Agentic analytics struggle with context, half of insights miss business impact

A MIT Technology Review study reports that fewer than half of insights produced by analytics agents actually improve business performance, with 48% leading to positive actions—a rise from 37% a year earlier—but still leaving the majority ineffective Raising the quality of insights in agentic analytics. Executives surveyed identified missing business context as the top cause, cited by 55% of respondents, while 45% pointed to fragmented or inaccessible data.

A small “insight leaders” cohort outperforms the average, achieving over 70% of agent insights that drive improvement and rating their agents’ quality higher, especially for prescriptive recommendations. However, 52% of organizations say agents generate more insights than can be evaluated, and only 24% can review all outputs, a process described as human-intensive.

To address these gaps, 60% of companies are investing in AI and data governance, and 45% are adding semantic layers to their data architecture, a move experts say is essential for higher-performing agents.

Our read: The findings underline that scaling agentic analytics without embedding robust semantic context and governance yields diminishing returns, and firms that prioritize these foundations are already seeing markedly better business outcomes.