Tech

Raising the quality of insights in agentic analytics

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 02/11/2026, 17:20:24 UTC.

Briefings

A recent report from MIT Technology Review's Insights arm reveals that while agentic AI is improving, only 48% of insights currently lead to business improvements, a 11-percentage-point increase from one year ago. Organizations labeled 'insight leaders' achieve 70% success rates, significantly outperforming the average. The primary barriers to high-quality insights are insufficient business context (cited by 55%) and fragmented data (45%), with governance and semantic layers being key strategic priorities.

Key points

• Only 48% of insights generated by analytics agents at surveyed companies lead to business improvements, representing a 11-percentage-point increase from the 37% average one year prior.
• Insight leaders achieve a 70% success rate for agent insights, significantly outperforming the overall average of 48% and showing higher quality marks for their agents' descriptive and prescriptive insights.

Context

This report from MIT Technology Review's Insights arm, published November 2, 2026, examines the quality of insights produced by enterprise agentic AI systems. It highlights that while descriptive insights are generally praised, prescriptive insights—which drive actions—remain mixed. The findings are critical for understanding the gap between AI capabilities and actual business value, with 'insight leaders' identified as a specific subset of organizations achieving superior performance.

Limitations

The evidence is limited to a survey of organizations surveyed for this report, with specific data points such as the 48% success rate and 70% success rate for insight leaders derived from this sample. The report does not provide causal claims or opinions on why these outcomes occurred, nor does it offer forecasts beyond the current survey period.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 02/11/2026, 17:23:31 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeidtdiqbpewzbv3ee47u7jevn3y7aqt4gogrighwfkn37lig5ajjze