What Is LLMO? How It Differs from SEO, GEO, AIO
— And What Google Actually Said

Hi, I'm Tsubasa.

If you've been reading about SEO recently, you've probably seen the terms LLMO, GEO, AIO, and AEO thrown around. They all seem to mean "optimizing for AI search," but nobody agrees on the definitions. To make things more confusing, on May 15, 2026, Google published an official document stating that "AEO and GEO are unnecessary as separate concepts — from Google's perspective, this is all just SEO."

This article sorts out the terminology, explains what Google actually said, and lays out what small businesses need to know and do.

What Is LLMO?

LLMO stands for Large Language Model Optimization. It refers to the practice of getting your content cited or recommended when AI systems — ChatGPT, Gemini, Claude, Perplexity, Google AI Overviews — generate responses to user queries.

Traditional SEO aims for high rankings in Google's search results pages. LLMO aims for your information to appear inside the AI-generated answer itself. For example, if someone asks Perplexity "What's a good photo retouching service?" and your company name appears in the response, LLMO is working.

AI systems generate answers using a mechanism called RAG (Retrieval-Augmented Generation) — they retrieve information from the web, then synthesize it into a response. The pages they retrieve tend to be the ones ranking well in search results. Google's own AI optimization guide, published May 15, 2026, confirms this: "SEO best practices continue to apply to generative AI features." The guide explicitly states that concepts like AEO and GEO are unnecessary — from Google Search's perspective, this is all SEO.

LLMO vs. GEO vs. AIO vs. AEO — What's the Difference?

Multiple terms exist for AI search optimization, but they all describe essentially the same thing. Different companies use different names for their services, which creates confusion for anyone searching for information.

Whether you search for "LLMO strategy," "GEO optimization," or "AIO best practices," you'll find the same category of services. The terminology doesn't matter — the underlying work is the same.

Google's May 2026 document settled this from their perspective: optimizing for Google's generative AI features (AI Overviews, AI Mode) is the same as optimizing for Google Search overall — it's all SEO, and there's no need for separate concepts like AEO or GEO. At least for Google's ecosystem, the official position is that LLMO/GEO/AIO are not distinct disciplines.

However, when targeting AI search outside Google's ecosystem — ChatGPT, Perplexity, Claude — each platform has its own crawler behavior and retrieval mechanisms. In that context, the term LLMO (focusing on LLMs specifically) is more accurate than treating everything as SEO.

The Relationship Between SEO and LLMO

SEO and LLMO are not competing concepts. LLMO sits on top of SEO as an extension.

What they share in common: creating high-quality content with first-party data and expertise, implementing structured data (JSON-LD) to make page content machine-readable, building E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness), and organizing site structure so crawlers can efficiently access information.

What LLMO adds: enriching FAQ content (AI uses Q&A pairs as building blocks for answers), clarifying entities (defining authors and organizations in structured data), and optionally setting up llms.txt (a site guide for AI crawlers). Google stated in May 2026 that llms.txt is "unnecessary" for Google's AI features, but for ChatGPT, Claude, and Perplexity, it remains useful and costs nothing to maintain.

That said, "similar tactics" and "the same thing" are different claims. Keita Takeuchi, CEO of LANY (a Japanese SEO/LLMO agency), argues that while the tactical overlap is real, the strategic framing is fundamentally different. In SEO, search volume and GA4 conversion data make it relatively clear "what to do to generate revenue." In LLMO, there's no visible prompt search volume and no proven causal link between AI recommendation share and sales. The work has to start from strategy — "what should LLMO look like for this business?" — rather than jumping to tactics.

The funnel position also differs. SEO captures the "entrance" to a website (top-of-funnel to mid-funnel). AI search operates closer to the "exit" — users interact with AI and decide what to choose (mid-funnel to bottom-of-funnel). Takeuchi describes LLMO as "closer to brand marketing than SEO." Applying SEO KPIs directly to LLMO may not work.

The practical takeaway: if your site's SEO fundamentals aren't in place, LLMO-specific tactics won't help. Fix the SEO foundation first, then layer on LLMO-specific elements.

Official Guidelines: Google, ChatGPT, Perplexity

As of May 2026, only Google has published an official content optimization guide for AI search. Other platforms publish crawler control documentation (robots.txt directives) but have not released systematic "how to get cited" guides.

An important distinction: asking Perplexity or ChatGPT "how do I optimize for your platform?" will produce an answer, but that answer is generated text — not the company's official position.

The underlying technical architecture is shared across platforms. Google has officially described Query fan-out — decomposing a user's query into multiple sub-queries and searching them in parallel. Perplexity uses similar query decomposition, and ChatGPT Search rewrites queries and performs multiple internal searches (documented in OpenAI's help pages). When your site has multiple similar pages on the same topic, each sub-query may pull up a different page, making it harder for AI to decide which one to cite. This risk exists across all AI search platforms.

Key LLMO Tactics

Structured Data (JSON-LD)

Structured data helps AI understand your page content. FAQPage (frequently asked questions), Article (publication metadata), and Person (author information) schemas are particularly useful — they provide the building blocks AI needs to construct answers. Google's May 2026 guide states that structured data is not a ranking factor and not required for generative AI features. However, it does enable rich results (FAQ dropdowns, breadcrumbs, author info) in search, which improves click-through rates. Think of it as an indirect benefit rather than a direct ranking signal.

For implementation details, see "How to Write Structured Data (JSON-LD): FAQPage and Article, Step by Step."

llms.txt

A specification proposed in September 2024. You place a text file at your site's root directory (example.com/llms.txt) that tells AI crawlers what your site contains. While robots.txt addresses search engine crawlers, llms.txt targets AI crawlers like GPTBot and Anthropic's crawler for Claude.

Google explicitly stated in May 2026 that llms.txt is "unnecessary" — Google's AI features won't give it special treatment. But for ChatGPT, Claude, and Perplexity, it provides useful context at zero cost. The decision to maintain it depends on whether you're targeting only Google or multiple AI platforms.

For setup instructions, see "What Is llms.txt? Google Says You Don't Need It — Why I Set One Up Anyway."

FAQ Content

AI generates "answers to questions," so FAQ-formatted content naturally aligns with how AI constructs responses. Adding FAQ sections to your pages and marking them up with FAQPage JSON-LD targets both Google's rich results and AI citations.

First-Party Data and Original Information

AI summarizes existing web content but cannot create new experiences or original data. Your own expertise, customer success stories, and proprietary research data are materials AI wants to cite. You don't need a budget for this — writing about your own experience creates differentiation.

Reducing Duplicate Content Within Your Site

In traditional SEO, cannibalization (multiple pages on the same topic competing within your site) wasn't penalized. Google's position was that duplicate content doesn't automatically result in a ranking reduction — it's an efficiency issue where page authority gets distributed.

That changed on May 15, 2026. Google's AI optimization guide listed "Reduce duplicate content" as a technical requirement for generative AI search. Because AI search uses query fan-out (decomposing queries into sub-queries and searching them in parallel), having similar pages on your site means different sub-queries pull up different pages. AI can't determine which one to cite, and the result is that none of your pages get cited.

Three countermeasures: designate one page as the hub for each topic and create clear hierarchy through internal linking. For pages targeting the same search intent, consolidate content or use canonical tags. For glossary or FAQ pages that overlap with main articles, set canonical URLs pointing to the main article.

The key distinction: having multiple articles on the same topic isn't inherently problematic. The problem is having multiple pages that answer the same user intent. A pricing overview (for people checking rates), an outsourcing guide (for people learning how to commission work), and a company comparison (for people choosing between two providers) serve different intents and won't cannibalize each other.

What Small Businesses Should Do First

You don't need to pay a consulting firm $2,000–5,000/month for LLMO services. The following steps are free and form the foundation of LLMO.

• Check your current AI visibility. Enter your company's keywords into ChatGPT and Perplexity. See whether AI recognizes your business — or doesn't.
• Implement structured data. Add FAQPage, Article, and BreadcrumbList JSON-LD to your site. WordPress users can handle this with free plugins like Yoast SEO or Rank Math.
• Add FAQ sections. For each page or product listing, add Q&A pairs addressing questions readers are likely to ask. This serves both rich results and AI citations.
• Set up llms.txt (optional). Google says it's unnecessary, but it's useful for ChatGPT, Claude, and Perplexity at zero cost.
• Verify with Google Search Console. Use the URL Inspection tool to confirm structured data is correctly recognized. The generative AI performance report also shows which of your pages AI search cites.

For a detailed breakdown of LLMO costs and budget-friendly options, see "SEO・LLMO対策の費用が高すぎる？｜AIO・GEOの相場と小規模事業者の選択肢" (Japanese).

Summary: LLMO (Large Language Model Optimization) is the practice of getting your content cited by AI search — ChatGPT, Gemini, Perplexity, Google AI Overviews. Multiple terms exist (LLMO, GEO, AIO, AEO), but they describe essentially the same work. Google's May 2026 official guide states that AEO and GEO are unnecessary concepts — it's all SEO. The practical overlap between SEO and LLMO is large, but the strategic framing and KPIs differ. For small businesses, start with the SEO foundation (structured data, FAQ content, site structure), then layer on LLMO-specific elements like llms.txt. Expensive LLMO consulting is not required — the foundational steps are free.

🇯🇵 この記事の日本語版を読む

よく読まれている記事

PICK

「おすすめ比較記事」をGoogleが規制？｜No.1表記の法的リスクとAI検索

No.1表記は景品表示法違反のリスクがある。Googleの規制・法的リスク・AI検索の3軸で、おすすめ比較記事の信頼性を整理した。

SEO・LLMO

PICK

写真レタッチの料金相場｜外注15社を比較【2026年版】

各社バラバラな料金体系を1ページにまとめました。外注先を比較するときの参考に。

レタッチ・画像加工

PICK

ささげ業務の代行会社おすすめ20選｜選び方と費用相場【2026年版】

撮影・採寸・原稿の頭文字をとった「ささげ」の意味から、代行会社20選の選び方・費用相場まで。EC担当者が外注先を探すときの備忘録。

EC・実務

PICK

有名レタッチ会社9社の法人情報を比較｜資本金・設立年・従業員数・取引先【2026年版】

法人9社の資本金は300万円〜1億円、9社すべてが取引先社名を公開。技術力以外に審査される法人格・設立年・取引先の公開状況を整理。

レタッチ・画像加工

Tsubasa

Working in the EC department of a Japanese apparel company. My responsibilities range from coordinating product photography (sasage) to managing retouching vendors, updating product pages, and supervising part-time staff. I'm not great at Photoshop, so I outsource serious image work to professionals — but after working with many vendors, I've become pretty good at choosing the right one and checking deliverables. This blog is a record of things I've learned on the job.