---
id: "question-measuring-ai-mentions"
type: "open-question"
source_timestamps: ["§ Shift 2: SEO and Website Design Matter Less and Less"]
tags: ["analytics", "kpis", "measurement"]
related: ["concept-engineering-recall", "framework-engineering-ai-recall"]
resolutionPath: "Development of new analytics tools that simulate user queries across major LLMs (ChatGPT, Claude, Gemini) to track brand share-of-voice in synthesized answers."
sources: ["geo"]
sourceVaultSlug: "hbr-seg-geo"
originDay: 3
articleStem: "hbr-ext-11-llms-overtaking-search"
sourceUrl: "https://hbr.org/2026/03/llms-are-overtaking-search-heres-how-to-adjust-your-online-presence"
sourceTitle: "LLMs Are Overtaking Search. Here’s How to Adjust Your Online Presence."
---
# How Can Brands Accurately Measure Mentions Inside AI-Generated Responses?

**Open question:** The authors advise measuring success by whether the brand and its experts are *mentioned, paraphrased, and associated with key ideas inside AI-generated responses* (step 7 of [[framework-engineering-ai-recall]]). But because LLM interactions are typically **private, ephemeral, and personalized 1-to-1 chats**, the source does not explain the technical mechanism or tooling to track these mentions at scale — the measurement gap at the heart of [[concept-engineering-recall]].

**Resolution path:** New analytics tooling that **simulates user queries across major LLMs** ([[entity-chatgpt-d11]], [[entity-claude-d11]], [[entity-google-overviews]]) to track brand *share-of-voice* in synthesized answers.

**Grounding (enrichment):** Real hurdle confirmed — McKinsey notes only **16% of brands** systematically track AI-search performance today and recommends panel testing plus platform-provided metrics. Until platforms expose standardized analytics, these KPIs stay approximate and potentially biased by test methodology (Semrush/McKinsey). Semrush and others already ship 'AI-visibility' toolkits as an early answer.


## Related across articles
- [[question-web-analytics-replacement]]
- [[concept-agentic-observability]]
- [[question-som-volatility]]