IABIGFR · Français
Open data, updated daily

AI-Readiness Index

Can AI search engines and assistants read, understand and cite the most popular websites? We audit the home pages of the Tranco top 10,000 on a transparent 100-point rubric, half of them every day.

Snapshot of 10,331 domains audited between November 7, 2026 and November 16, 2026 · every domain is re-audited every 2 days.

Headline numbers

77median AI-readiness score out of 100 (5,589 home pages scored)
11.5%are JavaScript-only shells that most AI crawlers cannot read
47.1%publish JSON-LD structured data
92.9%let every AI search crawler reach the home page
16.8%serve a valid llms.txt
51.7%have complete Open Graph tags

Percentages are of the 5,589 home pages we could read. The other 4,742 domains were left out: 2,097 unreachable or not running a website, 1,729 answered with a bot challenge or refused our request, 471 returned an error or no HTML, 445 disallow crawlers like ours in robots.txt (respected).

What the data says

  • The weakest area is Sitemap and llms.txt (on average 3.5 of 8 points); the strongest is AI search access (24.4 of 25).
  • 11.5% of readable home pages show fewer than 50 words without JavaScript: to most AI crawlers, which do not run JavaScript, they look almost empty. 70.7% serve 300 words or more in the raw HTML.
  • 47.1% publish JSON-LD structured data and 41.9% describe who they are (Organization, WebSite…); 0.5% ship at least one JSON-LD block that does not parse.
  • 56.2% have exactly one H1 heading, while 33.3% have none at all.
  • 16.8% serve a valid llms.txt, and only 14.4% organise it in sections with at least five links.
  • 0.2% use the non-standard noai / noimageai directives, and 1.5% set their home page to noindex.
  • Median home page: 152.2 KB of HTML, 685 readable words, 3.0% of the HTML is visible text; 337 home pages weigh more than 1 MB of HTML alone.

Score distribution

0–9: 1 sites10–910–19: 2 sites210–1920–29: 118 sites11820–2930–39: 505 sites50530–3940–49: 350 sites35040–4950–59: 379 sites37950–5960–69: 672 sites67260–6970–79: 1,065 sites1,06570–7980–89: 832 sites83280–8990–100: 1,665 sites1,66590–100

A 37.3% · B 26.4% · C 16.3% · D 8.7% · E 11.2% (A ≥ 85, B ≥ 70, C ≥ 55, D ≥ 40, E below 40)

Embed this chart

A PNG version of the chart, refreshed daily. Free to reuse under CC BY 4.0 with the credit link:

/charts/readiness-score-distribution.png

Average points per criterion

CriterionMaxAverageShare of max
AI search access2524.497.5%
Server-rendered content2015.376.8%
Title, description, canonical, language129.276.9%
Open Graph63.761.0%
Structured data156.845.3%
Heading structure106.564.5%
Sitemap and llms.txt83.544.2%
Indexable43.998.5%

Signal by signal

Signal on the home pageSitesShare
All AI search and browsing crawlers allowed on the home page5,19192.9%
At least 300 words readable without JavaScript3,95470.7%
JavaScript-only shell (under 50 words without JS)64511.5%
Title of 10 to 70 characters4,20675.3%
Meta description of 50 to 160 characters3,00153.7%
Any meta description4,52080.9%
Canonical URL3,80468.1%
Language declared (<html lang>)4,90687.8%
Exactly one H13,14056.2%
No H1 at all1,86333.3%
Two or more H2 headings3,66665.6%
Open Graph title, description and image2,89251.7%
Open Graph image3,17856.9%
Twitter/X card2,75149.2%
JSON-LD structured data2,63147.1%
JSON-LD identity (Organization, WebSite…)2,34441.9%
At least one invalid JSON-LD block270.5%
Sitemap (declared or /sitemap.xml)4,07372.9%
llms.txt93716.8%
llms.txt with sections and 5+ links80614.4%
Home page set to noindex841.5%
noai / noimageai directive110.2%
hreflang alternates1,51727.1%
Block at least one AI training crawler85315.3%

By popularity rank

Non-overlapping tranches of the Tranco ranking.

Tranco rankScoredMean scoreJS shellJSON-LDOpen Graphllms.txt
1–1004769.321.3%38.3%57.4%19.1%
101–1,00044274.510.2%50.9%59.7%25.1%
1,001–5,0002,11773.313.1%47.2%52.8%19.0%
5,001–10,0002,98373.310.5%46.6%49.7%13.9%

By top-level domain

TLDs with at least 30 scored home pages.

TLDScoredMean scoreJS shellJSON-LDOpen Graphllms.txt
.com2,7917610.5%54.4%56.8%22.0%
.org30371.35.6%34.0%43.2%11.2%
.net28070.410.0%39.6%45.7%13.2%
.ru15266.48.6%31.6%47.4%9.2%
.jp12868.910.9%34.4%62.5%4.7%
.io12682.36.3%58.7%69.8%40.5%
.de11876.45.1%51.7%40.7%11.0%
.edu9479.63.2%39.4%47.9%4.3%
.gov6471.34.7%18.8%50.0%1.6%
.cn6054.523.3%11.7%18.3%3.3%
.br5979.110.2%59.3%52.5%18.6%
.fr5674.57.1%50.0%46.4%1.8%
.uk5575.87.3%45.5%50.9%7.3%
.it5374.311.3%58.5%39.6%11.3%
.tv5170.213.7%49.0%41.2%3.9%

By site language

Language declared in <html lang>, for languages with at least 30 scored home pages (683 home pages declare none).

LanguageScoredMean scoreJS shellJSON-LDOpen Graphllms.txt
en2,95978.18.6%53.3%57.2%21.9%
fr71671.224.3%46.2%53.9%17.5%
ja16571.710.3%39.4%67.9%6.1%
ru15774.13.8%40.1%56.7%8.3%
de12978.23.9%55.0%45.7%11.6%
es12279.27.4%60.7%57.4%8.2%
zh8464.820.2%35.7%38.1%15.5%
pt7782.29.1%70.1%57.1%18.2%
it5578.73.6%63.6%47.3%7.3%
fa4177.67.3%58.5%34.1%7.3%
ar3680.42.8%61.1%52.8%13.9%
ko3467.817.6%26.5%70.6%14.7%
tr3181.63.2%58.1%61.3%29.0%

Most common schema.org types

WebSite 31.3%, Organization 27.6%, WebPage 15.9%, BreadcrumbList 10.2%, ImageObject 3.7%, NewsMediaOrganization 3.7%, FAQPage 3.3%, SoftwareApplication 3.2%, CollectionPage 3.1%, ItemList 3.0%, Corporation 2.2%, SiteNavigationElement 1.3%, Person 1.0%, VideoObject 0.9%, Article 0.9%

Beyond the home page

Our audit worker also crawls up to 30 inner pages of index domains when it is not busy with user requests: 4,624 sites and 123,973 inner pages so far (last 45 days). Average scanner score: 86.7 on home pages, 84.8 on inner pages.

  • 86.6% of inner pages serve at least 150 words without JavaScript.
  • 49.2% carry JSON-LD, 81.4% a meta description, 77.9% a canonical URL, 73.6% exactly one H1.
  • 2.6% of inner pages are set to noindex.
  • 95 of 4,624 sites block at least one AI search or browsing crawler on some inner pages, through path rules in robots.txt.
  • 1,757 sites have near-duplicate inner pages (90%+ identical text) among those sampled.

The 100 best-prepared home pages

Highest scores in the current snapshot (ties broken by Tranco rank). A high score means the home page is easy for AI crawlers to reach and understand; it says nothing about the quality of the site itself.

#DomainTranco rankScoreWordsJSON-LD typesllms.txt
1fastly.net401001,6732yes
2intuit.com981006967yes
3samsung.com1231002,2983yes
4forter.com1821001,27610yes
5hubspot.com2601001,8734yes
6centinel-analytica.com2711009394yes
7datadoghq.com2751001,9114yes
8unisys.com3491003,1271yes
9dynatrace.com3631006316yes
10atlassian.net3781001,3625yes
11gitlab.com3991001,1504yes
12businessinsider.com4901001,3632yes
13taboola.com5051007434yes
14atlassian.com5671001,1794yes
15wix.com5941001,3713yes
16optimizely.com6251003072yes
17supabase.co6841006754yes
18surveymonkey.com6851007541yes
19auth0.com7661001,9845yes
20statuspage.io7681005765yes
21cloudinary.com7921001,5944yes
22netlify.app7951008276yes
23mixpanel.com8341006263yes
24unpkg.com8711005953yes
25no-ip.com9221002,3001yes
26revenuecat.com1,0451001,1092yes
27netgear.com1,0601004592yes
28terra.com.br1,1051005863yes
29usercentrics.eu1,1061002,0903yes
30singular.net1,2461001,4836yes
31haberler.com1,3321002,4554yes
32notion.com1,3641003053yes
33adroll.com1,3931006012yes
34posthog.com1,4921007632yes
35markt.de1,5391001,0281yes
36therapservices.net1,5681001,4637yes
37surfshark.com1,8011001,2521yes
38rapid7.com1,8121001,1272yes
39hostinger.com1,9011002,4887yes
40intentiq.com1,9731001,3855yes
41crowdstrike.com2,1001002,9313yes
42utoronto.ca2,1021001,0692yes
43livechatinc.com2,1061001,1434yes
44yotpo.com2,1121001,9237yes
45railway.app2,1651001,0874yes
46notion.so2,1841003053yes
47mgid.com2,2471006751yes
48semrush.com2,3261001,3251yes
49jumpcloud.com2,3321009504yes
50ringcentral.com2,3331001,2552yes
51spot.im2,3461006654yes
52wpengine.com2,3771009025yes
53hubspot.net2,3831001,8734yes
54brightspace.com2,3931001,3915yes
55notion.site2,3991003053yes
56scw.cloud2,4141002,6784yes
57websupport.sk2,4271001,2163yes
58samsungelectronics.com2,4631002,2983yes
59urbanairship.com2,5101003,2116yes
60pubnative.net2,5571009965yes
61zscaler.com2,5921006995yes
62duosecurity.com2,6631001,1445yes
63geocomply.com2,6791001,0513yes
64rudderstack.com2,6831001,5742yes
65sportradar.com2,7311001,1204yes
66cookieyes.com2,7461007305yes
67savvis.net2,7831001,4691yes
68mountain.com2,8681001,7847yes
69cloudx.io2,8711003733yes
70cutt.ly2,8741003,6336yes
71redis.io2,9331001,1313yes
72phys.org2,9401003,1632yes
73clerk.com2,9961006992yes
74zapier.com3,0081001,6884yes
75a2hosting.com3,0211001,6641yes
76drpc.org3,0531008494yes
77elespanol.com3,1561006,1865yes
78elevenlabs.io3,1741001,3616yes
79proximus.be3,1791006862yes
80vtex.com3,1901001,1853yes
81sprinklr.com3,2221001,5321yes
82ably.io3,2231007626yes
83postman.com3,3151006432yes
84hotmart.com3,3441005027yes
85adsbynimbus.com3,3481005222yes
86amplify.com3,3791008956yes
87adyen.com3,4421005151yes
88demandbase.com3,4681001,0679yes
89sanity.io3,5481001,2672yes
90replit.com3,5861001,2361yes
91nanit.com3,6171001,0836yes
92razorpay.com3,6211003,4152yes
93owhealth.com3,6241001,1653yes
94opsgenie.com3,6531001,1235yes
95carfax.com3,6611003833yes
96netlify.com3,7381008276yes
97tanium.com3,7431006113yes
98colorlib.com3,7571006783yes
99mygaru.com3,7691006172yes
100signifyd.com3,8051001,2535yes

Daily snapshots

DayScoredMeanJS shellJSON-LDllms.txt
November 16, 20265,58973.311.5%47.1%16.8%
November 15, 20265,51773.311.6%47.1%16.7%
November 14, 20265,43173.411.4%47.2%16.8%
November 13, 20265,44473.311.4%47.2%16.8%
November 12, 20265,44173.311.4%47.1%16.8%
November 11, 20265,44073.311.5%46.9%16.7%
November 10, 20265,44973.311.5%46.8%16.7%
November 9, 20265,46273.311.4%46.8%16.7%
November 8, 20264,56473.111.4%46.5%16.9%
November 7, 20263,66173.311.8%47.1%17.3%

Look up a domain

See how a domain of the top 10,000 scored in its latest audit, criterion by criterion. For any other site, run the AI visibility checker or a full-site audit.

Download the data

Aggregates and the top-100 leaderboard, under CC BY 4.0: credit “IABIG AI-Readiness Index, iabig.com” with a link to this page.

  • readiness.json — current snapshot (all aggregates, breakdowns, leaderboard) and daily history
  • readiness.csv — adoption of every signal by rank tranche

Methodology

Domains. The 10,000 registrable domains of the latest Tranco list used by our AI Crawler Census. Each day we audit 1/2 of them, taken across all ranks, so every domain is re-audited every 2 days and each day's snapshot covers every rank tranche. The figures above combine the latest audit of each domain from the last 14 days.

Requests. For each domain we read robots.txt (https, then https://www., then http) and fetch the home page only if robots.txt allows our crawler (iabig-readiness/1.0 (+https://iabig.com/ai-readiness-index/), token iabig-readiness, falling back to the * group), following up to 5 redirects and reading up to 2 MB of HTML. We then request /llms.txt and, if robots.txt declares no sitemap, /sitemap.xml. 12 domains at a time, from one server in Europe; pages are analysed as they were served to us, without running JavaScript, exactly like most AI crawlers see them. Bot-challenge pages (Cloudflare, Akamai, DataDome…) are excluded rather than scored.

Score (100 points).

  • AI search access (25 points) — share of the AI search and browsing crawlers we track (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot, Googlebot, Bingbot…) that robots.txt allows on the home page.
  • Server-rendered content (20 points) — readable words in the raw HTML, without JavaScript: 300+ = 20, 150+ = 12, 50+ = 6.
  • Title, description, canonical, language (12 points) — title of 10–70 characters (4, or 2 if present), meta description of 50–160 characters (4, or 2), canonical URL (2), <html lang> (2).
  • Open Graph (6 points) — og:title, og:description, og:image (2 each).
  • Structured data (15 points) — at least one valid JSON-LD block with a schema.org type (10), including an identity type such as Organization or WebSite (5).
  • Heading structure (10 points) — exactly one H1 (6, or 3 if several), two or more H2 (4, or 2 for one).
  • Sitemap and llms.txt (8 points) — sitemap declared in robots.txt or at /sitemap.xml (4), valid llms.txt (2) organised in sections with 5+ links (2).
  • Indexable (4 points) — no noindex in the robots meta tag or X-Robots-Tag header.

Grades: A ≥ 85, B ≥ 70, C ≥ 55, D ≥ 40, E below. The rubric measures how easily an AI crawler can reach and understand the home page, not the quality, honesty or usefulness of a site. robots.txt is parsed with the RFC 9309 parser of our robots.txt tester.

Limits. One page per domain (inner pages are sampled separately, see above); sites may serve different content by country, device or user agent; a JavaScript-only home page may still be indexed by crawlers that render JavaScript, such as Googlebot. No page content is stored: only the measured signals.

Improve your own score: run the AI visibility checker on a page, or a full-site audit of up to 30 pages, then read the guides on JavaScript sites and llms.txt. Follow robots.txt policy changes of top sites in the AI policy change tracker.