MFG Wellness | Where Healthcare Meets Marketing
A practice can have a good-looking website, a decent Google ranking, and still be invisible when a patient asks an AI assistant about care options nearby. Those are different systems with different requirements, and the gap between them catches a lot of practices by surprise.
The good news is that the reasons are mostly mundane, and Google has now published documentation explaining several of them.
Reason one: the crawlers cannot get in
An engine cannot use a page it cannot read. Google’s own guidance on AI features and your website lists allowing crawling in robots.txt, and by any CDN or hosting infrastructure, as a starting requirement.
That second clause catches people. A robots.txt file can be perfectly permissive while a firewall, a security plugin, or a content delivery network quietly blocks automated traffic. Nobody notices, because human visitors are unaffected.
Different engines use different crawlers, and each is allowed or blocked separately. A site can be readable by one and closed to another without anyone having made that decision on purpose.
Reason two: the page was never indexed
Google is explicit that meeting the requirements does not promise anything. Its documentation states plainly that indexing and serving are not guaranteed, even for a page that follows every published recommendation.
For practices, the common causes are dull. A page exists in two places with slightly different addresses, so the signal splits. Internal links point to old versions. A service page was built and never linked from anywhere, so nothing points a crawler toward it. Google’s guidance specifically names making content findable through internal links on your own site.
None of that is exotic. All of it is invisible until someone checks.
Reason three: the page does not say who you are
An engine assembling an answer about a local provider needs to establish, from the page itself, what the business is, where it is, and what it does. A homepage that opens with a value statement and a stock photograph supplies none of that in a form a machine can use.
The fix is unglamorous and effective. Say what the practice is, where it is located, and what it treats, in plain prose, near the top of the page. Keep that wording identical everywhere it appears. Inconsistency across pages is worse than plain writing.
Reason four: the answers are not extractable
This is the one that separates pages that get used from pages that get passed over.
An engine building an answer needs a passage it can lift out and use without the surrounding context. A page written as a flowing brochure, where the useful information is spread across four paragraphs and depends on what came before it, does not offer one.
A page that asks a question in a heading and answers it directly in the next forty to sixty words does. The answer needs to work when read cold. No pronouns referring back to the heading, no “as mentioned above,” no assumption that the reader saw the paragraph before.
That is a writing decision rather than a technical one, and it costs nothing to implement.
What Google says you do not need
This part is worth reading carefully, because a good deal of what gets sold as AI optimization runs against it.
Google’s guide to optimizing for generative AI features states that you do not need to create new machine readable files, AI text files, or markup to appear in these features, and that there is no special schema.org structured data required. It lists overfocusing on structured data as a specific pitfall, while noting structured data remains worthwhile as part of general search work.
Structured data still helps engines understand a page, and FAQ markup pairs naturally with content already written as questions and answers. What it does not do is act as a switch that turns AI visibility on. Anyone selling it that way is selling something Google’s own documentation contradicts.
Google also points to the Generative AI performance report in Search Console as the way to see how content performs in these features. That is a measurable check rather than a promise.
Reason five: the site says one thing and the listings say another
An engine building an answer about a local provider is reconciling several sources at once. Your website, your Google Business Profile, directory listings, an insurance panel page, an old profile from a practice you left years ago.
When those disagree about your address, your phone number, or your practice name, the engine has to decide which to believe. Sometimes it picks wrong. Sometimes it hedges and names a competitor instead, because their information is consistent and yours is not.
This is tedious to fix and there is no clever way around it. Someone has to list every place your practice appears online and make the details match. Old listings from a previous location are the usual culprit, followed by variations in how the practice name is written.
The version on your own website should be treated as the master copy. Everything else gets corrected to match it.
What this means for a practice
The sequence is boring and it matters. Confirm crawlers can reach the site, including through whatever sits in front of it. Confirm every service page is indexed and reachable by internal links. State plainly who you are and where. Make the listings agree with the site. Write answers that survive being lifted out of the page.
Only after those are settled does anything more advanced pay off. Schema on an unindexed page does nothing. A beautifully written answer on a page a crawler cannot reach does nothing.
For the patient behaviour driving all of this, see what patients actually ask AI before they book. For whether your own site is even the right place to be publishing, see your own website versus being written about elsewhere.
Frequently Asked Questions About AI Visibility
Why does my practice rank on Google but not appear in AI answers?
Ranking and AI citation are different systems with different inputs. A page can rank on relevance and links while offering no passage an engine can lift out and use. Google also states that indexing and serving are never guaranteed for any page.
Do I need to add schema markup to appear in AI results?
Google’s documentation says no special schema.org structured data is required for its generative AI features, and lists overfocusing on structured data as a pitfall. Structured data remains useful for general search work, so it is worth having rather than worth relying on.
What is llms.txt, and do I need one?
It is a plain-text file describing a site’s structure for AI agents. Adoption is growing, and Google states you do not need to create AI text files to appear in its generative features. Treat it as optional rather than foundational.
How do I know whether AI crawlers can reach my site?
Check robots.txt, then check your firewall, security plugin, and CDN separately. Google specifically calls out hosting infrastructure alongside robots.txt. A site can be open in one place and blocked in another without anyone intending it.
Can I measure whether this is working?
Partly. Google Search Console includes a Generative AI performance report for its own features. Other engines publish less. Many practices also run a monthly check by asking the same buyer questions and recording what comes back.
How long does any of this take to show up?
There is no reliable timeline, and anyone offering one is guessing. Indexing alone can take weeks. Because the work here is mostly fixing access and rewriting pages, the sensible measure is whether the fixes are done rather than when results arrive.
MFG Wellness: Clearing the Blockages Before Writing Anything New
MFG Wellness works exclusively with healthcare providers, from York, Pennsylvania. Most of what this article describes is repair work rather than creative work, and it has to happen before a single new page is worth writing.
How we work on it:
- AEO for Healthcare: Crawler access checked at robots.txt and at whatever firewall or CDN sits in front of it, indexing confirmed page by page, listings reconciled against the site, and only then the rewriting that makes each answer stand alone
- Medical Public Relations: Engines also read what other sites say about a practice, so earned coverage puts material about you in places that do not depend on your own site being crawled correctly
Not sure whether the crawlers can even reach your site? Request a quote and we will check robots.txt, your hosting layer, and what is indexed today.
Works Cited
Google. “AI Features and Your Website.” Google Search Central, developers.google.com/search/docs/appearance/ai-features. Accessed 31 Aug. 2026.
Google. “Google’s Guide to Optimizing for Generative AI Features on Google Search.” Google Search Central, developers.google.com/search/docs/fundamentals/ai-optimization-guide. Accessed 31 Aug. 2026.
Disclaimer
MFG Wellness is a healthcare digital marketing agency in York, Pennsylvania. We are not a law firm and do not provide legal or compliance services.
This article describes published platform documentation as it stands at the time of writing. Search platforms change their guidance, and none of this describes a guaranteed result. Current as of 31 August 2026.
This article is general information and is not legal or compliance advice. Healthcare regulations change and apply differently depending on your arrangement. Confirm current requirements with your compliance officer or counsel.
