A new study from Confluent reveals that 67% of UK IT leaders see siloed, fragmented data systems as the primary obstacle to fully leveraging AI and machine learning (ML).
The survey of 550 UK IT decision-makers highlights how disjointed data ownership and management hinder the development and deployment of effective AI solutions.
Key challenges identified include:
• Fragmented data ownership across disparate systems: 67%
• Uncertainty around data lineage, quality, and timeliness: 63%
• Lack of skills and expertise in managing AI workflows: 61%
• Difficulty integrating new data sources seamlessly: 57%
• Limited infrastructure for real-time data processing: 50%
Despite these hurdles, 82% of respondents agree that AI success depends heavily on enterprise data to realize its full potential. Without high-quality, real-time data access, the growth of AI-driven applications and analytics is severely constrained — 63% strongly believe AI use in business will grow significantly, and 41% see an emerging role for AI agents within their organizations.
To address these issues, decision-makers see Data Streaming Platforms (DSPs) as a key enabler. About 64% agree that DSPs simplify access to diverse data sources, 55% believe they help track data provenance and lineage, and 51% see DSPs as vital for ensuring data quality and timeliness.
“Data fragmentation has long been a challenge, but in the AI era, it is more critical than ever. Without seamless, real-time access to high-quality data, AI initiatives risk failure, operational performance suffers, and companies fall behind. To unlock AI’s full potential, organizations must focus on breaking down data silos and modernizing their infrastructure,” said Richard Jones, VP of Northern Europe at Confluent.
This research underscores that modern data architecture — especially the adoption of DSPs — is essential for organizations aiming to harness AI effectively and stay competitive in an increasingly digital landscape.
