Illustrative scenario. The clinic below is a composite sketch, not a real client, and no results are stated or implied. Use this as a planning template; the only real numbers are the ones your own baseline and re-measurement produce.
What is the situation?
Imagine a two-dentist practice in a competitive suburb: excellent patient retention, a dated website, a Google Business Profile last touched two years ago, reviews trickling in unanswered, and pricing that exists only as “call us”. Patients increasingly ask assistants “best dentist near me for anxious patients”, “does X take my insurance”, “how much is a crown at X” — and voice assistants compress the shortlist to one or two names. The hypothesis to test: the clinic loses patients at the answer layer before its clinical quality ever gets considered. See voice assistants and local discovery.
How would the 90-day POC be structured?
Week 1 — baseline
Freeze a 40-60 question set in the clusters patients actually use: service-and-fit (anxious patients, kids, implants), logistics (insurance, parking, hours, emergency slots), price anchors, and trust verification (credentials, hygiene standards). Phrasing comes from front-desk call logs. Run the set across Google AI surfaces, ChatGPT and Perplexity, plus a voice-assistant spot check, since local queries skew conversational. Score the four dimensions per engine, screenshot every answer, and audit the NAP baseline: name, address, phone and hours across every directory the engines cite.
Weeks 2-10 — intervention
- Profile completeness (weeks 2-3): rebuild the Google Business Profile — services, photos, attributes, Q&A — and sync hours and NAP across the directories the baseline showed engines citing. Local GEO fundamentals are in GEO for dental practices.
- Answer-first service pages (weeks 3-7): one page per service with the questions patients ask answered in text: what it costs or what drives the cost, what insurance handling looks like, what the first visit involves, honest “who this suits”. FAQPage schema mirroring visible content.
- Review operations (weeks 2-10): a compliant invitation flow — every patient invited the same way, no gating, no incentives — and a response ritual where every review, including the bad ones, gets a professional reply. See review management.
- Consistency fixes (weeks 4-8): correct the stale directory entries, the old phone number on a forgotten profile, the insurance list that changed last year. Grounding reads contradictions as unreliability.
Week 12 — re-measurement
Re-run the frozen set, same engines, same rubric, and produce the delta report with the screenshot archive. The method follows the anatomy of a 90-day POC.
What would be measured?
| Dimension | What it answers for this clinic |
|---|---|
| Visibility (0-100) | Does the clinic appear when patients ask assistants for recommendations in its suburb? |
| Citation Rate (0-100) | Do answers cite the clinic’s site and profiles, or only aggregators and competitors? |
| Semantic Coverage (0-100) | Do answers carry the right services, insurance and hours — or stale data from old profiles? |
| Brand Perception (0-100) | Is the clinic framed as gentle and trustworthy, and are any stated facts wrong? |
Per engine, monthly, frozen set, screenshot evidence — the open standard in the measurement standard.
What decisions would this POC inform?
Whether the practice manager can own GEO as a monthly ritual (usually yes, at local scale) or needs agency support; whether price transparency on the site changes citation behavior for cost questions; and whether the voice-assistant layer needs separate attention. For a local clinic the POC also has a quieter benefit: the baseline itself — seeing what assistants currently say about you — is worth the week even if nothing else ships.
What pitfalls would this scenario need to avoid?
- Review bursts before a measurement date; velocity spikes read as manipulation and invalidate the evidence.
- Fixing Google and ignoring the directories engines actually cite for the category.
- Hiding all pricing; “call us” cannot be quoted, so the engine quotes the competitor’s range instead.
- Measuring only Google; local patients also ask ChatGPT and voice assistants.
Local GEO is the smallest-scope version of the discipline and therefore the best place to learn it: one location, one question set, one quarterly ritual. The POC tells the clinic exactly where it stands — and for most local practices, standing somewhere measurable at all is already ahead of the street.