Generative AI has transformed enterprise analytics by making data tools accessible across organizations. Agentic AI promises to automate workflows further, yet the quality of insights generated by these systems often falls short. A recent report indicates that fewer than half of agent-generated insights lead to actual business improvement. While descriptive insights generally receive praise, prescriptive insights—those intended to drive action—receive mixed reviews.
Current performance metrics
Agentic analytics is only just starting to deliver measurable results. On average, 48% of insights generated by analytics agents at surveyed companies lead to actions that improve the business. This marks an advance from the 37% average reported one year ago. However, it also means that more than half of all agent-generated insights fail to provide clear value. Failure to produce insights that improve the business deprives organizations of their desired return on investment.
Companies are taking a closer look at the quality of insights produced by their analytics agents, and exploring ways to improve their performance. A key finding is that better-quality insights lead to a significant business advantage. For example, at a select group of organizations, known as “insight leaders,” AI agents outperform the average, with over 70% of their insights leading to business improvements. In contrast, the average company sees less than half of its agent-generated insights leading to positive actions. The gap in quality is particularly noticeable when it comes to prescriptive insights, which are used to inform decision-making.
Related Post: Google’s Secret AI Dictation App Goes Live: Revolutionary Offline Voice-to-Text Technology Unveiled
Low-quality insights are often the result of insufficient context, with executives citing a lack of business context as the primary reason for subpar insights. In contrast, fewer “insight leaders” report struggling with access to context. Additionally, the sheer volume of insights generated by analytics agents can be overwhelming, with 52% of organizations receiving too many insights to properly evaluate and act upon. Only 24% of companies are able to assess all agent insights, with most relying on a human-intensive, case-by-case approach.
Addressing the gap
The surveyed companies are prioritizing initiatives that enable better agent insights by strengthening AI and data governance. They are also adding a semantic layer to their data architecture. Experts interviewed for the report agree that the semantic layer is essential to ensure high-performing agents. A semantic layer helps define the meaning of data, allowing agents to interpret information more accurately.
This approach addresses the root causes of poor insights, such as fragmented data and a lack of context. When analytics agents deliver insights of substandard quality, the main reason is a lack of business context, as cited by 55% of surveyed executives. Fragmented or inaccessible data is the other major factor, according to 45%. They are also adding a semantic layer to their data architecture (45%).
