State of AI Visibility · 2026

We scanned 38 leading online stores. 24% are invisible to AI search.

When a shopper asks ChatGPT, Claude, or Perplexity for what they sell, most of these stores can’t be read or recommended — they ship none of the files AI agents look for. The average store scores 31/100 today, but could reach 89 with a two-minute fix.

24%Invisible to AIScore 20 or below — effectively unreadable to AI shopping assistants.

71%Have any agent fileServe even one of agents.json, llms.txt, or agent-instructions.md.

31Average scoreOut of 100. The leaders reach the 80s; most sit far lower.

50%Have product schemaStructured data AI search can read directly off the page.

Where the stores land

score /100 · 38 stores

0–20
9
21–40
24
41–60
4
61–80
0
81–100
1

Furthest ahead

today → with a kit

1
Mejuri
mejuri.com
8595
2
MVMT
mvmt.com
6085
3
Alo Yoga
aloyoga.com
45100
4
Our Place
fromourplace.com
45100
5
Vessi
vessi.com
45100
6
Everlane
everlane.com
4095
7
Kosas
kosas.com
4095
8
Olipop
drinkolipop.com
4095
9
Outdoor Voices
outdoorvoices.com
4095
10
Ritual
ritual.com
4095

Where does your store land?

Free check, no signup. See your score in seconds — and the exact number you’d reach.

Scan my store

Method

We scanned the public homepages of 38 recognizable online stores with the same read-only checks the BridgeToAgent audit uses: whether each site serves agents.json, llms.txt, and agent-instructions.md, exposes schema.org product data, and is reachable by automated readers. “Potential” is the score each site would reach after adding those files — measured against the same checks, not a sales forecast. Illustrative; updated periodically. Last run June 1, 2026.