FAQ

It can, and legibility is the part of that you control directly. What an assistant recommends depends on several things at once: authority, citations, freshness, what other sources say about you, and how closely you match the question asked. Being readable does not win the recommendation on its own. It is a prerequisite, because nothing can be quoted that cannot be parsed. Visibility Mesh measures how readable you are and shows you what to fix. Start with a free scan to see where you stand.

Run a free scan. In a few minutes you get a 0 to 100 legibility score, your maturity band, and your top issues with fixes. It reads your website the way an AI crawler does, so you see what the machines see, not a guess.

Visibility Mesh makes websites legible to AI. We measure what AI can understand on your site, show you what it cannot, and scan again after each fix so you can see what changed. On Shopify the app writes those fixes on its Growth plan, and you approve each one. On any other platform you get the measurement and the ordered roadmap, in plain English, worst first.

Ahrefs and Semrush began as tools for ranking in classic search, and both now also track how brands appear in AI answers. Visibility Mesh measures what AI answer engines can read and trust on your website and fixes what stops them. Many teams keep their SEO tool and add Visibility Mesh, since the two answer different questions.

Run a free scan at visibilitymesh.com with your website URL and email. You get your score and top fixes by email, and you choose a plan only if you want the full picture. You do not need a credit card to start.

It is for businesses whose buyers find them through search and AI, including ecommerce and DTC stores, the marketing and IT teams who share the work, and agencies that run it for clients. The scan works on any website, and Shopify stores get the deepest analysis. The platform support page shows what we do on each platform.

AEO, Answer Engine Optimization, is making your website readable and quotable to the AI systems that answer shoppers directly instead of showing a page of links. If an assistant cannot read your pages, it cannot mention you when a shopper asks.

For Google, mostly yes. Google's guide to generative AI features in Search, updated July 10, 2026, says that optimizing for generative AI search is still SEO. ChatGPT, Perplexity and Claude send their own search crawlers, so AEO adds a check that each of them can reach, read and quote your pages. The SEO foundation comes first, and the comparison of SEO, AEO and backlinks explains the order.

Make it machine readable. Allow AI crawlers, add complete structured data, write product and brand copy that answers first, and link your pages so an AI can follow them. A free scan tells you which of these you are missing, in priority order.

Give the AI what it needs to trust and quote you: complete, structured product data (title, description, attributes, price, availability), a clear brand identity, and consistent facts across pages. We score each of these and show you the gaps.

There are several possible reasons, and legibility is the one to rule out first, because it is the one you can act on this week: blocked crawlers, missing or incomplete schema, thin descriptions, weak internal links, or an unclear brand identity. It is not the only cause. Authority, corroboration from other sources, freshness, and how closely you match the question all matter too, and those move slowly. Guessing which of them is your problem wastes weeks. A scan tells you what is measurably wrong on your website today.

Be the clearest, most complete and most machine readable source for your products and topic. AI can only quote what it can parse, so start there. We measure how readable your pages are today and list the specific changes that raise the score.

Fix the highest impact issues from your scan first, then scan again to see what changed. The fastest wins are usually crawler access, complete structured data, copy that answers first, and internal linking. The report puts them in priority order.

The fixes cover three things. Structure lets an AI system tell what each page is. Content puts the answers a shopper asks for on the page in plain words. Links connect your organization, products, articles, FAQs and breadcrumbs with consistent identifiers, so AI sees one business instead of scattered pages. On Shopify, the app applies approved fixes on its Growth plan. On Shopify and WordPress, an AI Visibility Setup builds the connected structure by hand. On other platforms your developer applies them from the fix plan.

It is shopping done through AI agents that browse, compare and buy for a shopper. Agents reward sites with clean, complete, machine readable products and policies. If your data is messy, an agent may pass over you. The scan shows the product and policy data an agent would need, and the fix plan shows what to change.

Give agents complete structured product data, clear pricing and availability, readable policies and a consistent brand identity. Run a scan to see which of that data is missing or hard to read on your website.

Visibility Mesh runs two different checks. The scan reads your site the way AI crawlers do and checks, crawler by crawler, whether your robots.txt lets in the crawlers of OpenAI, Anthropic and Perplexity. The AI mentions check asks ChatGPT, Claude and Perplexity your questions and records who they name. Gemini and Google AI Overviews are not checked today, and both are listed as Building on our public roadmap.

It does when you start a check. The AI mentions check asks your questions of ChatGPT and Claude with web search and of Perplexity, and it records whether your brand is named. You start it yourself in the customer portal, and the Shopify app runs it as AI Mentions. Nothing runs on a schedule yet, and Gemini and Google AI Overviews are not checked. The scan itself measures something different: how completely AI crawlers can read your website.

For Perplexity, allow PerplexityBot in robots.txt and let Perplexity's published IP addresses through any firewall. Perplexity says PerplexityBot is the crawler that surfaces and links websites in its search results. For Google, keep your pages indexed and eligible to be shown with a snippet. Google says there are no additional requirements to appear in AI Overviews or AI Mode, and that the best practices for SEO still apply. The guide to how each AI finds your store goes platform by platform.

Let the search crawlers in if you want to be found. OpenAI says OAI-SearchBot is the crawler that lets a site appear in ChatGPT search, while GPTBot collects content for training, so blocking GPTBot does not take you out of ChatGPT search. Whether to allow training crawlers is a separate business choice. Our scan checks eight AI crawlers one at a time, GPTBot and OAI-SearchBot among them, and shows which ones your robots.txt blocks. The guide to robots.txt for AI crawlers has the full decision.

Start with a scan to get a prioritized list, then work top down. Most fixes are completing structured data, tightening titles and descriptions, fixing canonicals, and improving internal links. Every fix we give comes with the steps and a way to verify it worked.

Shopify themes usually emit product schema already; the gaps are missing fields like availability, price validity, or reviews. Our detailed audit shows the exact fields you are missing and where to add them, so you are not guessing.

llms.txt is a short plain text guide at your domain root that tells AI systems what your site is and where the important pages are. It is optional. Google says Google Search does not use it and that it neither helps nor harms visibility there. We have not found an AI company that documents reading it either, so treat it as housekeeping. The article on llms.txt explains more.

robots.txt tells crawlers, including AI crawlers, what they may read. Crawlers that follow robots.txt will not fetch what it blocks. Blocking a training crawler such as GPTBot does not remove you from ChatGPT search, and fetchers that act on a user's request, such as ChatGPT-User and Perplexity-User, may not follow it at all. Our scan reports, for each of eight AI crawlers from OpenAI, Anthropic and Perplexity, whether it is allowed or blocked. It reports Google-Extended on its own, as a control for Gemini training and grounding. The robots.txt guide has the detail.

Lead with one clear sentence stating what the product is and who it is for, then cover attributes, materials, sizing, and use cases in plain language. Avoid burying the facts under marketing copy. AI quotes the facts it can extract quickly.

Link related products, collections and educational pages to each other with descriptive anchor text, so an AI can follow the relationships. An AI system has a harder time placing a page that nothing else links to.

Publish complete Organization structured data (legal name, address, founder, and links to authoritative profiles), keep your details consistent across pages, and connect a real author identity. Give the company one Organization entity with a stable @id and sameAs links to the profiles you control, and make sure those profiles state the same facts and link back. Markup alone does not guarantee recognition, since search engines and assistants also weigh what other sources say about you. We measure your entity completeness and tell you what is missing, and the page on entity recognition explains the idea.

A common cause is boilerplate, like a country and currency selector, dominating the page text and leaving little real content for an AI to read. A descriptive H1 and clear homepage copy fix it. Your scan will tell you if this is the issue.

It can. Slow pages and poor Core Web Vitals make crawling and rendering less reliable, which weakens your technical foundations. We factor measured performance into the score so you know if speed is holding you back.

Write real questions as headings and answer each one in its first sentence, in plain text on the page. FAQPage markup is optional: Google stopped showing FAQ rich results on May 7, 2026, and we have not found an AI company that documents using the markup.

It does. A canonical tag is a strong signal that tells Google which URL to index, so a tag that points elsewhere can keep the right page out of search. Check the Google selected canonical in URL Inspection in Search Console, fix the template or app that writes the tag, and point redirects, the sitemap and internal links at the same URL.

You may not. The findings are written in plain English for you, and the project plan turns them into developer tasks with steps and acceptance checks. You can do the work yourself or hand it off. On Shopify, the app on its Growth plan can apply approved fixes for you.

You enter your website URL and email, and we scan a representative sample of your key pages. Then we email you a scorecard with your maturity level and your top issues, each with fix steps. You do not need a credit card, an install or a call.

A scan usually takes a few minutes. You see live status while it runs, and your results arrive by email and in your account.

An AI visibility score is a number a vendor calculates, and each vendor measures something different. Some count how often assistants mention a brand. Ours measures how completely an AI crawler can read and understand your website, from 0 to 100 across five categories. It is a diagnostic of legibility, and it does not count how often AI names you. The AI Legibility Framework explains the method.

There are four bands. Starting Point (0 to 39) means AI struggles to read your website or place it as a trustworthy source. Developing (40 to 59) means AI crawlers can read parts of the website and important gaps remain. Established (60 to 79) means most of what we measure is readable to AI crawlers and a few fixes are still open. Leading (80 to 100) is the top band, where nearly everything we measure is readable to AI crawlers. The current spread of scores across our benchmark is on the Research and Press page.

A perfect 100 is not the goal. It is a moving target, because AI systems and standards keep changing. What matters is your band and the direction you are moving. We publish our own score in the open, and we measure our site by the same rules we use on yours.

A low score is a map of what to fix. It means there is readable content an AI is missing today, and every point is a specific fix you can make. Many sites start in Developing, and the value is knowing what to change and watching the number climb.

You will usually get a score within about 2 points of the first one. Two scans of the same unchanged site rarely return the exact same number, but in most cases they land that close together. If your score moves by more than that, look at what changed on your site since the last scan.

No, and be skeptical of anyone who does. Nobody controls what AI systems say. What we control is whether your website is fully legible to them. We measure that, show you what is broken and scan again after each fix so you can see what changed.

We score five categories: AI Readability, Semantic Architecture, Technical Foundations, Schema and Structured Data, and Authority and Operational Discipline. Together they answer whether an AI can read, follow, trust and cite your website.

Your website changes constantly with new products, theme updates and app installs, and each change can affect legibility. Scan again once you have changed something, so the new score tells you whether the change worked. If nothing has changed, a new scan usually lands within about 2 points of the last one and tells you very little.

Your next scan shows what changed on your pages. How soon AI systems pick up the changes depends on their own crawling schedules, and the full answer is on our page about growing traffic from search and AI assistants.

You can scan any public website, since we only read what is publicly visible. It is a fast way to benchmark yourself against the sites AI already favors.

Not while it is password protected, because the scanner only sees the password page. Remove the password or use a publicly reachable URL, then scan.

Every plan and one time assessment has a fixed price, and all of them are on the pricing page. Each plan is a one time purchase with optional monthly monitoring on its product page, and the free scan costs nothing. The AI Visibility Setup comes in two parts, Foundation and Schema Mesh. Each has a fixed price on its own product page. Monthly AI Visibility Management is priced by quote, and you can ask for one through the quote form. The Shopify app plans and their free trial are on the Shopify App Store listing.

Plans are measured in pages, not credits, so there is no conversion math. Essential reads 100 pages on every scan, Advanced 150 and Agency 400. Catalog assessments are sized by product count.

A one time assessment gives you the full picture once, which suits a project where you want to find and fix everything now. A subscription suits a website that keeps changing and needs reading again as it does. Start with the free scan, then pick based on whether this is a project or an ongoing discipline.

Run the free scan first. If you want the complete findings once, take the Full Assessment. If you publish often and want monitoring, choose a subscription sized to your page volume. Catalog assessments are for stores where the catalog is the business.

You can. The free scan is the preview: a real score, your maturity band, and your top issues with fix steps. The paid tiers expand that into the full findings, the complete fixes and a roadmap you can assign. You see what the findings look like before you spend anything.

It is free. You get a score and your top fixes at no cost, and you do not need a card. The free scan reads a sample of your key pages rather than the whole site, so the paid tiers read more of your website and give you the complete list with every fix.

Generic advice tells you what matters in general. We tell you what is wrong with your website specifically, in priority order, with steps and a rescan that shows what changed. You pay to stop guessing and to spend your fixing time where it moves the score.

Yes, on the Catalog and Agency products. After purchase you get a secure link to upload your product export (Matrixify or Shopify CSV or XLSX) and optionally your theme. It is optional; skip it and we assess your live site automatically.

The Catalog Assessment reads your Shopify or Matrixify product export and counts every product once. It reports how many products carry a description, an SEO title, an SEO description and an image, next to a scan of your live pages. It cannot read the exports of other platforms yet.

You can upload product data as CSV or XLSX (a Matrixify or Shopify export) and your theme as a .zip. We check every file for safety before we process it.

It depends on the tier. You get a scorecard report, a detailed audit with fixes and verification steps, and a project plan you can assign to a developer, and the higher tiers add an executive deck. Everything is in plain English and ordered by impact.

Results arrive by email and in your account. Report links are private and time limited for security, so your findings are not exposed publicly.

You can. The project plan is built as assignable tasks with owners and steps, so a developer can work straight from it without a translation layer.

We measure and tell you what to fix, with steps and verification, and we scan again so you can see what changed. On Shopify, our app can also make the changes on its Growth plan: you approve each one, we write it, and you can undo it in one click. On Shopify and WordPress, including WooCommerce stores, we can also make the changes ourselves and manage the work through Monthly AI Visibility Management, which is sold by quote. On both platforms, a done for you AI Visibility Setup builds the structure by hand. Each of its two parts has a fixed price on its own product page. On Webflow, custom built sites and every other platform, you or your developer make the changes from the plan.

You get a confirmation, your scan starts, and your deliverables arrive by email and in your account. A standard scan usually takes minutes, and a full assessment takes up to two business days. Catalog and Agency buyers also get a secure link to upload product data, so the assessment counts products as well as the pages we crawl: up to 1,000 or 5,000 on the Catalog Assessments and up to 10,000 on the Agency plan.

Subscriptions are billed monthly and renew automatically until you cancel. We charge a one time assessment once.

You can cancel at any time from your account at account.visibilitymesh.com or by emailing support@visibilitymesh.com. Your plan stays active through the end of the current paid period and does not renew. There are no retention hoops.

Deliverables are digital and produced on demand, so sales are generally final once a scan has run. If we fail to deliver, or a charge was made in error, email us within 14 days and we will run the scan again or refund where appropriate.

You will not. After you cancel, your plan runs to the end of the current paid period and is not renewed, and you are not charged again.

We keep the exposure small. We scan only what is publicly visible, uploads go to private storage with time limited access, results are stored to your account, and you can ask us to delete them at any time.

We access only what is publicly visible on your website, the same surface any visitor or AI crawler sees. For deep assessments you can choose to upload your own product or theme files.

You can. Email support@visibilitymesh.com and we will delete it.

We do not sell your data. We use it to produce your assessment and, in aggregate with nothing that identifies you, to build industry benchmarks. Your website is never identified in those.

You do not. The scan works on any public website, including WordPress, WooCommerce, Wix and custom sites. The optional deep uploads read Shopify exports, so Shopify stores get the most detailed analysis. The platform support page shows what we do on each platform.

It does. The scan of the live site works on any platform. Only the optional catalog and theme uploads are tailored to Shopify exports today.

We do. The scan reads any public site regardless of language or market, since legibility signals like structured data, clean architecture and crawler access work the same everywhere. Findings are delivered in English. The scan reads sites in any language, but the AI Visibility Setup and Monthly AI Visibility Management write all content in English only.

They can. Reviews and ratings, especially when marked up with structured data, are a trust signal AI systems read. They strengthen the authority side of your legibility, which is often the hardest part to earn.

We scan our own site with our own engine and publish the result each month on the Our Own Score page. You should be able to see how our own website scores before you trust the score on yours.

You should not have to take our word for it. We publish our own score every month, and we say how the score behaves: scan the same unchanged site twice and the two scores usually land within about 2 points of each other. Every figure we publish comes with its method, population and date, and we do not report metrics we will not stand behind.

We do. The Agency plan reads 400 pages on every scan, assesses up to 10,000 products from a product export, and returns an executive deck you can take into a client meeting. White label and a client dashboard are not available yet, so the deliverable carries our name. If those matter to you, talk to us and we will tell you where they are.

Email support@visibilitymesh.com, or use the form on our contact page. Our team replies within one business day.

No website in our live benchmark has reached the Leading band so far. As of September 2026, there are 754 websites in the benchmark, and the highest score any of them has reached is 71.5. The Leading band starts at 80. We publish that because it is the difference between a measurement and a checklist: a checklist is finished when you run out of boxes, and this one is not finished for anybody yet. It also means the field in front of you is wide open.

The median site scores 43.8 out of 100, and 67% score below 50. The top 10% start at 58.3. That is measured across 754 real websites as of September 2026, ecommerce and not, scored by the same engine that scores yours, so it is a comparison rather than an opinion. Most sites are far less readable to AI than their owners expect, which is why the first scan is usually the surprising one.

We will not. We never write a value we did not read on your website. Where a field cannot be filled from your real content we leave it out, because markup that contradicts your page is worse than no markup, and it is visible to everyone in your search results. A rating you did not earn is the fastest way to lose the trust you are paying us to build.

We count a visit by the referrer on the visit itself, classified on our server against known AI assistants and AI search engines. That tracking is part of the Shopify app on its Pro and Growth plans, where a theme app embed reports the visits. For a website on another platform, read AI referrals in your own Google Analytics, and the GA4 guide shows how. A visit that merely claims to have come from an assistant is never counted, since any script could send that. We would rather undercount, and tell you how to fix the undercount, than show you traffic your website never had. If the count is zero, we report zero.

There are four things we will never do. We never show you a score you can max out by ticking boxes, since that measures your effort rather than your website. We never invent a number to fill a gap. We do not hand you the same block of structured data we hand everyone else, because identical markup gives an answer engine nothing to choose on. And we never promise a ranking, a citation or revenue, since none of those are ours to promise.

Yes, Visibility Mesh is on the Shopify App Store. It runs the same engine as the scan on this website, so the number always matches, and it adds the things that need a connection to your store: it reads your real catalog, publishes structured data to your storefront, tracks visits arriving from AI assistants, and keeps your score history so you can see the work paying off. Plans and the free trial are on the Shopify App Store listing.

On the Growth plan, the app changes only the fields you approve, one at a time. Those are product and collection titles and descriptions, SEO titles and meta descriptions, image alt text and tags, plus an FAQ page or a blog post when you approve it. Every change names the finding it fixes, we log the before and after value, and you can undo any of it in one click. Nothing is written until you click approve.

It cannot edit theme files, robots.txt, llms.txt, URL redirects or navigation menus. On the Growth plan it writes the product and collection fields you approve, such as titles, descriptions, SEO fields, image alt text, tags, and care and fit facts, and it can publish FAQ pages and blog posts. Its structured data reaches your storefront only after you switch on the app embed in the theme editor. Theme, robots and redirect changes arrive as step by step instructions for you or your developer, or as part of an AI Visibility Setup.

Shopify only lets an app add anything to your storefront through a theme app embed. Until that is on, the structured data we generate is saved on our side and reaches none of your live pages. The app reads your live storefront and shows you the true state on every screen, so you are never guessing whether it worked. You can check it yourself at any time: view your page source and search for visibility-mesh-embed.

Only if you switch on a type your theme already covers. The app's structured data reaches your storefront through its theme app embed, and each type has its own switch. Product is off by default, because themes such as Dawn and Horizon already print Product markup on every product page, and an app embed cannot remove markup a theme prints. When you run the app's structured data check, it reads your live pages, tells you which switches to leave off and flags any type that appears twice. The app never invents ratings, prices or review counts to fill its markup.

It is the same score, from the same engine. The app never calculates a score of its own. It asks the same service this website asks, so the two cannot drift apart. Every scan is stamped with the version of the methodology that produced it, so two of your own scores are always compared on the same rules.