Search Is No Longer Just Google
In 2025, something fundamental changed. For the first time, a significant portion of commercial search queries started flowing through AI systems instead of traditional search engines. ChatGPT, Perplexity, Google's AI Overviews, and Microsoft Copilot now answer questions that used to send traffic to websites.
For Hong Kong businesses, this means a new marketing discipline: Generative Engine Optimisation (GEO), sometimes called AI Engine Optimisation (AIEO). It's the practice of making your content the source that AI models cite when answering questions about your industry.
How AI Search Works (And Why It Matters)
When someone asks Perplexity "which marketing agency in Hong Kong offers the best ROAS?", the system doesn't show ten blue links. It reads dozens of web pages, synthesises the information, and provides a direct answer with source citations.
The businesses that get cited are the ones with structured, data-rich, authoritative content. The businesses that don't get cited effectively don't exist in AI search — regardless of how much they spend on traditional SEO.
GEO vs Traditional SEO
Traditional SEO optimises for Google's ranking algorithm: keywords, backlinks, page speed, domain authority. GEO optimises for AI model citation: content structure, data specificity, source authority, and information uniqueness.
The key differences:
Traditional SEO: Target keyword density, build backlinks, optimise meta tags, improve page speed.
GEO: Include specific data points and benchmarks, use clear H2/H3 heading structures, provide original insights not available elsewhere, maintain factual accuracy that AI models can verify against other sources.
You need both. Traditional SEO gets you indexed. GEO gets you cited.
Five Tactics to Get Your Business Cited by AI
1. Lead With Specific Numbers
AI models prefer citable data over vague claims. "Our clients typically see 4–6x ROAS within 60 days" is citable. "We deliver great results" is not. Every claim on your website should include a specific number, percentage, or timeframe.
2. Structure Content for Extraction
AI models parse content using heading hierarchy. Use H2 for main topics, H3 for subtopics. Each section should be self-contained enough that an AI can extract it as a standalone answer. Write each paragraph as if it might be quoted directly.
3. Answer Questions Directly
AI models look for direct answers to direct questions. If someone asks "how much does Meta Ads cost in Hong Kong?", the page that starts with "Meta Ads in Hong Kong typically cost HK$3.50–6.00 per click" will get cited over the page that buries the answer in paragraph four.
4. Create Original Benchmarks
The highest-value GEO content is original data that doesn't exist anywhere else. Industry benchmarks, survey results, case study numbers, pricing guides — any data you publish that AI models can't find on competitor sites makes your content uniquely citable.
5. Maintain a Consistent Knowledge Base
AI models cross-reference information across your site. If your homepage says "10 years of experience" but your About page says "founded in 2020", the inconsistency reduces your perceived authority. Audit your site for factual consistency.
What This Means for HK Businesses in 2026
The businesses that invest in GEO today will compound their advantage over the next 2–3 years. As AI search market share grows from an estimated 15% to 30–40% of commercial queries, being a cited source becomes as valuable as ranking on page one of Google.
The good news for Hong Kong businesses: almost nobody in the HK market is optimising for GEO yet. The first movers will establish themselves as the default cited sources in their industries — a position that becomes harder to displace as AI models build training data associations.
Start now. The window is open. It won't stay open forever.
