AI Projects Fail 80% of Time Due to Poor Data Management, New Report Reveals

Dec 31, 2025
dbt Labs
Article image for AI Projects Fail 80% of Time Due to Poor Data Management, New Report Reveals

Summary

New O'Reilly report reveals AI projects crash at an alarming 80% failure rate due to chaotic data management, while major companies like Walmart and Block deploy structured frameworks and 10-step roadmaps to rescue their artificial intelligence initiatives from costly disasters.

Key Points

  • AI projects fail over 80% of the time primarily due to fragmented, inconsistent, or poorly governed data according to a new O'Reilly report
  • Major companies like Walmart, Block, and NBIM are using MCP and structured context frameworks to scale AI safely with unified metrics, lineage tracking, and policy-driven access controls
  • The report provides a 10-step roadmap for building compliant AI-ready data stacks that include conversational analytics, governance frameworks meeting EU AI Act requirements, and semantic layers for reliable AI outputs

Tags

Read Original Article