Buyer's guide · what we do

What does a GEO agency actually do for a dental practice, day to day?

Not a fixed checklist. A GEO agency tracks how AI engines answer a set of patient questions, and what it works on next is decided by what those answers show, not by a package tier.

Abstract illustration: three vertical bars of different heights, the tallest one solid black.

A GEO agency finds the questions patients actually type into ChatGPT, Perplexity, Gemini and Google AI Overviews about a practice like yours, agrees that question set with you, then checks how each engine answers them every day. What it works on next (page copy, technical fixes, review requests) is decided by what those daily answers show, not by a fixed monthly checklist.

Citevio calls this sequence its Citation-to-Chair Protocol, and it runs the same four steps for every practice: find the real buying questions first, agree the set with the practice before tracking starts, analyze how the four engines answer that set every day, and let the results, not a pre-written plan, decide what gets worked on next. The rest of this page breaks that down into what actually lands in your inbox each month and who is responsible for acting on it.

What deliverables should I get each month?

The shape of the work is closer to an ongoing diagnostic than a one-time SEO project. Every month you should be able to see, in plain language, what each engine said about the practice this month versus last month, which specific gap that points to, and what Citevio proposed to close it. If a report only shows a score with no underlying answers attached, that's not enough to act on: ask to see the actual engine responses. Comparing what different kinds of agency actually deliver is a separate question, covered in dental marketing agency vs. general marketing agency, and the promises worth refusing are listed in red flags when hiring a dental SEO agency.

Some of this is more consequential than it sounds. Recent reporting from Search Engine Land (2025) found that ChatGPT's local answers are still built on Microsoft Bing's index, so a page that ranks well on Google but isn't indexed by Bing can be invisible to ChatGPT regardless of Google performance. Google Search Central (accessed August 2026) explains that pages eligible for AI Overviews must be indexed and eligible to show a snippet in Google Search. Checking for exactly that kind of gap is part of the technical review below.

What's on Citevio's monthly deliverable list?

DeliverableWhat it actually isWho acts on it
Engine answer logDated record of what ChatGPT, Perplexity, Gemini and Google AI Overviews said, for each tracked question, that monthCitevio compiles it; you see the actual answers, not just a score
Priority list for the monthWhich pages, listings or gaps to fix next, ranked by what that month's answers showed was actually missingCitevio recommends; you or your web team implement by default
Draft copy & schemaReady-to-paste page edits, FAQ answers and structured-data blocks addressing the priority listWritten by Citevio, published by you or your team
Crawlability checkWhether the pages being changed are actually readable to AI crawlers, not just to Google's rendererCitevio flags; a developer fixes if needed
Review-pattern notesWhat the practice's review volume and recency currently look like, and what to ask patients. Citevio has not measured whether changing them changes what an engine answersFront-desk team sends the requests
Monthly reportThis month's engine answer log set against last month's, in writingSent to you; no call required to read it

The crawlability check matters because the bot that decides whether ChatGPT can cite a page isn't the same one that trains the model. OpenAI's own crawler documentation distinguishes OAI-SearchBot, which controls whether a page can appear in a ChatGPT search answer, from GPTBot, which is used for model training. A site can allow one and block the other, and checking both is part of this deliverable, not just one.

How often does that check actually turn something up? In Citevio's own reading of US dental practice robots.txt files, 493 of 3,929 sites listed under the Dentist category (12.5%) blocked at least one of the 13 AI and data crawlers tested at the root, and 327 of 3,929 (8.3%) blocked at least one crawler run by OpenAI, Anthropic or Perplexity.

That narrower category cut sits inside a broader open research base. Citevio's June 2026 dataset covers 527 US dental practice websites across seven metros and nine nearby towns. Citevio's dental market research is published openly under CC-BY-4.0 with a DOI, so the numbers can be checked and reused; client data is never published. The permanent DOI is 10.5281/zenodo.23143214.

Citevio AI crawler access study, robots.txt read on 22 July 2026. Dentist-category listings only: n = 3,929 of 6,497 readable US dental practice robots.txt files. Counts and CSV: open data. Method: how we measure AI visibility.
What the robots.txt file doesSites% of 3,929
Blocks at least one of the 13 AI and data crawlers tested at the root49312.5%
Blocks at least one crawler run by OpenAI, Anthropic or Perplexity3278.3%

That is a minority of practices, not most of them, which is the point of checking rather than assuming. A blocked crawler is one reason a practice can be missing from an answer; it is not the only one, and Citevio has not measured whether unblocking a crawler changes what an engine answers.

Why the list changes every month

The deliverables above aren't a package you buy once and repeat. Each month's priorities follow from what that month's tracked answers actually showed. A practice that was named on Perplexity in July but dropped off in August gets different work than one that's been steady. What doesn't change mid-stream is the question set itself: "The question set is fixed and agreed with the client at the start, and logged. It does not change mid-month, because a set that changes cannot be compared month to month." Full methodology on how we measure AI visibility. Citevio has not measured whether any single input, such as reviews, page speed or schema markup, changes what an engine answers; the tracked answers are the measurement, not a proven cause.

On what terms — who owns the work, and how do you get out?

Almost everything delivered in this kind of engagement lives on your property: pages on your website, structured data in your templates, corrections to your Google Business Profile. That changes what the important contract question is. It is not how long the term runs. It is what leaves with you — the page copy, the schema blocks, the agreed question set, and the raw engine answers collected while the work was running.

Those four are worth naming separately, because the first two usually come back without an argument and the second two often do not. Copy and markup are visibly on your site. A question set and a body of dated engine answers can sit in a supplier's tool, and if they do, leaving means starting the measurement from zero somewhere else. Ask for both in writing before you sign, in whatever format the supplier can actually export.

  • The question set — the literal wording of every tracked question, as a file you keep.
  • The answer history — what each engine returned, with dates, not a summary chart of it.
  • The page work — copy and schema, already on your site rather than in a portal.
  • Profile access — your Google Business Profile stays under your ownership, with the supplier removed as a manager on exit.

A long term is not automatically a bad deal, and it should not be treated as one. Six-month and twelve-month agreements are normal in marketing services and can buy a lower rate. What makes a term dangerous is not its length but what happens at the end of it: a twelve-month agreement that returns everything is a better deal than a monthly one that returns nothing. Pricing structures based on a share of revenue — a percentage of treatment income rather than a flat fee — are not common in this particular corner of the market, which is a reason to ask exactly how it would be calculated and audited, not a reason to rule it out.

How much of this is normally published rather than discussed on a call is measurable. Citevio's side-by-side of eight dental AI visibility firms found six of the eight state their contract terms somewhere on their own website — read on 9 August 2026, with the eighth row added on 17 August 2026 — while price and the number of questions tracked are published by two. The column-by-column version is in dental AI visibility agencies compared, and the questions to ask when a cell is blank are in how to verify an agency's AI visibility claims.

Citevio's own terms — billing, notice, setup fee and what happens if you cancel — are not restated here, because a page that mixes "what to expect from any agency" with "what we do" stops being useful for the first purpose. They are published in full on what GEO for a dental clinic costs, where they can be compared line by line against anyone else's.

Citevio publishes its pricing: Visibility is $1,400/mo plus $1,900 setup, Authority is $2,900/mo plus $3,500 setup, and Dominance is $5,500/mo plus $7,500 setup. Citevio does not guarantee placement, but it does guarantee the fee: if at least two of the four engines have not named the practice in one of the commercial-intent questions agreed at onboarding within 45 days of confirmed setup, the setup fee is refunded in full.

Frequently asked questions

Who implements the website changes, the agency or me?

By default, you or your existing web person publishes it. Citevio writes the specific change (new FAQ copy, a schema block, a page edit) as something ready to paste in, and nothing goes live on your site without your sign-off. Some practices prefer Citevio to publish directly instead of routing everything through a web person; that requires a scoped, editor-level login to specific pages rather than full admin access, and it is agreed per engagement rather than assumed.

What access does the agency need from me?

At minimum, read access to what is already public: your website, your Google Business Profile listing, and any existing analytics you are willing to share. That is enough to start the discovery and daily analysis work. Having Citevio implement changes directly adds a scoped, editor-level website login. Social media and ad accounts are never requested, because they are outside Citevio's scope.

How much of my time will this take?

Most of the ongoing work does not need your time at all; the daily engine checks and the analysis behind them happen without you. Where your time actually goes: one conversation at the start to agree the question set, and a short review of the monthly report and priority list. Opting into direct implementation adds a longer one-time setup conversation, and the day-to-day after that still does not need you.

Do you write content, or just audit it?

Both. The daily engine analysis is the audit half: it identifies where a page, a review pattern or missing structured data is costing a practice a mention. The other half is drafting the fix: page copy, FAQ answers, schema markup. Citevio does not hand over a list of problems and leave; specific replacement copy is part of what is delivered.

Do you handle social media or ads too?

No. Citevio is an AI search visibility (GEO) agency, and the work is limited to how ChatGPT, Perplexity, Gemini and Google AI Overviews describe and recommend a practice, not social posting or paid ads. A practice running both needs a separate vendor or an in-house person for social and ads; the two do not overlap in what Citevio delivers.

Does the crawlability check confirm my site works for Google, or does it test AI crawlers separately?

It tests AI crawlers specifically, not just Google's renderer. A page can render fine in Google's crawler and still be unreadable to the crawler an AI engine actually uses, so that gap is checked directly rather than assumed from a normal SEO or PageSpeed test.

Within that check, do you test the bot that trains ChatGPT, or the one that lets it cite your page in a search answer?

Both, because they're not the same bot. OpenAI's own crawler documentation separates OAI-SearchBot, which controls whether a page can appear in a ChatGPT search answer, from GPTBot, which is used for model training — a site can allow one and block the other, so checking both is part of the deliverable rather than just one.

Is a six-month or twelve-month contract normal for this kind of work, or a red flag on its own?

Length on its own isn't the signal. Six- and twelve-month agreements are normal in marketing services and can buy a lower rate; what actually makes a term risky is what happens when it ends. A twelve-month agreement that hands back everything you're owed — the question set, the answer history, the page work, your Google Business Profile access — is a better deal than a monthly one that returns nothing.

Is being charged a percentage of my treatment revenue instead of a flat fee normal for this kind of work?

Not currently — pricing based on a share of revenue rather than a flat fee is uncommon in this corner of the market. That's a reason to ask exactly how it would be calculated and audited before signing, not automatically a reason to rule it out.

Does correcting my Google Business Profile mean giving you manager access, or just letting you view it?

Manager-level access, specifically for the corrections — not just the read access described for the initial review. Our terms specify removing the agency as a manager from the listing when the engagement ends, which only applies if manager access was granted for the GBP work in the first place; the profile itself stays under your ownership throughout.

See what the engines say about your practice first

Before you compare what any agency proposes to do, get an objective baseline: where your clinic stands across ChatGPT, Perplexity, Gemini and Google AI Overviews today. Free, no call.

Run my free AI scan

Delivered within 48 hours · no sales call — ever