Research and Press
Research and Press
The State of AI Visibility 2026
Visibility Mesh measures how legible a website is to AI answer engines such as ChatGPT, Perplexity, and Google AI Overviews. Our first public study scanned 216 live Shopify storefronts and scored how readable each one is to those systems.
Key findings
- The median store scored 46 out of 100 on the AI Visibility Score.
- Two thirds of stores scored below 50.
- Only 27 of the 216 stores, one in eight or 12.5 percent, were fully legible to AI. Each of them passed all five measured layers.
- The single highest scoring store of 216 reached just 71. None reached the top maturity band.
- Entity Integrity, the machine readable product and brand data AI engines cite, was the weakest layer, failed by 72 percent of stores.
These are not bad stores. They are invisible ones. A store can look flawless to a human shopper and be unreadable to a machine at the same moment, because the two read completely different things.
Margareta Petrovic, Founder of Visibility Mesh
Read the full study
State of AI Visibility 2026, the full research report
Methodology in brief
All scores come from live scans, not cached data. Of 229 stores scanned, 13 were automatically diagnosed as unreadable and excluded from scoring rather than counted as zeros. Visibility Mesh scores are repeatable but not identical: when we scan the same unchanged store twice, the two results usually land within about 2 points of each other. Findings are a point in time snapshot. Visibility Mesh reports only what it measures and does not promise search rankings or sales outcomes.
The living benchmark
The 2026 study above is frozen, so its published figures never move. Behind it runs a benchmark that keeps growing as we scan. The per site scores are published as an open dataset, AI Legibility Scores (596 websites, CSV on GitHub, CC BY 4.0, refreshed with each release). These are its populations, stated separately, because they are not interchangeable and a number is only as good as the population it came from. All figures measured September 30, 2026.
- 952 websites scored to date. Every distinct site our engine has completed a scored scan of, across all versions of the methodology. Our own sites and development stores are excluded from every figure on this page.
- 754 websites in the live benchmark. The subset scored on the current methodology version, counted once each at their latest score. It mixes our research scans with scans that merchants ran through our Shopify app and a few deeper paid scans. This population spans ecommerce and non ecommerce websites.
- 431 of those 754 are online stores, and 377 of the stores are on Shopify, by platform detected during the scan rather than assumed. Four sites could not be classified and are counted only in the 754.
- 216 Shopify storefronts in the published study. The frozen press cohort above, and a separate population from the 377 detected stores in the live benchmark.
- 43.8 median score across the whole 754. 67 percent score below 50, and the top 10 percent start at 58.3. The online store sub-cohort runs slightly lower: a median of 40.5, with 76 percent below 50. In September 2026, most of that gap came from stores whose owners scanned them through our Shopify app, since those merchants start well below the stores we scanned for research.
- Nothing in the Leading band. The highest any site in the benchmark has reached is 71.5. Leading starts at 80.
We name the population next to every number because the same engine produces all of them, and an ecommerce median and an all web median are different claims. Where we cannot separate a population cleanly, we say so rather than pick the flattering label.
Methodology, versions and limits
Every score comes from a live automated scan of public pages, read the way AI crawlers read them, across five categories: whether AI can find a site, read it, follow it, quote it and trust it. The total is the share of available points, from 0 to 100. The five metrics are specified in the AI Legibility Framework v1.0, and the four bands in the AI Visibility Maturity Model.
The live benchmark counts only sites scored on the current methodology version, one latest score per site. Each published study was measured once, on its own date and its own population, and is not revised when the method changes. A study figure and a benchmark figure are two separate measurements, so read them side by side rather than as a trend.
A score measures legibility: what AI systems can read and trust on public pages at the time of the scan. It does not measure rankings, traffic, or how often an assistant names a site, and it is not a promise of any of them. We publish our own score under the same method.
Published research
- The State of AI Visibility 2026: 216 Shopify storefronts, published June 28, 2026. This is the frozen study summarized at the top of this page.
- The State of AI Visibility on the Non Ecommerce Web: 303 websites across 25 verticals, published July 13, 2026.
- We Scored 245 Shopify Stores on What AI Can Read: 245 Shopify storefronts, published August 22, 2026.
- AI Legibility Scores: the open dataset of 596 websites from the live benchmark, generated September 21, 2026, CC BY 4.0.
Independent product verification
The Visibility Mesh Shopify app is listed on the Shopify App Store. The listing names Amaretta LLC as the developer and shows that the app launched on July 10, 2026. Merchant ratings and reviews are collected and published by Shopify on that listing, so the listing is the source for any rating we quote, and its figures change as new reviews arrive. A listing on the Shopify App Store is not an endorsement by Shopify.
Founder columns
Margareta Petrovic, the founder of Visibility Mesh, writes for the Forbes Business Council on artificial intelligence, trust, and how businesses are read by the systems that now recommend them. These are member contributed columns that she wrote, not Forbes newsroom coverage of Visibility Mesh. Her background is on her author page and her Forbes Business Council profile.
- Before You Buy More Traffic, Fix The Proof, Forbes Business Council, September 23, 2026. The four question proof test behind it.
- Brand Visibility Mistakes That Can Erode Customer Loyalty, Forbes Business Council, September 11, 2026.
- Your E-Commerce Storefront May Look Perfect, But AI Likely Can't Read It, Forbes Business Council, August 4, 2026.
- AI Is Coming For Fake Value, Forbes Business Council, June 26, 2026.
Press release
- Only 13% of Shopify Stores Can Be Read by AI Search, New Study Finds, The National Law Review, press release on the 2026 study.
Other references
OngoingAI, an independent Shopify app intelligence site, lists the Visibility Mesh app with its developer, pricing, rating and category positions. OngoingAI tracks those figures daily, so they change often, and its listing is not a review or an endorsement.
About Visibility Mesh
Visibility Mesh gives organizations a single AI Visibility Score out of 100, the specific gaps, and the steps to fix them. Visibility Mesh is a brand of Amaretta LLC, which operates this website, the web application and the Shopify app. Margareta Petrovic is the founder.
Media contact
Margareta Petrovic, Founder
press@visibilitymesh.com
High resolution charts and media assets available on request.
Get your own score
Any website owner can get a free AI Visibility Score at visibilitymesh.com.