Raising the quality of insights in agentic analytics
Reported by MIT Technology Review1 source ↓
A MIT Technology Review Insights report says 48% of analytics-agent insights at surveyed companies lead to actions that improve the business, up from 37% a year earlier. It identifies missing business context and fragmented data as key limits.
Key points
- Only 48% of agent insights lead to actions that improve business outcomes, the report says
- At insight-leading organizations, more than 70% of agent insights lead to business-improving actions
- 55% of surveyed executives cite insufficient business context as a cause of poor insights
- 52% of surveyed organizations say agents produce more insights than they can properly evaluate
- Companies are prioritizing stronger AI and data governance (60%) and semantic layers (45%)
Details
The MIT Technology Review Insights report says the share of analytics-agent insights leading to business-improving actions rose from 37% one year earlier to 48%. The report says the figure means more than half of agent-generated insights still do not lead to actions that improve business outcomes.
The report describes a small group of “insight leaders” where more than 70% of agent insights lead to business-improving actions. It says the difference in insight quality is especially wide for prescriptive insights, which are used to guide actions.
Among surveyed executives, 55% cite insufficient business context as a reason for substandard insights, while 45% point to fragmented or inaccessible data. At 52% of surveyed organizations, agents generate more insights than staff can properly evaluate; fewer than one-quarter can evaluate them all.
The surveyed companies are focusing on AI and data governance (60%) and adding a semantic layer to data architecture (45%), according to the report. MIT Technology Review says the report was produced by its custom content arm, Insights, rather than its editorial staff.
Written from coverage as of