Your next patients aren't Googling.
They're asking AI.

The free check is immediate. The technical layers — crawler access, structured data, index health — usually resolve within weeks. Movement in what the engines actually say takes longer, because they have to re-read and re-trust the practice: our setup-fee guarantee is written against a 45-day window for a first naming, and change across the wider question set is a matter of months rather than days. You get before/after reporting from the first cycle, so you can see the direction long before the finish line.

ChatGPT, Perplexity, Gemini and Google AI Overviews — each pulls from different data sources (Bing index, Google Business Profile, review platforms, community content), so each gets its own playbook.

No. Ads are beginning to appear around AI answers, but the recommendation itself isn't for sale. Engines assemble it from public signals: index presence, reviews, mentions, site structure. The answer can't be bought. It can be earned, systematically.

Three things.

• Our own data. We run our own engine scans and keep every result. Our advice comes from measurements we made, not marketing blogs.
• Sources that stay current. Our system re-tests the engines on a loop; when they change, the playbook changes with them.
• Full-fat packages. Review automation, competitor monitoring, content refreshes: things others sell as paid add-ons are built into the plans.

Never. Citevio is async by design: you get short recorded video walkthroughs, written reports and a dashboard. Questions are answered in writing, usually within one business day.

Yes. We only invite real patients to leave honest reviews, we never filter who gets asked, and we never write reviews or offer incentives. This keeps you safe under the FTC's fake-review rule — where penalties can reach $53,088 per violation, the maximum civil penalty published by the FTC in the Federal Register in January 2025 — and Google's review policies.

There are two limits and they are different. Across all plans, Citevio works with a maximum of three cosmetic or Invisalign practices per US state at any one time; when a state reaches three it closes to new clients until a slot opens. That is a capacity limit: every account is run personally by the founder rather than handed to an account manager, and the cap is what keeps it that way. On the Dominance plan it goes further: we contractually commit to working with only one cosmetic/Invisalign practice in your metro area for as long as you're an active client. Your competitors cannot hire us. One city, one clinic.

It is our published method, and every plan runs on it. Six layers, run as a loop: Discover, the question set patients actually ask; Read, whether an engine can crawl and parse the site at all; Match, whether it can connect the practice to the right question; Trust, whether it will say the name out loud; Answer, whether there is a sentence worth quoting; and Chair, whether the citation turns into an appointment. Whichever layer the daily measurement shows breaking most often is the layer we work next. The full method is published — including the parts a competitor could copy.

SEO ranks a page inside a list of links. GEO decides whether the practice is named inside the answer. An engine does not hand back ten options for the patient to weigh; it usually names a handful of practices, and everyone else is simply absent. Layers 1 and 2 of our protocol use the same foundations classic SEO uses — structured data, crawl access, name/address/phone consistency — engineered for AI crawlers specifically. It is not a new discipline invented to sound different. What changes is which bots you are optimizing access for, and what counts as winning. More: dental SEO vs GEO.

Yes, and it is free. The instant check reads your website and scores it in seconds. The free deep scan goes further: we ask ChatGPT, Perplexity, Gemini and Google AI Overviews the questions patients in your city would actually ask about practices like yours, then send you the real queries, screenshots, the competitors named instead of you, and a short recorded walkthrough of where you stand. No call, ever.

Yes, anywhere in the US. The work happens on signals that are read online, so distance is not a constraint. Two limits are: a maximum of three cosmetic or Invisalign practices per state at any one time, and on the Dominance plan, one practice per metro. What we need from you is local — your treatments, your city, and the questions your front desk actually fields.

No. Patient names, phone numbers, appointment times and treatment details never reach us. That is written into our agreement rather than offered as reassurance: review invitations go out from your own appointment system, under your own account, in your practice's name. What we do ask for at onboarding is a short factual record — accepted insurances, hours, price ranges, whether you are taking new patients — because without it there is no such thing as a wrong AI answer, only one we happen to disagree with.

Usually not. Most of what an engine needs is already on the site; it is either unreadable to crawlers or missing the specific answer a patient asked for. We work with the site you have, and if something genuinely has to change we tell you what and why before touching it. Every answer-ready page we write is built for the patient reading it first and structured for the engine second — a page that only works for crawlers gets neither.

Cosmetic and Invisalign practices in the United States, and nothing else. The narrowness is deliberate: the questions patients ask about veneers, implants and clear aligners are specific, and the published city baselines your result is compared against are dental. If a prospect is outside that, we say so rather than take the work.

The month's delivery is whichever layer the measurement says is breaking, rather than a fixed checklist. Alongside it you get a report showing which questions named your practice and which named a competitor, every finding written with its ratio — named in two of three runs on this date, not named at all — what moved since last month, an AI-sourced line drawn from the booking form, and what we are fixing next. Short recorded video walkthroughs instead of meetings.

Because AI answers are not deterministic. The same question, asked twice in the same hour, can name different practices, so one reading of one engine is noise rather than a measurement. A 2025 Profound analysis of roughly 80,000 prompts per engine found that 40 to 60% of the sources an engine cites for a question change within a single month. Every question in a client's tracked set therefore goes to all four engines across at least three separate runs, and a finding without its ratio does not go into a report. If you rerun a question yourself and get something different, that difference is part of the finding, not a contradiction of it.

That is the more urgent half, and we measure it separately. A missing mention is an absence; a wrong answer is a leak. Alongside the visibility set we run a smaller set that asks about the practice by name — insurance accepted, opening hours, price ranges, location, whether you are taking new patients — and sort every answer into correct, wrong, stale, or gap. An engine that says it does not know is honest and low risk. An engine that fills the gap with something plausible is recorded as wrong, because that is what it is.

No one can sell that, and we will not pretend otherwise. Google states that knowledge panels are automatically generated; there is no application form and no date anyone can put on it. What we can build is the conditions: the legal name published in your structured data, the Secretary of State filing and public registry record matched character for character, and third-party sources carrying the same details. Graph presence follows from those or it does not. We price it as the work, never as the outcome.

Pricing is published openly rather than quoted on a call: Visibility at $1,400 per month, Authority at $2,900 per month and Dominance at $5,500 per month. Every plan runs the same Citation-to-Chair Protocol; what changes is how much of the trust and answer layers we work each month, and Dominance adds the contractual one-practice-per-metro commitment. Review automation, competitor monitoring and content refreshes are built into the plans rather than sold as add-ons.

No. No bought links, no private blog networks, no fake or incentivised reviews, and no volume of generated pages. Some of those tactics still move traditional metrics, and they are exactly the signals an engine's trust layer is being trained to discount. We would also have to hide them from you, which is its own answer.

Every account is run personally by the founder, Muhammed Veysel Erin, rather than handed to an account manager — and the three-practices-per-state cap exists to keep it that way. You are never passed to someone who has not read your measurements.