Snowflake report: 48% of agentic analytics insights improve business

The report identifies a significant gap between average performers and a small group of "insight leaders," where over 70% of insights lead to positive outcomes. Executives cite a lack of business context as the primary reason for low-quality insights, with 55% of respondents pointing to this factor. Another 45% blame fragmented or inaccessible data. Additionally, 52% of organizations report that agents generate too many insights to be properly evaluated, creating an overload that hinders effectiveness.

To address these issues, companies are prioritizing better AI and data governance, cited by 60% of respondents, and adding semantic layers to their data architecture, cited by 45%. Experts interviewed for the report agree that semantic layers are crucial for ensuring high-performing agents. The content is noted as custom content produced by MIT Technology Review's Insights arm, not its editorial staff.