Google has changed. When someone searches for a medical professional in Australia, they increasingly see an AI-generated summary sitting above the organic search results. Not a list of links. A paragraph. One narrative, assembled by Google's AI from whatever sources it deems most relevant.
Google calls these AI Overviews. They appear on 88% of healthcare-related searches. And for medical professionals whose online presence includes negative press, a regulatory finding, or an outdated complaint, that AI summary can compress a complex situation into a single damaging sentence.
This is not a future problem. It is happening now.
What Triggers an AI Overview
Not every Google search produces an AI Overview. The trigger depends almost entirely on how Google classifies the intent behind the query.
Navigational queries ("Dr Jane Smith anaesthetist") rarely trigger them. Google interprets this as someone looking for a specific person's website. It shows the ten blue links as usual. Trigger rate: roughly 1-10%.
Informational queries are different. Add a geographic qualifier ("Dr Jane Smith anaesthetist Australia") and Google reclassifies the search as someone wanting to learn about this person, not just find their website. The trigger rate for healthcare informational queries jumps to 96%.
That single word, "Australia", changes everything.
Question-phrased searches ("who is Dr Jane Smith") are 84% more likely to trigger an AI Overview. Longer queries (seven or more words) trigger them at 46%, compared to 9.5% for single-word searches.
The practical implication is straightforward. Your name plus your profession plus a country or context qualifier will almost certainly produce an AI summary. And that summary will be the first thing anyone reads.
Where Does the AI Get Its Information?
This is where it gets concerning for anyone who assumed ranking on page one was enough.
In mid-2025, 76% of AI Overview citations came from pages that also ranked in the top 10 organic results. By early 2026, that figure had dropped to 38%. Google now runs what researchers call "query fan-out", issuing multiple related sub-searches and scoring sources on relevance, authority, freshness, and engagement before assembling its answer.
The AI pulls from beyond the first page. It pulls from Reddit threads, LinkedIn posts, Medium articles, news archives, professional directories, and structured data sources like Wikidata. A single negative article buried on page two of organic results can resurface in the AI summary if Google's system considers it relevant.
The sources that carry the most weight include Wikipedia and Wikidata (disproportionately weighted across every major AI system), high domain authority platforms like LinkedIn, Medium, and YouTube, and pages that include structured data, statistics, and cited sources. Fresh content (updated within two months) earns 28% more AI citations than older pages.
And 44% of all AI citations come from the first 30% of a page's text. What you put in your opening paragraph matters more than anything buried further down.
What Is GEO?
Generative Engine Optimisation (GEO) is the discipline of optimising your content and online presence so that AI search systems accurately represent you.
Traditional SEO asks: how do I rank higher in Google's list of links? GEO asks a different question: when Google writes a paragraph about me, what does it say?
The techniques are measurable. Academic research (Aggarwal et al., KDD 2024) tested nine GEO strategies across thousands of queries and found:
- Adding specific statistics to content: +32% AI visibility improvement
- Adding quotations from authoritative sources: +22-41% improvement
- Citing sources within content: +9-30% improvement
- Using structured data (JSON-LD schema markup): +73% selection boost
- Keyword stuffing: -8% to -10% (it actively hurts)
That last point matters. The tactics that worked for gaming old-school SEO do not work here. AI systems penalise content that reads like it was written for an algorithm rather than a person.
Why This Matters for Medical Professionals
For most professionals, their Google results are a background concern. Patients search, find the practice website, and book an appointment.
But when negative content enters the picture (a media article, a regulatory finding, an AHPRA notation, even an unfair online review), the AI Overview compresses it. A tribunal finding that resulted in a reprimand with no conditions on practice becomes a bare statement: "Dr X was found guilty of unsatisfactory professional conduct." No context. No nuance. No mention of the outcome.
And because AI summaries get shared, screenshotted, and fed back into future AI training data, they create a reinforcement loop. The AI's summary becomes a source for the next AI's summary. Breaking that cycle requires more than just publishing positive content and hoping it ranks.
What Can Be Done
GEO is not a single fix. It is a systematic approach to shaping how AI systems understand and represent a person or practice.
Structured data tells AI systems exactly who you are. Person schema markup on your website (your name, credentials, location, education, professional memberships) feeds directly into Google's entity recognition. Without it, the AI is guessing.
Entity presence matters enormously. A Wikidata entry (the structured data backbone behind Google's Knowledge Graph) makes you 3.2 times more likely to get a Google Knowledge Panel and 2.7 times more likely to be cited in AI Overviews. Wikidata does not require Wikipedia-level notability. Any identifiable professional with verifiable credentials can have an entry.
Content structure determines what gets extracted. Front-load your key information. Make paragraphs self-contained (134-167 words is the optimal range for AI extraction). Use clear headings. Include data and cited sources. Comparison tables are cited at 2.5 times the rate of equivalent prose.
Platform presence creates the source pool the AI draws from. Publishing on high domain authority platforms (LinkedIn, Medium, Substack) gives the AI more positive sources to work with. Research shows AI systems exhibit a "systematic and overwhelming bias" toward third-party (earned) media over content published on your own website.
Freshness keeps you relevant. Pages updated within two months earn significantly more AI citations than stale content. A website you built in 2021 and never touched is losing ground every month.
The Shift
Traditional reputation management was about controlling a list. Push the bad results down, push the good results up. That still matters. But the AI does not present a list. It presents a narrative.
Controlling that narrative requires a different approach. It requires structured data that tells the AI who you are. It requires content designed to be extracted, not just read. It requires presence across the platforms the AI trusts most. And it requires ongoing attention, because the AI's summary can change with every crawl.
At Narrative Digital, GEO is built into everything we do for clients. Every article is structured for AI extraction. Every client site includes enhanced schema markup. Every strategy includes Wikidata, cross-platform publishing, and monthly AI monitoring alongside traditional SERP tracking.
The professionals who act on this now will shape how AI represents them for years to come. The ones who wait will be shaped by whatever the AI finds on its own.
Narrative Digital provides reputation management and practice marketing for medical professionals across Australia. If you would like to understand how AI search is representing you or your practice, contact us for a confidential assessment.