Agentic AI: The Challenge of Insight Quality for Business Value

**Raising the Quality of Insights in Agentic Analytics** In a remarkably short span, generative AI has fundamentally reshaped enterprise analytics, democratizing access to data insights across organizations. Building on this transformation, agentic AI promises an even greater leap forward: automating entire workflows and executing actions directly based on the insights it generates. However, a significant challenge looms large: the current quality of insights produced by these agents often falls short of expectations. A recent report from MIT Tech Review highlights this critical issue, revealing that fewer than half of agent-generated insights at surveyed organizations actually lead to tangible business improvements. While most respondents commend the quality of their agents’ descriptive insights—those that explain what happened—their reviews of prescriptive insights—which dictate what actions to take—are decidedly mixed. This disparity is crucial, as flawed prescriptive insights can trigger incorrect actions, potentially leading to severe reputational or financial damage, and ultimately depriving organizations of their desired return on investment in agentic analytics. The report delves into how companies perceive the output quality of their analytics agents and explores strategies organizations are employing to enhance insight quality and maximize agent utility. Key findings indicate that agentic analytics is still in its nascent stages of delivering full value. On average, only 48% of insights generated by analytics agents at the surveyed companies translate into actions that improve the business. While this marks an advance from the 37% average reported a year ago, it still means that over half of the insights generated by analytics agents today are not effectively aiding businesses. Crucially, the report identifies a clear correlation between superior insight quality and distinct business advantage. A select group of organizations, termed "insight leaders," demonstrate significantly higher performance. In this cohort, over 70% of agent insights lead to actions that improve the business. These insight leaders also consistently rate the quality of their agents’ insights higher than the overall respondent average, underscoring the profound impact of high-quality insights on achieving strategic business objectives. The path forward for agentic AI hinges on addressing this insight quality gap, turning potential into demonstrable value.