// Monthly Findings
Data drops
One stat. One methodology link. One copy-paste citation. Published monthly between quarterly reports. Every drop is derived from the open Agent Readability Leaderboard.
// 2026-05-26 · category-shift
Infra vs Dev Tools: the agent-readability gap inside AI
Dev Tools leads at 72/100; Agents trails at 57/100.
Read drop →// 2026-05-19 · blocking
27% of AI companies are effectively invisible to ChatGPT, Perplexity, and Claude
27% of 390 AI companies score below 55/100 on agent-readability — the threshold below which AI engines silently skip the page.
Read drop →// 2026-05-12 · adoption
llms.txt adoption stalls at 37% across top AI companies
63% of 390 tracked AI companies still ship no usable llms.txt — eighteen months after the spec landed.
Read drop →// 2026-05-05 · adoption
67% of AI companies ship insufficient JSON-LD for reliable citation
67% of 390 AI companies ship JSON-LD too thin for AI engines to verify entity claims.
Read drop →// 2026-04-28 · adoption
Only 13% of AI companies clear the agent-native bar
13% of 390 AI companies score 85/100 or higher — the threshold above which AI engines reliably cite the source by name.
Read drop →