AI Agents: The Challenge of Reliable Data for Businesses

The Rise of Agentic AI and Its Challenges

Business and technology leaders are now convinced that the era of agentic AI has indeed begun. This technology, which enables AI agents to take autonomous actions, is being rapidly adopted by organizations. However, the promise of significant return on investment hinges on a robust data infrastructure. Current data systems, often inadequate, pose a major barrier to fully leveraging this technology.

Agentic AI imposes unprecedented requirements on enterprise data systems. Unlike traditional systems that merely respond to queries, these agents must take actions based on data from across the organization. This includes both structured and unstructured data, enriched with the right business context. To be effective, agents need real-time access to operational systems, such as those in supply chain or human resources. Unfortunately, even recently updated data systems struggle to meet these new demands.

The Urgency of Modernizing Data Systems

As AI agents integrate into business operations, the need to modernize data systems becomes crucial. Gartner predicts that by 2027, these agents could automate or augment 50% of business decisions. To avoid depriving agents of the necessary data for quick and accurate decisions, companies must eliminate bottlenecks associated with legacy systems.

A survey of 300 data and technology executives highlights how legacy systems limit the effectiveness of AI agents. It reveals that only a few organizations, labeled as "data leaders," manage to overcome these limitations. These leaders are paving the way by creating a data environment conducive to the prosperity of agents and the development of trustworthy systems.

Key Findings: Data Access and Trust

• Currently, few companies provide agentic AI with sufficient access to data. On average, AI accesses only 45% of enterprise data. This figure drops to 30% or less for "data laggards," while "data leaders" ensure access to over 70% of their data, thus achieving better results with their agents.

• Trust in agents' decisions is directly linked to data readiness. Only 50% of surveyed organizations trust the accuracy of their AI agents' decisions. In contrast, 100% of data leaders trust their agents, underscoring the importance of a reliable data foundation for reliable AI.

• Data leaders find it easier to scale and accelerate their agents. Two-thirds of laggards report that legacy systems limit the scalability (66%) and speed (68%) of AI agents. Leaders, having overcome these constraints, report only an 8% limitation.

Preparing for the Future: Data Infrastructures and AI

The pressure to make data infrastructures ready for agents is strong. In the next two years, all respondents plan to use agentic AI, with 69% expecting widespread use. Without addressing the constraints of data systems, agentic AI will not be able to deliver on its promises of speed and efficiency.

Improving data access and business context is a major priority. Key initiatives include enhancing access to both structured and unstructured data for AI agents, as well as data and AI governance. Data leaders are also focusing on automating data management, a crucial element for enabling agents to operate at scale.