Patients ask now. Answer engines decide.
By late 2025, AI Overviews appeared on roughly 89 percent of informational health queries on Google, up from 59 percent two years earlier (BrightEdge, 2025), and on a fast-growing share of the commercial searches practices compete for. This is no longer a fringe surface; for health questions it is the default. No one can guarantee an AI citation; Source is the work that makes a practice the obvious source to cite, and the monitoring that shows where it is and is not being surfaced.
This is not Engine with a new label. Engine optimizes for what patients search. Source optimizes for what they ask. The two surfaces share a foundation but diverge in execution, reporting cadence, and the engineering that earns each citation.
- Source defined
- Sourceis a healthcare answer-engine optimization service. The work covers schema-graph engineering across the practice’s site, llms.txt authoring and maintenance, monthly citation monitoring on Google AI Overview, ChatGPT, Perplexity, and Claude, E-E-A-T author and reviewer attribution rollout, and quarterly AEO posture reporting. The deliverable is citation readiness and a measurable posture, not a guaranteed citation.
Build the foundation. Monitor the surface. Hold the lead.
Practices with up to 25 priority URLs; single-location operations; ready to be discoverable to AI surfaces but not in active citation competition with multi-location category leaders.
- Schema graph audit + completion (Organization, MedicalBusiness, Physician, Article, FAQPage)
- llms.txt authoring + monthly maintenance
- Quarterly AEO posture report (which AI engines cite the practice and where)
- E-E-A-T author/reviewer attribution rollout (NPI + board cert in schema)
- Monthly structured-data validation report
- Up to 25 priority URLs covered
Established practices in active AEO competition. Multi-vertical or multi-service offerings where citation share-of-voice meaningfully changes patient acquisition.
- Everything in Foundation, plus:
- Monthly AI Overview / ChatGPT / Perplexity / Claude citation monitoring
- Per-engine share-of-voice tracking inside the practice's category
- Schema-graph extensions (MedicalProcedure, Service, Review)
- Entity boundary mapping: explicit service exclusions so AI engines don't invent offerings
- Quarterly competitive AEO read against three named competitors
- Priority schema engineering: 48-hour fix on graph errors
- Up to 100 priority URLs covered
Multi-location practices, hospital-affiliated programs, and named-brand category leaders where AI citation share is a strategic surface and not just a marketing surface.
- Everything in Growth, plus:
- Per-location schema graphs (multi-location AEO)
- Real-time citation alerts (notification when AI Overview placement changes)
- GBP-to-schema integration; structured data syncs to live operating data
- Custom llms.txt segmentation by patient journey
- Monthly written narrative on AEO posture trajectory
- Up to 200+ priority URLs covered
Different surface. Different cadence. Different work.
The temptation for any agency in 2026 is to bolt “AI Overview optimization” onto an existing SEO retainer and call it covered. We resisted that for the same reason we resisted bolting paid social onto Site maintenance: the operating cadence is too different, and bundled work becomes work nobody owns.
Engine publishes content. Source engineers structured data and monitors citations. Engine reports monthly on keyword position and traffic delta. Source reports monthly on citation share-of-voice across four engines. Engine’s lever is editorial volume. Source’s lever is graph completeness and entity-level identification. Two products, run side by side, never substituting.
Practices that buy Source without Engine get visibility on the answer-engine surface but not on classic search. Practices that buy Engine without Source get classic search but watch their AI Overview share-of-voice slip toward zero. Most established practices end up on both. Some start on one and add the other in year two.
About Source.
- How is Source different from Engine?
- Engine optimizes for what people search. Source optimizes for what they ask. The two surfaces overlap less than you'd think. Google's AI Overview, ChatGPT, Perplexity, and Claude pull citations from a different ranked layer than blue-link organic search. A practice can rank #1 in classic search and never get cited in any AI answer, or vice versa. Engine and Source run side by side; they don't substitute for each other.
- Why is this its own product line?
- Different operating cadence. Engine is a publishing-and-content motion measured in months and quarters. Source is a graph-engineering and citation-monitoring motion measured weekly. Different deliverables, different reporting cycle, different expertise. Bundling it into Engine would dilute both.
- Is AEO real or marketing hype?
- Real, and no longer marginal. By late 2025 AI Overviews appeared on roughly 89% of informational health queries on Google, up from 59% two years earlier (BrightEdge, 2025); SE Ranking had clocked health queries near 65% a year prior (Sept 2024). On the commercial find-a-provider searches a practice actually competes for, AI Overviews run lower, around 13 to 18% and climbing (SE Ranking, 2026), and that is where the demand is. The patient who asks ChatGPT "what's a good concierge medicine practice in Tampa" is making a real demand-side query, and the practices cited in that answer compound visibility while uncited practices don't. Source is the work that decides which side of that line a practice lands on.
- What's the deliverable each month?
- A monthly AEO posture report: which queries surfaced the practice in AI Overviews / ChatGPT / Perplexity / Claude that month, where citations came from, what changed. Plus the work itself: schema graph completions, new llms.txt entries, structured-data fixes, E-E-A-T attribution updates. Quarterly: a deeper review of citation share-of-voice in the practice's category and a recommendation for what to publish next quarter.
- Do we need Engine or Site before Source?
- Not strictly required, but strongly recommended. Source works on the foundation a real schema-equipped site provides. Practices on legacy WordPress sites with no schema graph need a Site build first or they're fixing the foundation while we work on the second-floor finishes. Practices on a Macbach-built site already have most of the schema scaffolding; Source accelerates from there.
- Why does this cost what it costs?
- Schema graph engineering is precise work, performed by engineers with credentials and verified by a structured-data validator on every change. Citation monitoring requires manual review across four engines monthly. E-E-A-T attribution and answer-first restructuring take editorial judgment. The fees reflect work hours on the levers that actually move citation, not retainer-economics. We include llms.txt authoring as housekeeping; we do not charge for it as though it were the thing earning the citations.
- Does llms.txt actually get a practice cited by AI?
- Honestly, not on its own. llms.txt is a proposed convention for listing your key pages in a flat file, and as of 2026 there is no evidence the major answer engines use it to decide citations. We author and maintain it because it is cheap, harmless, and the standard may matter later, but we would be misleading you to sell it as the lever. What actually moves AI citation is the rest of Source: a complete, entity-resolved schema graph, credentialed E-E-A-T attribution, answer-first content structure, and the monitoring that tells us what is working. That is where the work and the fees go.
- How do you stop AI from inventing services we don't offer?
- By defining the boundaries explicitly. Answer engines will fill gaps with assumptions, so the most-overlooked move in medical AEO is telling them what a practice does not do, in plain content and in schema: no Medicaid, no walk-in urgent care, no pediatric cases, whatever applies. We map those exclusions deliberately so the model has nothing to hallucinate, which both protects the practice from disappointed patients and sharpens its entity. It is a small, specific piece of the work that most agencies never think to do.
Start with the audit. Source follows naturally.
The Practice Audit grades your current schema posture and citation visibility on every dimension Source would address. After the report lands, the right tier (Foundation, Growth, or Dominance) usually becomes obvious within ten minutes of reading.