---
id: "claim-ai-visibility-fragmented"
type: "claim"
source_timestamps: ["§ AI Recommends What It Can Interpret"]
tags: ["research-finding", "platform-variance"]
related: ["claim-ai-infers-positioning-externally"]
confidence: "high"
testable: true
sources: ["geo"]
sourceVaultSlug: "hbr-seg-geo"
originDay: 3
articleStem: "hbr-new-25-get-ai-to-surface-your-brand"
sourceUrl: "https://hbr.org/2026/06/how-to-get-ai-to-surface-your-brand"
sourceTitle: "How to Get AI to Surface Your Brand"
---
# AI brand visibility is highly fragmented across platforms

In a study of **15 retail categories** using identical prompts across **GPT-4o, Claude, and Gemini**, researchers found that "AI visibility" is highly fragmented. Out of **716 unique brands** surfaced, only **8.4% appeared consistently across all three platforms**. Most brands appeared on only one platform, meaning a brand that seems dominant in ChatGPT might be entirely absent in Claude.

This indicates that traditional "visibility" metrics do not translate neatly into the AI era, and that models require highly structured, credible evidence to converge on a single brand as a reliable answer. It sets up the companion finding that [[claim-ai-infers-positioning-externally|AI infers positioning from third-party data]].

**Confidence:** high · **Testable:** yes.

> Enrichment note: There is no independent, published replication of the exact 15-category / 716-brand / 8.4% statistic yet — the study appears original to this HBR article. However, the *directional* claim is strongly consistent with independent evaluations showing substantial cross-model divergence in answers and named entities for identical prompts (different training data, safety filters, search integrations), mirroring classic search fragmentation across Google, Bing, and DuckDuckGo.


## Related across articles
- [[claim-llm-processing-styles-vary]]
- [[concept-ai-model-segmentation]]
- [[claim-model-idiosyncrasy]]