Skip to content

The Hub for Public Relations Expertise & Insights

News

High-authority outlets are no readier for AI agents than small ones, Connectively's scores of 908 sites show

PTPR ThriveNovember 18, 20265 min read

Domain Authority says almost nothing about AI agent readiness: Connectively scored 908 outlets, found a 0.08 correlation, and 57.5% scored below 40.

Domain Authority predicts almost nothing about AI agent readiness, Connectively found when it scored 908 publications that post journalist questions on its platform, in scans run from October 22 to November 2, 2026. Across the 906 with every score, the Pearson correlation between Domain Authority and agent readiness was 0.08.

Key points

  • Connectively, the journalist-request platform owned by Featured (which also publishes PR Thrive), scored 908 publications with an Agentic Score, a 0 to 100 measure from Vercel's Is Agentic tool.
  • The average score was 39.1 and the median 37; 57.5% scored below 40, and 5.2% scored 70 or higher.
  • Domain Authority and the Agentic Score correlated at 0.08, and Domain Rating at 0.10, while the two authority metrics correlated with each other at 0.87.
  • Of the 143 publications with a Domain Authority of 60 or higher, 79 (55.2%) scored below 40.
  • The figures leave out 44 publications Featured owns, so its own sites do not shape the result.

What is an Agentic Score?

It is a 0 to 100 rating of whether an AI agent can find a website, read its pages and act on what it finds. The score comes from Is Agentic, a free tool from Vercel whose scans are run by Ora against the open AgentReady standard. Essential checks carry 80 of the 100 points, Recommended checks 20, and bonus signals add up to 5.

Connectively treats 70 or higher as agent-ready and below 40 as not agent-ready. Those are its own working thresholds: Is Agentic publishes no grade bands and calls every score a snapshot.

Does Domain Authority predict agent readiness?

No. Publications with a Domain Authority of 70 or higher averaged 39.9, and those below 20 averaged 36.9, only three points apart on a 100-point scale. Ahrefs' Domain Rating fared no better: the 40 to 69 band averaged the highest score, 42.6, ahead of the top band at 40.0.

Moz Domain Authority Publications Average Agentic Score Score 70+ Below 40
0–19 355 36.9 2.5% 64.2%
20–39 265 39.7 3.4% 55.5%
40–69 177 42.2 11.3% 50.8%
70–100 110 39.9 8.2% 51.8%

Bands by Moz Domain Authority leave out the single publication without that score, and the correlations use only the 906 that carry all three scores.

Ahrefs Domain Rating Publications Average Agentic Score Score 70+ Below 40
0–19 312 36.3 1.9% 65.4%
20–39 200 39.0 6.0% 57.5%
40–69 201 42.6 6.0% 51.7%
70–100 194 40.0 8.8% 51.0%

Bands by Ahrefs Domain Rating leave out the single publication without a rating.

The report's explanation is that SEO metrics never look at what the score tests: crawler blocking, articles that need JavaScript to appear, and content kept behind a login or paywall.

Why do so many news sites score low?

Many large publishers block AI crawlers on purpose. The Reuters Institute found that by the end of 2023, 48% of the most widely used news websites across ten countries were blocking OpenAI's crawlers, and 79% of the top sites in the US were. Blocking AI user agents alone costs a site the first Essential check, Connectively says.

Content behind a paywall or sign-in fails another check. Articles loaded through scripts and challenge pages for non-browser traffic cost points too, and the report finds big commercial publishers use those far more often than small editorial sites do.

Does a low score mean AI will not cite a publication?

No. The report is explicit that the score measures readiness, not citations, and that authority and reach still decide much of what AI engines cite. It points to Featured's AI Citation Report, which found 34.5% of Perplexity citations went to domains under a Domain Authority of 40.

What does this mean for PR professionals?

Our analysis: A media list sorted by Domain Authority alone now answers one question, reach with people and search engines, and leaves a second unanswered. For clients whose goal is AI visibility, the readiness of the outlet matters as well.

  • Read the two scores together. In the report's own example, a question from a site with a Domain Authority of 20 and an Agentic Score of 85 reaches AI agents, while one from a site with a Domain Authority of 80 and a score of 15 reaches readers but turns agents away. Both are worth answering; they do different jobs.
  • Check the open questions. Of 192 publications with questions open on Connectively on November 2, 107 (55.7%) scored below 40 and 22 (11.5%) scored 70 or higher.
  • Audit the media list, not just the pitch. Our earlier coverage of why a media list for AI visibility has to get longer and of the media list as PR's least audited document sets out how.

Frequently asked questions

How can I see a publication's Agentic Score?

On Connectively's Questions page, open the sliders menu and tick Agentic Score and Domain Rating. Both columns sort with a click, and Connectively has a step-by-step guide.

How often are the scores refreshed?

Connectively refreshes every publication's Agentic Score, Domain Authority and Domain Rating every 15 days. Is Agentic runs a fresh scan whenever someone asks for a report and the stored one is more than six hours old.

Can I check my own website's score?

Yes. Is Agentic is free and scores any site you enter against the same checks.

Is the sample representative of all news media?

No, by the report's own description. The outlets that post on Connectively lean toward trade titles, niche sites and digital-first publishers open to expert input, and the report names none of them.

Sources