2 2026 (CC BY 4.0).
Headline statistics
63% of 390 AI companies are missing or under-serving llms.txt.
67% ship insufficient JSON-LD for reliable AI citation.
13% clear the agent-native bar (score ≥ 85).
27% score below 55 — effectively opaque to ChatGPT, Perplexity, and Claude.
66% fail the first-byte speed threshold AI crawlers timeout against.
Dev Tools leads at 72/100 average; Agents trails at 57/100.
All figures derive from the open dataset at /api/public/leaderboard.json. Re-verify any number by re-scoring its source row at /check.
Author bio
Short (50 words): Gudmundur Eyberg Kristjansson is the founder of grow.contact, an agent-native web agency that builds fixed-price, 48-hour marketing sites engineered for AI citation. The quarterly State of the Agent-Readable Web report is the firm's primary research output, drawing on an open dataset of 390 AI companies.
Long (120 words): Gudmundur Eyberg Kristjansson runs grow.contact, the GEO and agent-native web agency that built and maintains the open Agent Readability Leaderboard — currently scoring 390 AI-industry companies across five signals (Semantic HTML, JSON-LD, llms.txt, Citability, Page Speed). The firm publishes a quarterly report and monthly data drops on AI-citation readiness, all under CC BY 4.0. Prior work covers fixed-price launch builds for AI startups (Tier 01 $2,400 / 48h, Tier 02 $4,800 / 5d), the free /check scanner, and an open-source CLI for CI-integrated agent-readability scoring. Based in Iceland; clients across the US and EU.
Downloadable assets
- Full dataset (JSON, CC BY 4.0) — 390 rows, five signal scores per row.
- Q2 2026 report (web) — canonical citation URL.
- Methodology — scoring formula, weights, limitations.
- Data drops feed — monthly single-stat findings.
- RSS feed — drops + reports as they publish.
Suggested topics for interviews
- Why 27% of the AI industry's own marketing sites can't be cited by AI engines
- The OAI-SearchBot vs GPTBot confusion that silently kills ChatGPT citations
- llms.txt: the cheapest GEO win and why most companies skip it
- What changes when AI engines become the primary source of inbound traffic
- How AI engines decide which competitor to cite when both rank similarly