---
id: "claim-ai-scaling-failure"
type: "claim"
source_timestamps: ["§ Effect #3: Undershot Company Targets"]
source_url: "https://hbr.org/2025/09/dont-let-ai-reinforce-organizational-silos"
source_title: "Don't Let AI Reinforce Organizational Silos"
tags: ["statistics", "ai-scaling", "failure-rates"]
related: ["concept-siloed-ai-implementations", "entity-kim-oosthuizen"]
confidence: "high"
testable: true
speakers: ["Kim Oosthuizen", "Graham Kenny"]
sources: ["tail2"]
sourceVaultSlug: "hbr-seg-tail2"
originDay: 2
articleStem: "hbr-tail-130-ai-reinforce-silos"
sourceUrl: "https://hbr.org/2025/09/dont-let-ai-reinforce-organizational-silos"
sourceTitle: "Don’t Let AI Reinforce Organizational Silos"
---
# 70 percent of AI initiatives fail to scale beyond initial deployment

**Claim:** 70% of AI initiatives fail to scale beyond their initial deployment. — *Confidence: high (as stated by the authors) · Testable: yes*

Based on research conducted by co-author [[entity-kim-oosthuizen]], 70% of AI initiatives fail to scale beyond their initial deployment. The authors assert the precise reason is that initiatives are implemented and measured within [[concept-siloed-ai-implementations]], preventing compound, cross-functional effects.

**Enrichment validation — IMPORTANT:** This specific statistic is **not independently validated** by the provided web sources. It should be treated as an *unverified author claim* unless the underlying Kim Oosthuizen research can be identified. Counter-perspective from the enrichment: experts would want the original study, its sampling method, and its definition of “scale” before using the figure as evidence. Downstream: attribute this number to the authors, not to consensus, and flag the sourcing gap if a user leans on it.