Unified Data and System Connectivity Identified as Critical AI Barriers
Fragmented systems and siloed data are a growing strategic risk for organizations seeking to deploy AI. A new survey of healthcare technology leaders reveals that nearly 90% see an urgent need to modernize connected systems. This challenge is echoed in other sectors, where experts caution that applying AI before establishing a unified data foundation leads to unreliable results and expensive projects that fail to deliver ROI.
- The challenge of data fragmentation is compounded in legacy enterprise environments, which often rely on tightly coupled architectures and batch-based processing, conflicting with the real-time, data-intensive nature of modern AI solutions. Data preparation, largely due to these siloed systems, can consume up to 80% of the time in an AI development lifecycle. - In response to data fragmentation, enterprises are adopting "data fabric" architectures, which create a unified layer to connect and manage data across disparate systems without requiring a complete overhaul of existing infrastructure. This approach supports the real-time, high-quality data access necessary for training reliable AI and machine learning models. - Agentic AI, which can set its own goals and make decisions, is seen as the next phase of enterprise AI, with Gartner predicting it will be a key part of 33% of business software by 2028. This shift may lead to new software pricing models based on a "per-agent" rather than a "per-user" basis. - As AI adoption grows, with over 73% of organizations using or piloting AI, robust AI governance frameworks are becoming critical. These frameworks are essential for managing compliance with regulations like GDPR and HIPAA, ensuring ethical use, and mitigating security risks in production environments. - The integration of AI into legacy systems introduces significant security and compliance risks, as older environments often lack the necessary controls for data privacy and auditable model outputs. A cross-functional governance council, including stakeholders from legal, compliance, and data science, is a recommended best practice to address these challenges. - Globally, AI adoption is accelerating, with one 2024 report indicating that 35% of businesses have fully deployed AI in at least one function and 42% are actively experimenting with it. In the European Union, the percentage of enterprises using AI increased from approximately 8% in 2023 to 13.5% in 2024.
