A new research study finds that nearly half of all artificial intelligence projects fail, underperform, or are delayed, despite major investments in AI. The survey indicates that poor data readiness is the leading roadblock to AI execution, driving increased costs, stalled innovation, and lost revenue. The study was released by data management company Fivetran and was conducted by Redpoint Content.
Even though 57 percent of organizations rate their data centralization strategy as highly effective, nearly the same proportion say that over half of their AI projects fail to deliver. The disconnect is data that is not fully centralized, governed, or made available in real time for AI models. From integration bottlenecks to pipeline maintenance burdens, enterprises are stuck managing infrastructure instead of delivering business value through AI.
Key findings:
- 42 percent of enterprises say more than half of their AI projects have been delayed, underperformed, or failed due to data readiness issues
- 68 percent of organizations with less than half of their data centralized report lost revenue tied to failed or delayed AI projects
- 67 percent of centralized enterprises allocate over 80 percent of their engineering resources to maintaining data pipelines
- 59 percent of enterprises say regulatory compliance is their top challenge in managing data for AI
Ambition Without Execution
AI underperformance is not just a technical problem. It is a business risk. The research found that:
- 38 percent of enterprises report increased operational costs due to AI project failures
- Reduced customer satisfaction and retention was the most common consequence of failed AI projects
The report calls on enterprises to modernize their data infrastructure with automated integration tools that reduce pipeline complexity and free up engineering resources. Among the top investment priorities cited by respondents:
- 65 percent plan to invest in data integration tools as their primary strategy to enable AI
- Nearly three-quarters of enterprises manage or plan to manage more than 500 data sources, amplifying the need for scalable, automated solutions
What’s really blocking AI success
The survey found that many enterprises are struggling to move beyond pilot AI projects because they cannot efficiently prepare, integrate, or operationalize their data. The data revealed several key pain points:
- 74 percent of enterprises manage or plan to manage more than 500 data sources, creating significant integration complexity
- 67 percent of highly centralized enterprises still spend over 80 percent of their data engineering resources maintaining pipelines, leaving little time for AI innovation
- 41 percent of organizations report the lack of real-time data access prevents AI models from delivering timely insights
- 29 percent of enterprises say data silos are blocking AI success
Until these challenges are addressed, organizations will continue to struggle with AI performance and fail to unlock the full value of their investments.
Regional and industry differences in AI readiness
These issues are not limited to any one sector. Industries like healthcare and retail are leading in AI readiness due to stronger automation and data integration strategies. Sectors such as finance and manufacturing continue to struggle with legacy systems and integration constraints.
Regional differences are also significant. The Asia-Pacific region leads all others with an AI readiness score of 8.8 out of 10, followed by the United States at 8.2. The United Kingdom trails with a score of 6.0 due to weak integration strategies and fragmented infrastructure.
The survey was conducted in the first quarter of 2025 by Redpoint Content. It gathered responses from 401 data leaders and professionals across the United States, United Kingdom, Europe, the Middle East, and Africa and the Asia-Pacific region. Respondents represented enterprises in technology, finance, healthcare, retail, and manufacturing, ranging from 500 to more than 5,000 employees. 