What is LLMO, and how is it different from SEO?

LLMO (LLM Optimization), also called GEO (Generative Engine Optimization), is the practice of structuring a website so AI language models can read, summarize, and cite its content accurately. Traditional SEO targets ranking highly in search engine results pages; LLMO targets being correctly surfaced and quoted when an AI synthesizes an answer directly. The two aren't in conflict, but the techniques differ in places.

Should political websites do LLMO?

Strongly worth considering. More people are asking political questions directly to AI assistants instead of searching and reading themselves. If a website's data structure is hard for AI to parse, information about a candidate or campaign may simply never surface — or worse, the AI may summarize inaccurate information from another source instead.

Concept · 28 July 2026

Is AI Reading Thai
Political Websites? This Is LLMO

As more people ask political questions directly to AI instead of searching the web themselves, the websites AI can "read" well are the ones that keep getting cited and surfaced. That's the idea behind LLMO.

Is AI reading Thai political websites? This is AI Optimization, or LLMO

🤔 What is LLMO?

LLMO stands for LLM Optimization — some call it GEO (Generative Engine Optimization). It's the practice of structuring a website so AI language models (ChatGPT, Claude, Gemini, Perplexity, and others) can read, summarize, and cite its content as accurately as possible.

It differs from traditional SEO (Search Engine Optimization), which aims to rank a page highly on a search results page. LLMO instead aims to have an AI directly synthesize the content into an answer. When someone asks an AI "who is Candidate No. 2 for the SSO board?", how good that answer is depends heavily on whether the source website's data is structured in a way AI can easily parse.

📊 Generic structure vs. AI-ready structure

😵 Generic structure (hard for AI)

Text mixed in with layout, no clear labels for what's a name, a title, or a policy — AI has to guess meaning from surrounding context, risking misinterpretation, especially with Thai proper nouns AI often confuses.

✅ AI-ready structure

Schema.org marks up exactly what's a Person, a Policy, an FAQ

{
  "@type": "Person",
  "name": "Boon Arayapon",
  "jobTitle": "Candidate No. 2"
}

🛠️ What this site has actually done for LLMO

📄

llms.txt — a new standard for AI crawlers

A single file summarizing the whole site's content in a format AI can ingest in one pass, instead of crawling page by page — helping AI grasp the site's overall picture faster and more accurately.

✓ Live at boonarayapon.com/llms.txt

🏷️

Schema.org structured data

Every key page carries JSON-LD clearly marking what's a Person, Article, FAQPage, or Event — letting AI distinguish candidate information, policy positions, and FAQs precisely, instead of guessing from plain text.

✓ Applied across all major pages

🤖

A dedicated AI Reference page

A page built specifically for AI to read — not for humans — containing key facts, FAQs, and disambiguation (e.g., correcting the misconception that "Mor Boon" implies he's a medical doctor).

✓ boonarayapon.com/ai-reference/

🕷️

An AI-crawler-permissive robots.txt

GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and other AI crawlers are all allowed full access to content — rather than blocking them, as many sites do out of concern about being used as AI training data.

✓ boonarayapon.com/robots.txt

Younger people don't search and read anymore
They ask AI and listen to the answer

💡 Why this matters for Thai politics

The SSO Board election is one most Thais have never heard of. More and more people will start looking for information via AI directly rather than searching themselves. If a candidate's website lacks AI-readable structure, there are two risks: AI may simply never find the information at all, or worse, it may summarize inaccurate information from another source instead. LLMO isn't just a technical concern — it's about maintaining control over the accuracy of one's own information in an era when people increasingly rely on AI.

📖 Want to see the detailed real-world example?

An honest, first-hand account — including mistakes made along the way

🤖 Read the full case study → 🔍 See the real AI Reference page →

Note: This article explains the general concept of LLMO/GEO alongside a real-world implementation on this website. It is not deep technical guidance for developers.

Is AI Reading ThaiPolitical Websites? This Is LLMO