Skip to content
VectorCite
Free audit

The state of AEO, 2026

We graded the VectorCite Atlas corpus of 103 category-leading brands (AI labs, dev infra, SaaS, fintech, ecommerce, media) with the quick grader. Here’s what the live grades show, recomputed every time this page refreshes.

Published June 1, 2026. The figures on this page come from the live leaderboard.

The corpus at a glance

Graded
88 of 103 brands
Average
86/100 across the corpus
A grades
55 brands at the top
F grades
0 brands failing the basics
Wikidata
88 brands with a verified entity

What wins AEO in 2026

The short version, computed from the live grades every time this page refreshes.

  1. Structured data on the homepage

    66 of 88 graded brands (most) ship JSON-LD on their homepage. It is the cheapest way to state facts an engine can quote.

  2. Access for AI search crawlers

    83 of 88 let every AI search and answer crawler reach their homepage. 2 block at least one AI search crawler outright, which keeps that engine from indexing them at all.

  3. llms.txt: published, but not scored

    62 of 88 (most) publish an /llms.txt. Those that do average 91/100, against 76 for the rest. The grade treats it as informational (an SE Ranking study of about 300,000 domains found it doesn’t change how AI systems cite pages), so read any gap as association, not cause.

  4. Wikidata identity and readiness

    Brands with a verified Wikidata entity average 85/100; those without one average 93. The grade doesn’t score Wikidata, and here the brands without an entity score higher: an entity record is no proxy for technical readiness.

  5. Titles and descriptions

    64 of 88 keep the homepage title between 30 and 70 characters, and 51 of 88 keep the meta description between 80 and 160. Both take minutes to fix.

Methodology

We ran the VectorCite quick grader against each of the 103 brands in the Atlas corpus. The quick grader checks 8 of the highest-leverage AEO signals: HTTPS, robots.txt and AI-bot rules, llms.txt presence, sitemap.xml, og:image, JSON-LD schema, <title> length, and meta description quality. The full 47-signal rubric is at /rubric and as machine-readable JSON at /api/v1/rubric.

The Atlas is a curated list of category-leading brands across 7 categories: AI labs, dev infra, SaaS, fintech, ecommerce, media and creator tools. 88 of the 103 have a verified Wikidata entity. The full Atlas and its selection criteria are public at /atlas.

Category leaders

The top-graded brand in each Atlas category, with the category's average quick grade.

Category leaders on the quick grade
CategoryAverageLeader
AI labsFull board84/100mistral.ai A, 100/100
Dev infraFull board91/100supabase.com A, 100/100
SaaSFull board88/100notion.so A, 100/100
FintechFull board90/100plaid.com A, 100/100
EcommerceFull board75/100shopify.com A, 100/100
Media + researchFull board68/100stripe.com/blog A, 93/100
Creator toolsFull board93/100framer.com A, 100/100

What to ship this quarter

In order: the cheapest wins first.

  1. Audit your homepage on the rubric.

    Free, about half a minute. It targets the cheapest wins first. Run the audit.

  2. Add JSON-LD if you don’t have it.

    Use our schema generator. Article, Organization and FAQPage at minimum.

  3. Publish /llms.txt.

    Five minutes of work, and most of your competitors haven’t done it. Check yours with our validator.

  4. Get a Wikidata entry.

    It needs 2–3 press mentions as citations; submit at wikidata.org. 88 brands with verified entities are listed at /atlas.

  5. Name AI bots in robots.txt.

    Even if you’d allow them by default, naming them signals intent and moves the needle.

Cite this research

The leaderboard data is MIT-licensed and citable. It refreshes every 6 hours; the corpus grows monthly.

Pull the dataset

curl https://vectorcite.digiocular.com/api/v1/leaderboard
curl https://vectorcite.digiocular.com/api/v1/rubric

How to cite

VectorCite (2026). State of AEO 2026: open benchmark of 103 category-leading brands on the VectorCite rubric. Retrieved from https://vectorcite.digiocular.com/research/aeo-2026

Where do you stand?

Audit one page against a real AI answer, or get the same 8-check quick grade these brands got for your own domain.

Grade your domainRun a free audit