MIT Sloan Management Review reports research by Jennifer Sloan and Vern L. Glaser, published 5 August 2026, examining how professionals generate insight when working with generative and agentic AI. The authors’ underlying study identified four pathways by which AI-assisted analysis produces surprising insight during expert work. A separate study of organisations as algorithmic assemblages examined how accessible data, configured capabilities and distributed agency shape what such systems reveal. The authors write that professional expertise, while making practitioners effective, also narrows the frame through which they view a problem, so that insights arise most often at the edges of that frame, in friction between competing interpretations. They describe conversational AI as surfacing such friction within a bounded exchange, while agentic systems, configured and directed rather than prompted, sustain analysis across larger data sets without losing that thread.
Source: MIT Sloan Management Review, 2026-08-05.