The shift to AI-first search
Something quiet happened over the last 18 months. A growing share of your potential patients stopped typing into Google and started asking ChatGPT: "What's the best aesthetic clinic in Kuala Lumpur for Botox?" Or asking Perplexity: "Which medspa in PJ has the best reviews for laser treatments?"
These aren't fringe users. They're the research-heavy, high-intent patients who spend time comparing before booking — exactly the patients worth acquiring. And AI is now their starting point.
Your clinic can rank #1 on Google and still be completely invisible to ChatGPT, Perplexity, and Google AI Overviews. These are different systems with different rules.
GEO — Generative Engine Optimisation — is the practice of making your clinic visible, credible, and recommendable to these AI systems. This guide is the complete technical breakdown of how it works.
How AI engines decide who to recommend
When a patient asks ChatGPT "best aesthetic clinic in KL", the AI doesn't simply search Google and read back the top result. It goes through a multi-step process:
- Web retrieval: Tools like ChatGPT's browsing and Perplexity actively crawl the web in real time using their own bots (GPTBot, PerplexityBot) to find current, relevant pages.
- Structured data extraction: The AI reads the page HTML and looks for structured signals — schema markup, clear entity definitions, FAQ content — that tell it who this business is and what they offer.
- Content quality signals: Is the content structured in a way the AI can extract citable answers from? Short, factual, clearly attributed content is cited far more frequently than dense paragraphs.
- Training data + fine-tuning: AI models are also influenced by what appears in their training data — business directories, review sites, forums where your clinic is discussed.
The implication: being cited in AI search is not about Google rankings. It requires a different technical and content strategy — and most specialist clinics have done none of it.
The 5 factors that determine your AI visibility
1. AI crawler access (robots.txt)
This is the most common — and most fixable — problem. If your robots.txt
file doesn't explicitly allow AI crawlers, GPTBot and PerplexityBot may not be
crawling your site at all. Every page they can't crawl is a page that can't be cited.
Your robots.txt must explicitly allow these bots:
GPTBot— OpenAI / ChatGPTPerplexityBot— Perplexity AIClaudeBot— Anthropic / ClaudeGoogle-Extended— Google AI Overviews & GeminiOAI-SearchBot— OpenAI's search crawler
Check yours now: visit yourdomain.com/robots.txt. If you see a blanket
Disallow: / or no mention of these bots, this is your first fix.
2. Schema markup (structured data)
Schema markup is JSON-LD code added to your website that tells search engines and AI systems exactly what your business is. For specialist clinics, the highest-ROI schemas are:
- LocalBusiness / MedicalBusiness: Your clinic name, address, phone, hours, services. This is how AI identifies you as a specific, real entity.
- FAQPage: Explicitly structured questions and answers about your treatments, process, and pricing. AI systems are optimised to extract and cite FAQ content.
- Service: Individual markup for each treatment you offer — Botox, fillers, laser, etc. — with description and pricing range.
- Review / AggregateRating: If you have reviews, this markup surfaces them in AI responses as social proof.
FAQPage schema has the highest ROI of any GEO investment for specialist clinics. It directly maps to how patients ask questions — and how AI answers them.
3. Content structure and extractability
AI engines don't read your website like a human does — they parse it for extractable, citable information. Content structured for extractability means:
- One clear H1 per page that states exactly what the page is about
- Logical H2/H3 hierarchy that breaks content into named sections
- Concise, factual answers near the top of each section — the AI "snippet"
- FAQ blocks throughout key pages — these are the most-cited format
- Listicle and comparison sections — "5 things to know about [treatment]"
A 2,000-word page with no structure is nearly unextractable. The same information organised with headers, lists, and a FAQ block becomes highly citable.
4. Entity clarity
AI systems work on entity recognition — they need to clearly identify who you are as a distinct business. Weak entity signals are why many clinics appear as "a medspa in KL" rather than "[Specific Clinic Name], a medspa at [address] known for [specific treatments]."
To strengthen your entity clarity:
- Your clinic name, address, and phone must appear consistently across your website, Google Business Profile, directories (Erufu Care, Findadoctor, WhatClinic), and social profiles.
- Your About page must clearly state who you are, where you are, what you specialise in, and your credentials — in plain, extractable language.
- Use LocalBusiness/MedicalBusiness schema with your full NAP (Name, Address, Phone) on every page.
5. Page crawlability and technical hygiene
Beyond robots.txt, AI crawlers face the same technical barriers that Googlebot does:
- JavaScript-rendered content — if your key content only loads after JS executes, crawlers may miss it. Use server-side or static rendering for critical pages.
- Missing sitemap.xml — without a sitemap, crawlers rely entirely on following links. A sitemap guarantees all pages are discoverable.
- Slow load times — AI crawlers have timeout thresholds. Pages that take over 3–4 seconds to load may be partially or fully skipped.
- No canonical tags — duplicate content without canonicals confuses AI engines about which version to cite.
Measuring your AI visibility: Share of Voice
GEO is measurable. Unlike some marketing channels, you can put a number to your AI search presence — and track it monthly. The primary metric is Share of Voice (SoV).
How it's calculated: Define 50–100 representative buyer-intent prompts ("best aesthetic clinic KL", "where to get Botox in Petaling Jaya", etc.). Run them across ChatGPT, Perplexity, and Google AI Overviews. Count how many mention your clinic. Divide by total prompts. That's your SoV %.
Track this monthly alongside:
- Citation frequency — total mentions, not just %. Tracks absolute volume.
- Sentiment — are you cited as "recommended" or mentioned in passing?
- Purchase-intent context — cited in "best [treatment] clinic" queries (high intent) vs generic brand mentions (lower intent).
I use Otterly.ai for tracking — it monitors mentions across ChatGPT, Perplexity, Google AI Overviews, Gemini, and Copilot on a scheduled basis and exports SoV reports.
Quick wins you can implement today
If you want to start improving your AI visibility before engaging a GEO consultant, these are the highest-leverage actions:
- Fix your robots.txt — add explicit
Allow: /for GPTBot, PerplexityBot, ClaudeBot, and Google-Extended. Takes 5 minutes. Potentially the single highest-impact action. - Add FAQPage schema to your homepage — 6–8 questions covering your most common treatment queries. Use Google's Rich Results Test to validate.
- Add LocalBusiness schema — include your full NAP, opening hours, and service list in JSON-LD format.
- Submit a sitemap.xml to Google Search Console and Bing Webmaster Tools. AI crawlers follow the same sitemap signals.
- Audit your entity consistency — check that your clinic name, address, and phone are identical across your website, Google Business Profile, and key directories.
These five actions can meaningfully improve your AI visibility within 6–10 weeks. They're also the foundation for everything more advanced.