# llmion > llmion 是 AI 品牌能見度(GEO / AEO)監測與優化平台:每週以即時搜尋模式詢問 ChatGPT、Google AI Overviews、Perplexity、Gemini、Claude,把品牌被 AI 推薦的程度量化為 0–100 分的 BVS(Brand Visibility Score),並依缺口生成 AI 會引用的內容、發布後自動追蹤效果。 > > llmion is an AI brand-visibility (GEO / AEO) monitoring and optimization platform. Every week it asks ChatGPT, Google AI Overviews, Perplexity, Gemini and Claude in live-search mode, quantifies how often and how well each brand is recommended as a 0–100 BVS (Brand Visibility Score), generates AI-citable content for the gaps it finds, and tracks the effect after publishing. 網站 / Site: https://llmion.com 語言 / Languages: 繁體中文(/)、English(/en)、日本語(/ja) 最後更新 / Last updated: 2026-11-15 作者 / Author: llmion ## 定義 / Definitions - **AI 能見度 / AI Visibility**:品牌在 ChatGPT、Google AI Overviews、Perplexity、Gemini、Claude 等生成式 AI 的回答中被主動提及、推薦或引用的程度。與 Google 排名是不同系統:AI 的推薦來自它即時檢索到的頁面與訓練資料。 The degree to which a brand is mentioned, recommended or cited in answers from generative AI such as ChatGPT, Google AI Overviews, Perplexity, Gemini and Claude. It is a different system from Google ranking: AI recommendations come from the pages it retrieves live and from its training data. - **BVS(Brand Visibility Score)**:llmion 定義的 0–100 分 AI 能見度指標。以 D1 品牌出現、D2 語義相符、D3 引用位置、D4 引用深度(含情感)、D5 來源連結五個維度計分,再依五個平台的市場影響力加權;≥80 強引用、≥60 可見、≥40 弱引用、≥20 低可見、<20 未被引用。 llmion's 0–100 AI-visibility metric. Scored on five dimensions (D1 brand mention, D2 semantic match, D3 citation position, D4 citation depth incl. sentiment, D5 source link) and weighted by the market influence of five platforms; ≥80 Strong, ≥60 Visible, ≥40 Weak, ≥20 Low, <20 None. - **AIF(AI-Friendly Format)**:內容在發布前的 AI 友善格式檢查,v2 共九個維度:頂部直接回答、數字密度、Entity 密度、FAQ 品質、佔位符完成率、字數、來源引用、結構化(表格 / 步驟 / 定義句)、新鮮度與作者。總分 100,≥60 可發布,Autopilot 自動發布需 ≥80 且零佔位符。 A pre-publish AI-friendly-format check. v2 has nine dimensions: direct answer at the top, number density, entity density, FAQ quality, placeholder completion, word count, source citations, structure (tables / steps / definition sentences), freshness and author. Out of 100; ≥60 may be published, Autopilot auto-publishes at ≥80 with zero placeholders. - **GEO / AEO**:Generative Engine Optimization(生成式引擎優化)與 Answer Engine Optimization(答案引擎優化):讓品牌內容成為 AI 搜尋與 AI 助理回答時的引用來源。llmion 的監測(BVS)與內容引擎(AIF)即為 GEO / AEO 的量測與執行層。 Generative Engine Optimization and Answer Engine Optimization: making a brand's content the source that AI search and AI assistants cite when they answer. llmion's monitoring (BVS) and content engine (AIF) are the measurement and execution layers of GEO / AEO. ## BVS 計算方式 / How BVS Is Calculated - 每個問題的 BVS = D1×25% + D2×15% + D3×15% + D4×20% + D5×25%(AI 無法即時搜尋、降級為記憶模式的回答改用 D1×30% + D2×25% + D3×20% + D4×15% + D5×10%) - 平台 BVS = 該平台所有問題 BVS 的平均值 - 最終 BVS = ChatGPT×30% + Google AI Overviews×25% + Perplexity×20% + Gemini×15% + Claude×10%(未監測 Google AI Overviews 時,其權重按比例分配給其餘平台) - Question BVS = D1×25% + D2×15% + D3×15% + D4×20% + D5×25% (answers where the AI could not search live and fell back to memory use D1×30% + D2×25% + D3×20% + D4×15% + D5×10%) - Final BVS = ChatGPT×30% + Google AI Overviews×25% + Perplexity×20% + Gemini×15% + Claude×10% (when Google AI Overviews is not monitored, its weight is redistributed proportionally to the other platforms) ### 五個維度 / Five dimensions - D1 品牌出現(25%,D1 Brand Mention):AI 的回答中有沒有提到你的品牌名稱或別名。 每題每平台採樣多次時,D1 = 提及率 × 100(例如 3 次中提到 2 次 = 67 分)。單次採樣時提到 = 100 分,沒提到 = 0 分。大小寫與全半形不敏感,支援別名比對。 - D2 語義相符(15%,D2 Semantic Match):AI 提到你的品牌時,前後文和你的品牌核心概念(Entity)有多相似。只看品牌首次出現前後各 300 字,不看整篇回答。 使用語義向量比對(text-embedding-3-small)。相似度 ≥ 0.85 = 100 分,0.70-0.85 = 60-100 分(線性),0.50-0.70 = 20-60 分,< 0.50 = 0 分。品牌完全沒出現 = 0 分。 - D3 引用位置(15%,D3 Citation Position):品牌名稱首次出現在 AI 回答的哪個位置。 前 1/3 = 100 分,中 1/3 = 60 分,後 1/3 = 25 分,未出現 = 0 分。 - D4 引用深度(20%,D4 Citation Depth):AI 提到你的品牌時,是主動推薦、並列選項、還是順帶一提,以及語氣是正面、中性還是負面。 主動推薦 = 100 分,並列選項 = 65 分,間接提及 = 35 分,未提及 = 0 分。由 AI 分析回答語境判定。若 AI 以負面語氣描述你的品牌(提到缺點、警告或不建議),該分數減半(例如並列選項負面 = 33 分);儀表板的問題表格會以紅色「!」標示負面描述。 - D5 來源連結(25%,D5 Source Link):AI 即時搜尋後實際引用的來源連結中,是否包含你的網站(含子網域)。 有引用 = 100 分,沒有 = 0 分。所有平台都計分(ChatGPT、Gemini、Claude 開啟網頁搜尋,Perplexity 與 Google AI Overviews 原生附來源)。AI 未即時搜尋(記憶模式)的回答沒有來源,該題改用 v1 權重。 ### 平台權重 / Platform weights - ChatGPT 30%:全球使用量最大的 AI 對話平台,對品牌推薦的影響力最高。監測走即時網頁搜尋模式。 - Google AI Overviews 25%:Google 搜尋結果頂部的生成式摘要,流量影響最大。未監測此平台時,權重按比例分配給其餘四個平台。 - Perplexity 20%:原生的 AI 搜尋引擎,每個回答都附來源連結,對流量貢獻最直接。 - Gemini 15%:Google 的 AI 助理,監測開啟 Google 搜尋 grounding,反映 Google 索引中的品牌狀態。 - Claude 10%:偏好有明確立場的專業內容,在 B2B 和專業領域影響力大。監測開啟網頁搜尋工具。 ### 等級 / Grades - Strong(強引用),80-100 分:AI 平台主動、頻繁地將你的品牌作為首選推薦。你的品牌已經在 AI 的推薦名單中佔據穩固位置。 - Visible(可見),60-79 分:品牌出現在多數相關問題的回答中,但不一定是首選推薦。持續優化可以提升到強引用。 - Weak(弱引用),40-59 分:偶爾出現在 AI 回答中,位置偏後或以間接方式提及。需要加強結構化內容和 Entity 密度。 - Low(低可見),20-39 分:極少出現在 AI 回答中,且多為間接關聯。需要從基礎開始建立 AI 友善的內容策略。 - None(未被引用),0-19 分:AI 平台回答相關問題時幾乎不提及你的品牌。這是最需要立即行動的狀態——但也代表巨大的成長空間。 - 聲量占比(Share of Voice)統計每一筆 AI 回答中,你的品牌與每個競品是否出現、誰最先出現。占比 = 該名稱出現次數 ÷ 所有追蹤名稱的出現次數;另提供提及率(出現的回答比例)與最先出現率。 完整說明:[BVS 分數完整指南](https://llmion.com/bvs) ・ [BVS Score Complete Guide](https://llmion.com/en/bvs) ## 核心頁面 / Key pages - [首頁 / Home](https://llmion.com/): 你的客戶正在問 AI「推薦哪家」你的品牌有 AI 能見度嗎? — 當你的潛在客戶問 ChatGPT「推薦一家東京的財富管理顧問」,AI 只會回答少數幾個品牌。這份名單來自兩個地方:AI 當下檢索到的網頁,以及模型的訓練資料。ChatGPT 搜尋、Perplexity、Google AI Overviews 都會附上引用來源,被引用的頁面就是 AI 推薦的依據。llmion 幫你進入這份名單:每週以即時搜尋模式監測你在 5 個 AI 平台的 BVS 分數,找出 AI 回答裡沒有你的問題,生成 AI 會引用的內容,並在發布後追蹤效果。 - [Home (English)](https://llmion.com/en): When your potential client asks ChatGPT to "recommend a wealth management advisor in Tokyo," the AI names only a handful of brands. That list comes from two places: the web pages the AI retrieves at that moment, and the model's training data. ChatGPT search, Perplexity and Google AI Overviews all cite their sources, and the cited pages are what the recommendation rests on. llmion helps you get on that list: it measures your BVS across 5 AI platforms every week in live-search mode, finds the questions where AI answers leave you out, generates content AI will cite, and tracks the effect after you publish. - [BVS 分數完整指南](https://llmion.com/bvs): Brand Visibility Score(BVS)是 0-100 分的 AI 能見度量化指標。了解 5 個評分維度、5 個平台加權、等級定義和提升策略。 - [BVS Score Complete Guide](https://llmion.com/en/bvs): Brand Visibility Score (BVS) is a 0-100 AI visibility metric. Learn about the 5 scoring dimensions, 5 platform weights, grade definitions, and improvement strategies. - [定價 / Pricing](https://llmion.com/#pricing): Free $0、Starter $29、Pro $79、Agency $299(每月,USD) - [常見問題 / FAQ](https://llmion.com/#faq): 10 題,含直接答案與補充說明 ## Blog - [GEO 選題:怎麼找出 AI 還答不好的內容缺口](https://llmion.com/blog/geo-topic-selection): GEO 選題的原則,是找出「客戶常問、AI 的答案還很弱、而你剛好最懂」的問題。先數 AI 引用了幾個不同網站來判斷稀缺度,再實際問 AI 一次,找出沒答到、答得模糊或答錯的內容缺口,最後把選定的問題加入量測清單。(2026-11-15) - [AI 引用來源怎麼決定?檢索式回答的五個階段](https://llmion.com/blog/how-ai-chooses-sources): AI 引用的網頁不是照搜尋排名決定的。ChatGPT、Gemini、Perplexity、Google AI 摘要會先把問題改寫成查詢、搜尋取回網頁、挑出最能回答問題的段落,再寫成附來源的答案。看懂這五個階段,就知道你的網頁在哪一步被漏掉。(2026-11-15) - [GEO 怎麼做?讓 AI 推薦你品牌的五個步驟](https://llmion.com/blog/how-to-do-geo): GEO(生成式引擎優化)是讓 ChatGPT、Gemini、Perplexity、Google AI 摘要在回答中提到你、引用你網頁的做法。從 AI 答案的來源講起,用五個步驟說明 GEO 怎麼做:量測、選主題、寫內容、讓 AI 讀得到、追蹤調整。(2026-11-15) - [讓 AI 讀得到你的網站:爬蟲、robots.txt 與結構化資料檢查清單](https://llmion.com/blog/make-site-readable-by-ai): 讓 AI 讀得到你的網站,需要做到三件事:robots.txt 放行搜尋用的 AI 爬蟲、文字直接寫在 HTML 裡、用結構化資料標示作者與日期。內含 OpenAI、Google、Perplexity、Anthropic 各家 AI 爬蟲的用途對照,以及一份做一次就好的網站檢查清單。(2026-11-15) - [怎麼量測 AI 能見度?問題設計、採樣次數與追蹤方法](https://llmion.com/blog/measure-ai-visibility): 量測品牌的 AI 能見度,就是用一組不含品牌名稱的固定問題,定期問 ChatGPT、Gemini、Perplexity、Claude、Google AI 摘要,記下有沒有提到你、位置、方式、語氣與引用來源。同一題問 3 次以上,發布內容約 4 週後再比較前後差異。(2026-11-15) - [GEO、AEO、SEO 是什麼?三者的定義與差異](https://llmion.com/blog/geo-aeo-seo): SEO 是讓搜尋引擎理解並排名你的頁面,AEO 是讓內容成為直接答案,GEO 是讓內容被 ChatGPT 等生成式 AI 引用。用一張表看懂 GEO、AEO、SEO 的目標、優化對象與成效指標。(2026-11-01) - [AI 引用優化:5 個讓 AI 平台主動引用你的內容的實戰技巧](https://llmion.com/blog/5-tips-ai-citation): 從頂部直接回答到引用可信來源,5 個讓 ChatGPT、Perplexity、Google AI Overviews 更容易引用你的內容的技巧,附 GEO 研究的實測數據。(2026-03-20,更新 2026-11-01) - [ChatGPT vs Gemini vs Perplexity:哪個 AI 平台對你的品牌最重要?](https://llmion.com/blog/chatgpt-vs-gemini-vs-perplexity): 比較 ChatGPT、Google AI Overviews、Perplexity、Gemini、Claude 5 個 AI 平台如何檢索與引用來源,以及 BVS 分數為什麼要按平台加權。(2026-03-18,更新 2026-11-01) - [什麼是 AI 能見度?為什麼你的品牌需要 BVS 分數](https://llmion.com/blog/what-is-ai-visibility): AI 能見度是品牌在 ChatGPT、Google AI Overviews 等 AI 平台的回答中被提及、推薦或引用的程度。BVS 分數用 5 個維度、5 個平台把它量化為 0–100 分。(2026-03-15,更新 2026-11-01) - [GEO Topic Selection: How to Find the Content Gaps AI Still Answers Poorly](https://llmion.com/en/blog/geo-topic-selection) [en]: The rule for picking a GEO topic is to find questions your customers ask often, that AI still answers poorly, and that you know best. Count how many different websites AI cites to judge scarcity, ask AI the question once to find what it misses, blurs or gets wrong, then add the chosen question to your measurement list.(2026-11-15) - [AI Citation Explained: How AI Chooses Which Sources to Cite](https://llmion.com/en/blog/how-ai-chooses-sources) [en]: AI citations are not decided by search ranking. ChatGPT, Gemini, Perplexity and Google AI Overviews rewrite the question into queries, retrieve pages, pick the passages that best answer it, then write an answer with cited sources. Understanding these five stages shows exactly where your page gets dropped.(2026-11-15) - [How to Do GEO: Five Steps to Get AI to Recommend Your Brand](https://llmion.com/en/blog/how-to-do-geo) [en]: GEO (Generative Engine Optimization) is the practice of getting ChatGPT, Gemini, Perplexity and Google AI Overviews to mention your brand and cite your pages. Starting from where AI answers come from, here is how to do GEO in five steps: measure, pick topics, write, make your site readable to AI, and track.(2026-11-15) - [Make Your Site Readable by AI: A Checklist for AI Crawlers, robots.txt and Structured Data](https://llmion.com/en/blog/make-site-readable-by-ai) [en]: Making your site readable by AI takes three things: let AI search crawlers through robots.txt, put your text directly in the HTML, and mark up author and dates with structured data. Includes what each AI crawler from OpenAI, Google, Perplexity and Anthropic does, plus a one-time site checklist.(2026-11-15) - [How to Measure AI Visibility: Question Design, Sampling and Tracking](https://llmion.com/en/blog/measure-ai-visibility) [en]: To measure your brand's AI visibility, ask ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews a fixed set of questions that do not include your brand name, and record whether you are mentioned, where, how, in what tone, and whether your site is cited. Ask each question at least 3 times, and compare before and after about 4 weeks after publishing.(2026-11-15) - [What Are GEO, AEO and SEO? Definitions and Differences](https://llmion.com/en/blog/geo-aeo-seo) [en]: SEO helps search engines understand and rank your pages, AEO makes your content the direct answer, and GEO gets your content cited by generative AI such as ChatGPT. One table compares the goals, targets and success metrics of GEO, AEO and SEO.(2026-11-01) - [AI Citation Optimization: 5 Proven Techniques to Get AI Platforms to Cite Your Content](https://llmion.com/en/blog/5-tips-ai-citation) [en]: From a direct answer at the top to citing credible sources: 5 techniques that make ChatGPT, Perplexity and Google AI Overviews more likely to cite your content, with measured results from the GEO study.(2026-10-31,更新 2026-11-01) - [ChatGPT vs Gemini vs Perplexity: Which AI Platform Matters Most for Your Brand?](https://llmion.com/en/blog/chatgpt-vs-gemini-vs-perplexity) [en]: How ChatGPT, Google AI Overviews, Perplexity, Gemini and Claude retrieve and cite sources, and why the BVS score is weighted by platform.(2026-10-31,更新 2026-11-01) - [What Is AI Visibility? Why Your Brand Needs a BVS Score](https://llmion.com/en/blog/what-is-ai-visibility) [en]: AI visibility is how often a brand is mentioned, recommended or cited in answers from AI platforms such as ChatGPT and Google AI Overviews. The BVS score measures it from 0 to 100 across 5 dimensions and 5 platforms.(2026-10-31,更新 2026-11-01) ## 機器可讀資源 / Machine-readable - [RSS](https://llmion.com/rss.xml) - [Sitemap](https://llmion.com/sitemap.xml) - [robots.txt](https://llmion.com/robots.txt): 明示允許 GPTBot、OAI-SearchBot、ClaudeBot、Claude-SearchBot、PerplexityBot、Google-Extended、Applebot-Extended 等 AI 爬蟲 ## Optional - 引用本站時請以 /bvs 頁面作為 BVS 定義與公式的來源,並註明 llmion 與最後更新日期。 - When citing, use the /bvs page as the source for the BVS definition and formula, and attribute llmion with the last-updated date.