Why AI-Built Dental Websites Don't Get Found, and How to Fix Them

TL;DR: A new or startup dentist should publish the practice website on a domain the practice owns, with a sitemap, pages search engines are allowed to crawl, Dentist structured data, and the same name, address, and phone on the site and the Google Business Profile. An AI website builder can produce the look quickly, and those checks still have to be done, because a site that looks finished can stay hard to find. In a hypothetical model built on cited industry averages, not a forecast, the 12-month gross-production gap is $41,785, and results vary.

Why does a beautiful dental website still get no patients?

A website has one job before any other: to be found. A good-looking site matters after someone lands on it, and getting them there is the hard part. A patient search can return ads, a map pack, an AI answer, and a handful of organic results. A new practice is competing with offices that have had years to build those signals.

Google, and tools that answer from the public web, need a few things before they can find a site at all. They have to be allowed to crawl it, able to read it, confident about what the business is and where it is, and given content that is actually useful. A generator is built to make a site look finished. The crawl file, the schema, and the matching phone number are separate jobs, and the owner usually does not know to ask.

Almost every problem below can be checked in a few minutes. Many of them can be fixed without rebuilding the site.

What often gets missed on a DIY or generated site?

The patterns below show up on do-it-yourself sites and on sites a generator drafted. Some are platform limits. Most are settings nobody turned on, or steps nobody knew to take. They are not a verdict on any product. Where a company documents its own behavior, that note is quoted as a fact to check, not as a score.

A few platforms publish the relevant details in their own docs:

• Wix's help center says Wix updates the sitemap automatically, and that you can open it by adding sitemap.xml to the domain. Wix's SEO page says the infrastructure includes server-side rendering and that Wix creates and maintains the XML sitemap.
• Squarespace's help center says Squarespace generates a sitemap automatically for a public site, on a custom domain or on a Squarespace subdomain. The same page says a sitemap is not available for a trial site, and that a private or password-protected site has to be set to public before the file loads.
• Framer's help center says Framer generates a sitemap for every published site, with no manual setup, including on a Framer subdomain.
• Lovable's hosting docs say apps created from May 13, 2026 use server-side rendering, so visitors and crawlers receive rendered pages. The same page says older apps serve rendered pages to verified search engines, social preview bots, and AI crawlers. Lovable's upgrade docs say an older React and Vite app sends the browser a mostly empty page, and that Lovable serves pre-rendered copies to verified crawlers. Tools outside that list can see very little.

None of those sentences is a ranking result. Product behavior changes. Check the live domain.

Is there a sitemap Google can read?

A sitemap is a file listing the pages you want search engines to know about. Google's sitemap documentation says sitemaps help crawl efficiency and are useful for large or new sites. A new practice, with few links pointing at it, is the second case.

• Check: Visit yourdomain.com/sitemap.xml. An error or an empty file means you do not have one Google can use.
• Fix: If the platform documents an automatic sitemap, confirm the file loads on the domain you actually use, then submit it in Google Search Console. If the site was generated as code and no sitemap exists, add one and submit it.

Is robots.txt or a noindex tag blocking Google?

robots.txt tells crawlers which URLs they may request. A noindex tag tells Google not to show a page in search results. Draft and staging setups often include one or both, and they can still be there on launch day. One blanket disallow can hide an entire site. Google's JavaScript guide notes that if robots.txt disallows a URL, Googlebot skips the request and will not render JavaScript on that page.

• Check: Open yourdomain.com/robots.txt and look for Disallow: /. In Search Console, inspect the homepage. It will say whether indexing is allowed.
• Fix: Remove a blanket disallow, turn off any "hide from search engines" setting, and request indexing in Search Console.

Does the site use schema markup?

Schema is a small block of code that tells a machine, in a shared vocabulary, what the page is. Google's local business documentation says to use the most specific LocalBusiness subtype you can. Schema.org lists Dentist as that subtype for a dental practice. The useful fields are the exact name, address, phone, hours, and geo coordinates, plus a link to the Google Business Profile. Many generated sites ship with none of this, or with only a generic website type. That is a gap to check, not a claim about every builder.

• Check: Paste the URL into Google's Rich Results Test or the Schema Markup Validator.
• Fix: Add Dentist markup that matches the site and the Google Business Profile character for character. FAQ markup is still valid code that helps a machine read questions and answers. Google's August 2023 note says FAQ rich results are shown only for well-known, authoritative government and health websites, so a typical practice site should not expect the visual extras.

Are the name, address, and phone consistent everywhere?

This is often called NAP consistency. If the site says "Suite 200," the Google Business Profile says "Ste. 200," and a directory still shows an old phone number, a search engine has less reason to treat the listings as one business. Generated footers, contact pages, and schema often disagree with each other on the same site.

• Check: Compare the header, footer, contact page, schema, Google Business Profile, and the directories you actually use, such as Yelp, Healthgrades, Zocdoc, and the ADA Find-a-Dentist listing.
• Fix: Pick one exact format and use it everywhere, including the punctuation. Prefer the practice's own local number over a call-tracking number that changes.

Are the title tags and meta descriptions unique?

The title tag is often the blue headline in Google. The meta description is the snippet under it. Generated sites sometimes leave every page titled "Home," or repeat one description across the site. Google's title link documentation says Google may build that link from other text on the page, such as a prominent heading, when the title is a poor fit. Duplicate titles also make it harder to tell your pages apart.

• Check: Search site:yourdomain.com and read the results. Repeated titles are easy to spot.
• Fix: Give each page a unique title, under about 60 characters, that names the service and the city. Write the description in plain English about what that page offers.

Did the build create a lot of thin pages?

Some tools will draft many near-identical pages in one sitting, such as "dentist in" every nearby suburb, or a stack of posts nobody has read. Google's spam policies call this scaled content abuse: many pages generated for the primary purpose of manipulating rankings and not helping users, no matter how they are created. One listed example is using generative AI tools to generate many pages without adding value. Google's people-first content guide looks at whether the site as a whole is helpful, so a large amount of low-value content can drag on the pages you do want found. Google's 2023 guidance on AI content says appropriate use of automation, including AI, is not against its guidelines. The problem is unreviewed volume, not the tool.

• Check: Count the pages. Many that say the same thing with a different town name, or posts you have not read, are the flag.
• Fix: Merge or delete thin pages. Keep a smaller set you would show a patient. AI is a reasonable drafting aid when a person reviews it.

Is the site accessible?

Accessibility means people using screen readers, keyboards, or zoom can use the site. It matters for patients and for legal risk, and many of the same fixes help a crawler read the page. The standard is WCAG 2.2. The contrast criterion asks for at least 4.5:1 for normal text.

Gaps that show up often on generated sites:

• Images with no alt text
• No "skip to content" link for keyboard users
• Light gray text on white that fails the 4.5:1 contrast minimum
• Headings chosen for size rather than order
• Form fields with placeholder text and no proper label

• Check: Run WAVE or Lighthouse, in Chrome DevTools, on the main pages.
• Fix: Most of these are settings or short edits inside the builder.

Is the site fast enough on a phone?

Core Web Vitals are three public measures of how a page feels. Largest Contentful Paint, a good experience within 2.5 seconds. Interaction to Next Paint, 200 milliseconds or less. Cumulative Layout Shift, 0.1 or less. Read them at the 75th percentile of real visits. Large hero videos, full-resolution photos, and stacked animation are common reasons a generated homepage misses those marks.

• Check: Run the homepage through PageSpeed Insights on mobile.
• Fix: Compress images, serve a modern format such as WebP, trim autoplay video and animation, and limit third-party widgets.

Can a crawler see the content without running JavaScript?

Some generated sites build the page in the visitor's browser. Google can run JavaScript. Its JavaScript guide says it queues pages for rendering after the crawl, and that the wait can be more than a few seconds. The same page says the initial HTML of an app-shell site may not contain the content, and that not all bots can run JavaScript. Server-side rendering or pre-rendering still helps users and crawlers. Whether a given product does that depends on the version. The Lovable notes above are one current example, and they are not a description of every generator.

• Check: Right-click the page and choose "View page source." If the services and the address are not in that text, some crawlers will not see them either.
• Fix: Serve complete HTML, either rendered on the server or pre-rendered, so the content is in the response.

Other checks worth ten minutes

• Canonical tags. A canonical tag names the URL you want treated as the main copy when duplicates exist. Without it, signals can split across copies.
• Search Console and Analytics. Without them you cannot see whether Google is indexing the site, which queries it appears for, or whether visitors book. Set both up on day one. The practice should hold the Search Console owner login.
• A link to the Google Business Profile. The profile powers the map pack. Link to it from the site, and link the site from the profile, with matching details.
• HTTPS and one hostname. The site should load on https://, and www and the bare domain should both land on one version.
• One page per major service. A single "Services" page that lists everything has little to rank for "dental implants" plus the city. Each major service deserves its own page.
• Local details in text. Mention the city and the neighborhood in ordinary sentences. Keep the address in text, not only inside an image. A map, parking notes, and nearby landmarks help a person and a machine.
• No orphan pages. A page with no internal links is hard for Google to find. Link every page from at least one other page.
• Real photos. Patients choose a person. Photos of the dentist, the team, and the office give a visitor, and an assistant summarizing the page, something specific to describe.
• llms.txt, optional. llmstxt.org proposes a plain-text summary for language models. It is not a standard. Google's guide to generative AI features says Google Search does not use these files, including for AI Overviews and AI Mode. Treat one as a bonus, not a ranking fix.
• A domain you own. A site that lives only on a builder subdomain, or a domain registered in a vendor's or a former employee's name, puts the work at risk. Register the domain in the practice's name, at a registrar the practice controls.

How much does page 1 versus page 2 matter in this model?

To make the checklist concrete, here is a model of two hypothetical new practices opening at the same time in the same mid-size US market. The search volumes use the Boise, Idaho market. This is a hypothetical model built on cited industry averages. It is not a study of real practices, it names no website company, and it is not a guarantee. Results vary. The position paths are assumed, not measured. The numbers were chosen to stay conservative, and then priced with a higher cited production figure.

• Practice A gets the technical groundwork in place: crawlable, fast, schema, consistent name, address, and phone, individual service pages, and a connected Google Business Profile. In the model it reaches page 1 for core searches around month 5, and position 3 by month 11.
• Practice B is a do-it-yourself build that still has the gaps above. In the model it sits on page 2 or 3 all year.

The assumptions

• Search demand. The volumes are a Google Ads search-volume pull for Boise, Idaho, retrieved October 14, 2026, through DataForSEO. There is no public page for that pull. "dentist near me" and "dentists near me" are one cluster of about 1,900 monthly searches. "dentist boise" and "boise dentist" are one cluster of about 880. The model counts the 880 plus 30 percent of the near-me cluster, about 1,450 core searches a month, because near-me results depend on where the person is standing. Service terms in the same pull add about 570 searches a month. Only Practice A has service pages, and those ramp up over the year.
• Click-through by position. The visit counts were built from First Page Sage's click-through report as it stood for 2025, then cut in half. That page, checked October 14, 2026, still says blended position 1 fell from 39.8 percent in the 2025 report to 7.1 percent. The current blended column on the same page is about 7.1 percent at position 1, 1.7 percent at position 3, and 0.2 percent at position 10. The full 2025 position-by-position curve is no longer printed there, so this article does not restate those older rates as if they were on the page. Page 2 and page 3 rates in the model (about 0.6 percent for positions 11 to 20, and about 0.2 percent for positions 21 to 30) are assumptions. The page does not publish them. Every organic click figure was then cut in half as an allowance for ads, the map pack, and AI answers. The visit counts were not rebuilt from the newer blended column. Backlinko's study of about 4 million results puts position 1 at 27.6 percent, which is another reminder that click rates move around by study.
• Map pack. The model assumed a local-pack click rate in the mid-teens, counted only 40 percent of those clicks as website visits, and capped Practice A's pack visibility at 25 percent of core searches by year end (Practice B at 5 percent). First Page Sage's current table on the same report puts local pack positions at 14.8 percent, 12.6 percent, and 11.9 percent. The visits below were not rebuilt from that newer table.
• Visitor to new patient. First Page Sage's 2025 conversion report, using client data from 2017 to 2024, lists dentistry at 3.0 percent of visitors becoming prospective patients and 59.8 percent of those becoming patients. That is about 1.8 percent of visitors becoming new patients.
• Value per new patient. The dollar columns use $4,220 of gross production per new patient, from Dentistry IQ's June 12, 2017 report of Sikka Software data. Sikka's figures, from more than 12,500 practices, put average gross production per new patient who received a comprehensive exam at $4,220.25 in 2016. The article's note says the ratio uses unique patients who received a comprehensive exam in the six months before the current month. It is gross production, not money collected, and it is a 2016 average. This model books that $4,220 in the month the patient is acquired, as a stand-in for first-year production. That is a simplification, not a measured twelve-month cohort. Lifetime value is a separate assumption: $4,500 per patient, chosen well below the $12,000 to $15,000 range stated in this Dental Economics article. A widely repeated "$653 a year from ADA data" figure could not be traced to a primary ADA page, so it is not used. The $4,500 lifetime figure is only a little above one year at $4,220, which is intentional: the model does not stack later years of Sikka-level production on top. Do not add the two valuations together. They describe the same patients.

The 12-month comparison

Patient counts are the model's, rounded to one decimal in the table. Dollars are the finer count behind each month, times $4,220, rounded to the nearest dollar. A row will not always equal the printed patient count times $4,220. Practice B's rounded monthly patient counts add to 1.3; the model total is 1.2, and the total row uses 1.2. In month 1 the model has Practice B ahead by $50, because of a small assumed map-pack share. That first month is left in the table on purpose. On a phone, scroll the table sideways.

The simple math

• Year one. Practice A gains about 11.1 new patients from search in the model, and about $46,791 of first-year gross production. Practice B gains about 1.2 patients and about $5,006. The 12-month gap is $41,785.
• Lifetime, same patients, different valuation. At $4,500 per patient, the 9.9-patient gap is $44,550 (11.1 times $4,500 is $49,950, and 1.2 times $4,500 is $5,400). That is not money on top of the $41,785. It is a more conservative value for the same people, and it sits far below the $12,000 to $15,000 range the Dental Economics article states.
• The run rate is the larger story. By month 12, Practice A is adding about 2.0 search patients a month and Practice B about 0.2. That is 21.6 more patients a year. At $4,500, each future year at that pace is about $97,200 in lifetime patient value. At $4,220, the same 21.6 patients are about $91,152 of first-year gross production. Again, pick one valuation. Do not add them.
• A smaller slice, organic only. Inside the model, moving from position 11 to position 3 on the core local searches is about 70 more visitors and about 1.25 more new patients a month, before the map pack or the service pages. That is about 15 patients a year. At $4,220, 1.25 patients a month is about $5,275 of gross production, or about $63,300 over a year.

These figures cover a handful of core terms in one market. They cut organic clicks in half, they ignore referrals and word of mouth, and they ignore searches that never become a click because an AI answer was enough. A real practice will land somewhere else. If the newer First Page Sage blended click rates were used in place of the 2025 curve, the visit counts, and the dollars, would be smaller. This page does not make that swap. It prices the existing patient counts and labels the whole thing as a model.

Quick checklist: can this site be found?

• yourdomain.com/sitemap.xml loads, and it is submitted in Google Search Console.
• robots.txt does not block the site, and no page you want found is set to noindex.
• Dentist schema carries the exact name, address, phone, and hours.
• Name, address, and phone match on the site, the Google Business Profile, and the directories.
• Every page has a unique title tag and meta description.
• There is no pile of near-duplicate or unread generated pages.
• Alt text, a skip link, readable contrast, logical headings, and labeled form fields.
• The homepage passes Core Web Vitals on mobile in PageSpeed Insights.
• The services and the address appear in "View page source," not only after JavaScript runs.
• Canonical tags are in place, the site is HTTPS, and www and non-www resolve to one host.
• Google Search Console and Analytics are connected, and the practice owns Search Console.
• The site links to the Google Business Profile, and the profile links back.
• Each major service has its own page, with local details in text.
• Every page is linked from at least one other page.
• The photos are of the real team and the real office.
• Optional: an llms.txt file, knowing Google Search does not use it.
• The domain is registered in the practice's name, on an account the practice controls.

Can I fix a generated website myself?

Often, yes. Many items on the checklist are settings: turning off "hide from search," editing titles, adding alt text, compressing images, and connecting Search Console. Others, such as schema, rendering, redirects, or cleaning up dozens of thin pages, take more care. A mistake can do real harm, including accidentally telling Google to hide the whole site.

Questions new and startup dentists ask

Should a new dentist use an AI website builder?

A new dentist can. An AI website builder is a fast way to publish a site, and Google's February 2023 guidance says appropriate use of automation, including AI, is not against its guidelines. The site still needs the findability work: a sitemap, pages search engines are allowed to crawl, Dentist structured data, unique titles, and a name, address, and phone that match the Google Business Profile. A finished look does not do that work by itself.

What does a new dental website need to get found?

It needs a domain the practice controls, a sitemap submitted in Google Search Console, a robots.txt file that does not block the site, and no noindex tag on pages the practice wants found. It needs Dentist structured data with the exact name, address, phone, and hours, and those same details on the Google Business Profile. Each service the practice wants new patients to find should have its own page. Google's sitemap documentation says a sitemap is useful for a new site with few links pointing at it.

When should a startup practice hire a dental website company?

Hire one when the practice wants a site it owns, on a domain it controls, and does not want to maintain the technical work itself. Titles, alt text, and connecting Search Console are jobs many owners can do. Schema, how the page is rendered, redirects, and cleaning up a pile of thin generated pages are the jobs that usually need someone who does them regularly. This is a question of scope. It is not a price.

How long does it take a new dental website to rank on Google?

It varies by market and competition. In this hypothetical model, a new practice in a mid-size market reaches page 1 for some core local searches around month 5 when the technical groundwork is in place. Without it, the same model keeps the practice on page 2 or 3 all year. That is an illustration, not a timeline anyone can promise.

Will Google penalize a site for using AI-written content?

Not for using AI as such. Google's spam policies describe scaled content abuse as many pages generated for the primary purpose of manipulating rankings and not helping users, including with generative AI tools, when the pages add little or no value. Google's 2023 guidance says appropriate use of automation, including AI, is not against its guidelines.

How do AI assistants choose a dentist to mention?

They can use what they can read on the public web: the practice site, the Google Business Profile, reviews, and directories. A site that is crawlable, clearly structured, and consistent with those other listings is easier to describe accurately. This article does not describe any one product's ranking formula.

Does each service need its own page?

For services you want new patients to find, yes. Someone searching for dental implants in their city is looking for a page about implants. A single list of every service gives Google little to match.

Is the dollar gap in the model a promise?

No. The 12-month gap of $41,785 is from a hypothetical model of two practices, using that model's patient counts and a 2016 average of $4,220 in gross production per new patient. It is not a forecast for any real practice, and results vary.

Sources

Pages below were opened on October 14, 2026. A source that did not load, or that did not support the claim, is not in this list. The $700 to $1,250 "first-year spend" range and a "$653 a year" ADA figure were dropped for that reason.

1. Google Search Central, "Learn about sitemaps": developers.google.com/search/docs/crawling-indexing/sitemaps/overview
2. Google Search Central, "Introduction to robots.txt": developers.google.com/search/docs/crawling-indexing/robots/intro
3. Google Search Central, "Block search indexing with noindex": developers.google.com/search/docs/crawling-indexing/block-indexing
4. Google Search Central, "Local business structured data": developers.google.com/search/docs/appearance/structured-data/local-business
5. Schema.org, Dentist: schema.org/Dentist
6. Google Search Central blog, "Changes to HowTo and FAQ rich results" (August 2023): developers.google.com/search/blog/2023/08/howto-faq-changes
7. Google Search Central, "Influencing title links": developers.google.com/search/docs/appearance/title-link
8. Google Search Central, "Spam policies" (scaled content abuse): developers.google.com/search/docs/essentials/spam-policies#scaled-content
9. Google Search Central, "Creating helpful, reliable, people-first content": developers.google.com/search/docs/fundamentals/creating-helpful-content
10. Google Search Central, "Google Search's guidance about AI-generated content" (February 2023): developers.google.com/search/blog/2023/02/google-search-and-ai-content
11. Google Search Central, "JavaScript SEO basics": developers.google.com/search/docs/crawling-indexing/javascript/javascript-seo-basics
12. Google Search Central, "Consolidate duplicate URLs": developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls
13. web.dev, "Web Vitals": web.dev/articles/vitals
14. W3C, "Understanding SC 1.4.3 Contrast (Minimum)," WCAG 2.2: w3.org/WAI/WCAG22/Understanding/contrast-minimum.html
15. llms.txt proposal: llmstxt.org
16. Google Search Central, "AI optimization guide" (Google Search does not use llms.txt): developers.google.com/search/docs/fundamentals/ai-optimization-guide
17. First Page Sage, Google click-through rates by position (page checked October 14, 2026; the 2026 blended table and the note that 2025 position 1 was 39.8 percent): firstpagesage.com/reports/google-click-through-rates-ctrs-by-ranking-position
18. Backlinko, organic CTR study: backlinko.com/google-ctr-stats
19. First Page Sage, patient conversion rate by practice type (2025 report): firstpagesage.com/reports/patient-conversion-rate-by-practice-type
20. Dentistry IQ, Kevin Henry, "How much is a new patient worth to a dental practice?" (Sikka Software data, June 12, 2017): dentistryiq.com/.../how-much-is-a-new-patient-worth
21. Dental Economics, "Patient attrition: 3 steps to finding a hidden gold mine": dentaleconomics.com/.../patient-attrition
22. Wix Help, "Understanding your site's sitemap file": support.wix.com/.../understanding-your-sites-sitemap-file
23. Wix, SEO overview (server-side rendering and sitemap, as stated by Wix): wix.com/seo
24. Squarespace Help, "View your site map": support.squarespace.com/.../view-your-site-map
25. Framer Help, "How to access your sitemap": framer.com/help/articles/how-can-i-access-the-sitemap-xml-file
26. Lovable Docs, "How Lovable hosts your app": docs.lovable.dev/features/hosting
27. Lovable Docs, "Upgrade a project to TanStack Start": docs.lovable.dev/features/upgrade-to-tanstack-start

DataForSEO Google Ads search volume for Boise, Idaho, was retrieved October 14, 2026, for the demand inputs above. It is a data pull, not a public article, so it has no URL here.

If you would rather hand the technical work to someone else, Rainey Dental Partners builds custom dental websites with this groundwork included from the start.

Back to all articles