If Someone Sells You GEO on Top of SEO, You're Paying Twice
The rep shows up every few years with a panel that promises to replace something you already order, and half the time it turns out to be the same assay with a new logo on the requisition form.
New test, or a new name for an old one. So you ask what it catches that the old one didn’t.
Sometimes there’s a real answer. Often there isn’t.
A version of that conversation is running through marketing right now, and getting it wrong means paying for a second panel while the pages patients actually read stay exactly as thin as they were.
The pitch is “GEO,” generative engine optimization, the practice of getting a site cited by ChatGPT, Perplexity, and Google’s AI Overviews on top of ranking in the old list of blue links.
Vendors are selling it as a new discipline, billed separately from SEO, sometimes as its own line item on the invoice.
So before you agree to pay twice for what might be one job, work out what actually changed.
The mechanism, in plain terms
AI answer engines aren’t out crawling the web and forming their own opinions. They draw on the same indexed web Google has spent two decades organizing, and much the same evidence guides them: which pages other reputable sites link to, which pages demonstrate real expertise, which pages actually answer the question being asked.
Independent studies have found AI citations and top Google rankings overlapping on the order of 85 to 90 percent.
That overlap has a cause. One underlying body of web content is feeding two different kinds of output.
Practically, that means a dermatology practice whose condition pages rank well for “eczema treatment Erie PA” is already most of the way to being the page an AI system cites when a patient asks an assistant the same thing.
The credentialed-author bio, the structured FAQ, the specific condition page instead of a vague “Services” page. All of that does double duty, one asset read by both systems.
So when a vendor proposes a “GEO package” that turns out, on inspection, to be a rebrand of the same content and technical work you’d want for ordinary search visibility, you’re being asked to pay for a diagnosis you already have.
The 10 percent that’s actually new
But there’s a real difference, and it’s worth naming precisely so you can tell it apart from the repackaging.
Traditional SEO optimizes for a single query producing a ranked list. AI systems handle a question differently.
Behind the scenes the system breaks that question into several smaller ones implied by what the patient actually asked, checks a source for each, and folds what it finds back into one reply.
A patient who asks an assistant “what’s the difference between a family physician and an internist for ongoing care” has kicked off more than one search.
The system is also quietly working out what an internist treats, how the two credentials differ, when a patient should choose one over the other, and what that decision looks like locally.
Each of those smaller questions is a separate opening to be the source it cites, even if your page never ranks first for the question as the patient typed it.
The genuinely new work is content built to answer those fanned-out sub-questions, beyond the headline query. For a medical practice, that looks like:
- FAQ sections that answer the adjacent questions a patient would ask a friend, beyond the one they typed into Google – “what to expect at a first visit,” “how is this different from urgent care,” “does insurance typically cover this”
- Condition and service pages that go one layer deeper than a single definition (causes, when to see a doctor, what treatment actually involves), written in the practice’s own words rather than a syndicated summary
- Content structured so each sub-question has a clear, extractable answer sitting right next to it. An answer three screens down inside a long paragraph is a lot harder to pull out
None of this replaces topical depth, named authorship, or structured data. It sits on top of them, an extension of the one program you’re already running.
On a page that’s already built well, it’s modest work.
What to ask before you sign a separate GEO contract
Ask any vendor proposing a distinct GEO line item what that specific work adds that the site’s existing SEO and content work doesn’t already cover.
An honest vendor can point to something concrete: a sub-question content gap they mapped, a new FAQ structure, a citation-tracking report showing where the practice is and isn’t turning up in AI answers.
A dishonest one will describe activities indistinguishable from ordinary content and technical SEO, with “AI visibility” language layered over the same deliverables.
Strip the acronyms out of the proposal and read what’s left.
If it describes good content and structured data work you’d want whether or not AI search existed, it’s the same work with a new name.
If it describes specific attention to how a question fans out into sub-questions your site doesn’t currently answer, it’s a real, incremental addition, and usually one small enough that it doesn’t justify a second monthly retainer.
So what do you do with your current contract?
Pull it up, along with any open proposal, and look for “GEO,” “AI visibility,” or “generative optimization” sitting there as a separate line item.
If it’s there, ask your vendor in writing to name the specific deliverables under that line that aren’t already covered by the SEO or content work you’re paying for elsewhere in the same agreement.
If what comes back is vague, or just a restatement of work they’re already billing you for, you’ve got your answer.
The practices getting real value out of this moment are the ones whose sites already answer patient questions clearly and specifically, which was true well before anyone coined the term GEO and remains the whole job now.
Questions practices ask about this
How do I find out whether my practice is showing up in AI answers right now?
Ask the assistants yourself, the same way a patient would. Open ChatGPT, Perplexity, and Google in a private or logged out window, type ten or fifteen questions a patient in your area would ask, and write down which sites get named and linked. Repeat it monthly. That log tells you more than a vendor dashboard, and it takes about an hour.
What should I ask a vendor before signing a separate GEO contract?
Ask them to name the specific deliverables under that line that aren't already covered by the SEO or content work elsewhere in the same agreement. A straight answer points to something concrete, like a mapped list of sub-questions your pages don't answer, a new FAQ structure, or a report showing where you get cited and where you don't. Vague activity language over familiar deliverables tells you plenty.
Can my staff write this content, or do I need to hire someone?
Your staff can write plenty of it, and the front desk is usually the best source of the questions worth answering. Have them keep a two week tally of what patients ask on the phone, then answer each one in two to four sentences under a heading that repeats the question. Hiring out makes sense when no one on your team has recurring hours for it, because this is ongoing work.
How long after adding this content before it shows up in AI answers?
Weeks to a few months is the realistic window to plan around, and it will be uneven. A page has to be crawled and indexed before any assistant can retrieve it, so confirm indexing in Google Search Console first. After that, some questions start pulling your page in fairly quickly while others take much longer, and treat any promise of a specific citation as a red flag.
Photograph: cottonbro studio / Pexels