Neo4j and MITTR released a report highlighting the critical role of knowledge in enterprise AI agents. The report, based on a survey of 300 executives, shows that knowledge gaps are a major barrier to deploying AI agents in production. Without sufficient knowledge, AI agents struggle to make reliable decisions, leading to stalled projects and lost competitive advantage.
The report found that data and knowledge weaknesses consistently stall AI agent progress. On average, only around a third (34%) of organizations’ agentic AI projects make it into production. Even high-tech firms face challenges, with legacy data systems, security concerns, and a lack of context being key points of failure.
Strong knowledge capabilities correlate with agent success. A small group of production leaders, where 61% of agentic projects advance beyond pilot, have stronger knowledge capabilities, especially in semantics. This advantage closely tracks with their higher production rates, indicating a clear link between knowledge and AI effectiveness.
"Data fragmentation was most commonly cited as a top challenge to expanding agents’ access to knowledge," said the report. "Production leaders, by contrast, are more likely to see security and privacy concerns as a major issue." This suggests that while data fragmentation is a widespread problem, some organizations are prioritizing security as a key challenge.
Most firms aim to strengthen the link between data and agents. Executives expect the biggest impact from improving the structural foundation between an organization’s data and its AI agents. A knowledge layer is seen as a prime way to achieve this, with investments in retrieval technologies, AI-ready APIs, and knowledge graphs being prioritized.
The report does not say how to fully resolve these knowledge gaps, and it raises questions about the long-term viability of agentic AI projects without robust knowledge systems. The findings suggest that addressing these challenges is essential for scaling AI in enterprise settings.
Source: mittr
