Analytical agents are generating more business-improving insights than they did a year ago, but most of their output still does not lead to useful action. A new report examines how enterprises assess the quality of insights produced by agentic analytics systems, the obstacles that reduce their effectiveness, and the steps organizations are taking to help these agents deliver greater business value.
Key findings from the report
Agent analytics remains in its early stages
On average, 48% of the insights generated by analytics agents at surveyed companies led to actions that improved the business. That represents an increase from the 37% of respondents who said they had taken such action a year earlier. Even so, more than half of the insights produced by today’s analytical agents still have no business use.
Higher-quality insights can improve business outcomes
Smaller organizations identified in the report as “insight leaders” said their AI agents perform substantially better than average. In these organizations, more than 70% of agent-generated insights lead to actions that improve the business.
Insight leaders also gave their agents higher ratings for insight quality than respondents overall. The gap was particularly pronounced for normative insights, which recommend what an organization should do.
Limited context affects insight quality
A lack of business context was the most frequently cited reason for substandard insights from analytics agents, identified by 55% of executives surveyed. Another 45% pointed to fragmented or inaccessible data as a major factor.
Few insight leaders reported that their agents struggle to access the necessary context, highlighting the importance of connecting analytical systems to relevant business information.
Too many insights can limit effectiveness
Analytical agents generate more insights than organizations can readily evaluate in 52% of the companies surveyed. Less than a quarter, or 24%, said they can assess all of an agent’s insights.
In most organizations, insight evaluation is handled case by case and remains a labor-intensive process. The volume of output can therefore make it harder for teams to identify and act on the most important recommendations.
Governance and semantic layers are priorities
Companies surveyed are focusing on stronger AI and data governance to improve the quality of agent-generated insights, with 60% identifying this as a priority. A further 45% are adding a semantic layer to their data architecture.
The experts interviewed for the report agree that a semantic layer is critical to achieving high agent performance because it helps analytical systems interpret data in a consistent business context.
This content was created by Insights, the custom content division of MIT Technology Review, and not by editorial staff. It was researched and written by humans, and any AI tools it may have used were limited to production processes under human supervision.
Source: www.technologyreview.com