AI and answer engines problems on small business websites
12 checks in this group, each a count from our own scanner over 6,450 free four-page scans in the last 90 days. The most common is No article or breadcrumb markup, on 77.3% of them.
| Finding | Share of scans | Scans | Sites | Severity |
|---|---|---|---|---|
| No article or breadcrumb markup | 77.3% | 4,986 | 4,093 | Low |
| Page content only appears after scripts run | 76.5% | 4,936 | 4,077 | Serious |
| Page regions are not marked semantically | 52.2% | 3,370 | 2,793 | Warning |
| No business entity in structured data | 44.4% | 2,862 | 2,365 | Warning |
| No structured data | 41.2% | 2,658 | 2,199 | Serious |
| No plain summary for machines to lift | 29.7% | 1,915 | 1,570 | Warning |
| Home page is thin on content | 21% | 1,353 | 1,131 | Warning |
| Structured data missing a required property | 9.5% | 610 | 494 | Warning |
| Structured data that no longer earns a rich result | 2.9% | 184 | 146 | Low |
| Structured data that does not parse | 1.1% | 68 | 65 | Serious |
| AI crawlers are blocked in robots.txt | 0.8% | 54 | 47 | Warning |
| Not all structured data was checked | 0.6% | 38 | 33 | Low |
No article or breadcrumb markup
77.3% of the free scans raised this: 4,986 scans across 4,093 distinct sites, 95% interval 76.1% to 78.4%. Severity: Low.
What it means
Article and breadcrumb structured data help an answer engine place a page within the site and credit its authorship.
How to fix it
Mark up long-form pages as Article and give every page below the home page a BreadcrumbList, so an engine knows what a page IS and how a reader would reach it.
How to check it yourself
View source on a long-form page and search for the words Article or BreadcrumbList inside a script tag. If neither appears, this markup is missing.
What fixed looks like
View source shows an Article or BreadcrumbList block naming the page and its place in the site.
Page content only appears after scripts run
76.5% of the free scans raised this: 4,936 scans across 4,077 distinct sites, 95% interval 75.3% to 77.7%. Severity: Serious.
What it means
The HTML the server sends carries almost no readable text, so a reader that does not run scripts sees an empty shell. Search crawlers usually render the page eventually; most AI answer engines read what the server sent, and here that is close to nothing.
How to fix it
Server-render or pre-render the pages that describe the business, so the positioning copy is in the first response. Put the actual words in the page, not only in images or scripts.
How to check it yourself
View source, not DevTools, on the page and search for your key sentences. If the visible text is missing from that raw HTML, it only appears after scripts run.
What fixed looks like
View source on the page shows the positioning copy already present in the raw HTML.
Page regions are not marked semantically
52.2% of the free scans raised this: 3,370 scans across 2,793 distinct sites, 95% interval 50.6% to 53.9%. Severity: Warning.
What it means
Without semantic regions (header, nav, main, footer), assistive technology and machine readers both have to guess at a page’s structure.
How to fix it
Wrap the page in real landmarks: <header>, <nav>, <main>, <footer>. A page built from generic <div>s gives assistive technology nothing to jump between.
How to check it yourself
Open DevTools, go to Elements, and check whether the page uses header, nav, main and footer tags or only generic div elements throughout.
What fixed looks like
The page markup uses header, nav, main and footer elements instead of generic divs.
No business entity in structured data
44.4% of the free scans raised this: 2,862 scans across 2,365 distinct sites, 95% interval 42.6% to 46.2%. Severity: Warning.
What it means
An Organization or LocalBusiness block tells an AI what the site is, what it sells and how to reach it, before it will name the business in an answer.
How to fix it
Add a LocalBusiness (or Organization) JSON-LD block with name, url, telephone and address. This is the block that tells an answer engine who you are.
How to check it yourself
View source and search for LocalBusiness or Organization inside a script type application/ld+json block. If neither appears, no business entity is declared.
What fixed looks like
View source shows a LocalBusiness or Organization block with name, url, telephone and address.
No structured data
41.2% of the free scans raised this: 2,658 scans across 2,199 distinct sites, 95% interval 39.4% to 43.1%. Severity: Serious.
What it means
Structured data (JSON-LD or similar) is the machine-readable version of a page: without it, an answer engine has to infer everything from prose.
How to fix it
Add JSON-LD structured data in a <script type="application/ld+json"> block. Start with the type that matches the page: LocalBusiness, Product, Article.
How to check it yourself
View source and search for script type="application/ld+json". If no such block appears anywhere in the page, no structured data is present.
What fixed looks like
View source shows at least one script type application/ld+json block matching the page's content.
No plain summary for machines to lift
29.7% of the free scans raised this: 1,915 scans across 1,570 distinct sites, 95% interval 27.7% to 31.8%. Severity: Warning.
What it means
A page that never states plainly what it is gives an answer engine nothing confident to quote.
How to fix it
Add one or two plain sentences near the top of the page stating what it is and what it offers, in ordinary prose rather than a slogan.
How to check it yourself
View source and read the first paragraph of visible text. If it never plainly states what the page or business is, there is nothing to quote.
What fixed looks like
The first paragraph in view source states plainly what the page is and what it offers.
Home page is thin on content
21% of the free scans raised this: 1,353 scans across 1,131 distinct sites, 95% interval 18.9% to 23.2%. Severity: Warning.
What it means
A page with little text gives an answer engine little to quote or summarize with confidence.
How to fix it
Link your important pages from the home page or main nav. A page reachable only by search is a page most crawlers will find late or not at all.
How to check it yourself
Load the home page and check whether your important pages are linked from it or the main navigation. Pages reachable only by search are not linked.
What fixed looks like
The home page and main navigation link directly to your important pages.
Structured data missing a required property
9.5% of the free scans raised this: 610 scans across 494 distinct sites, 95% interval 7.4% to 12%. Severity: Warning.
What it means
An entity marked up without the property a rich result depends on will not qualify for it, even though markup exists.
How to fix it
Fill in the required properties for the type you declared - a LocalBusiness needs at least name, address and telephone. A type with empty properties is ignored.
How to check it yourself
View source, find the JSON-LD block, and check whether it includes the properties its type requires, such as name, address and telephone for LocalBusiness.
What fixed looks like
View source shows the JSON-LD block with all the properties its declared type requires filled in.
Structured data that no longer earns a rich result
2.9% of the free scans raised this: 184 scans across 146 distinct sites, 95% interval 1.2% to 6.4%. Severity: Low.
What it means
The markup is present but no longer matches what the target rich result requires, so it goes unused.
How to fix it
Compare the JSON-LD block's properties against the current requirements for that rich result, and add or correct whatever no longer matches.
How to check it yourself
View source, find the JSON-LD block, and check its properties against what that markup type currently requires for a rich result.
What fixed looks like
View source shows a JSON-LD block whose properties match what its rich result currently requires.
Structured data that does not parse
1.1% of the free scans raised this: 68 scans across 65 distinct sites, 95% interval 0.1% to 7.2%. Severity: Serious.
What it means
Malformed structured data is invisible to anything reading it: a syntax error is the same as having none.
How to fix it
Fix the JSON-LD so it parses: a trailing comma or an unescaped quote makes the whole block invisible. Paste it into Google's Rich Results Test to see it fail.
How to check it yourself
View source, copy the JSON-LD block, and paste it into the DevTools console with JSON.parse. An error means it does not parse.
What fixed looks like
The JSON-LD block parses cleanly in the DevTools console with no error from JSON.parse.
AI crawlers are blocked in robots.txt
0.8% of the free scans raised this: 54 scans across 47 distinct sites, 95% interval 0.1% to 8.1%. Severity: Warning.
What it means
A robots.txt rule blocking AI crawlers keeps the site out of anything those systems answer from, whether or not that was deliberate.
How to fix it
Allow the answer-engine crawlers in robots.txt (GPTBot, ClaudeBot, PerplexityBot, Google-Extended) if you want to appear in AI answers. Blocking them is a choice worth making deliberately.
How to check it yourself
Open example.com/robots.txt and look for Disallow lines naming AI or answer-engine crawlers specifically, rather than a blanket rule for all bots.
What fixed looks like
Opening example.com/robots.txt shows no Disallow line naming AI or answer-engine crawlers.
Not all structured data was checked
0.6% of the free scans raised this: 38 scans across 33 distinct sites, 95% interval 0% to 10.2%. Severity: Low.
What it means
Some markup on the page could not be fully evaluated in this pass.
How to fix it
Review every JSON-LD block on the page against the schema.org type it declares, checking required properties and valid syntax by hand.
How to check it yourself
View source and count the JSON-LD blocks on the page, then check each one individually against the schema.org type it declares.
What fixed looks like
View source shows every JSON-LD block on the page reviewed individually against its declared type.
How we measured this
A count of what our own scanner found, not a survey and not an estimate: each share above is over every free four-page scan in the last 90 days that loaded at least one page, which is 6,450 scans. Of those, 6,316 were run by Go Voltic itself, our research sweep over a pool of small-business sites and our own monitoring, and 132 came from visitors using the free scan; a site counts once in the distinct-sites figure whoever scanned it. 2 predate the label that tells the two apart and are counted in the total without a side. Deep scans are a separate population and never in this denominator. Figures on this page were computed on 2026-10-15.
The methodology explains the denominator, the distinct-sites count and the 95% interval. The benchmarks page puts these shares beside every other finding, with the whole table as CSV and JSON.
Other groups of findings
Measured on the same scans: Structure, Trust and privacy, Accessibility, Speed.
Check your own site
The free scan reads four pages and reports every finding it raises, including every one on this page. It takes about ten seconds and asks for nothing but the address.