Skip to content
MIT Technology Review·

📊Analytics Agents Hit 48% Business Impact, Up From 37% in One Year

Half your agent insights still go nowhere. Here's why.

TL;DR

Analytics agents now drive business-improving actions on 48% of their insights, up from 37% a year ago. That still means more than half get ignored or fail. For teams betting on agentic analytics, the fix is context and governance, not more agents.

Analytics agents are working, but barely more than half the time. Across surveyed organizations, an average of 48% of agent-generated insights now lead to actions that improve the business. A year ago that number was 37%, so progress is real. It's also a long way from good. If you're wiring agents into dashboards or pipelines, this is your problem to own. Descriptive insights get praised. Prescriptive ones, the recommendations that tell someone what to do, are where quality falls apart. Flawed insights don't just waste time. They can cause reputational or financial damage, and an agent that confidently suggests the wrong move at scale is worse than no agent. The root causes are boring and fixable. 55% of surveyed executives blame lack of business context, and 45% point to fragmented or inaccessible data. At a majority of organizations, agents generate more insights than people can evaluate, and fewer than a quarter can check all of them. Only a small cohort tops 70%. Their edge is governance (cited by 60% as a priority) and a semantic layer (45%).

Analytics Agents Hit 48% Business Impact, Up From 37% in One Year — MIT Technology Review

Key Points

1

Across surveyed organizations, the average share of agent insights leading to business-improving actions reached 48%, up from 37% in the past year.

Keep reading — it's free

Enter your email to keep reading — plus the best of AI & tech, daily. Free, forever.

Also get
or

Already a member? Sign in

agentic-aianalyticsdata-governancesemantic-layerenterprise-ai

Comments

Subscribe to join the conversation...

Be the first to comment