Conor Grennan. NYU Stern Dean of Students and AI adoption educator writing at ai-mindset.ai. His core thesis: AI adoption is a behavioral problem, not a technical one. Organizations underinvest in defining what "better work" looks like per role before deploying AI, then struggle to demonstrate ROI after the fact because they have no baseline to compare against.
Grennan writes primarily for leaders navigating enterprise AI rollouts where the pressure to show results outpaces the discipline to set up the measurement conditions for results to be verifiable.
Operating themes
- Target-setting precedes measurement ROI cannot be measured retrospectively if no definition of success was established before deployment.
- AI fits the person, not the process Unlike workflow software, AI's gains appear in qualitative improvements to individual work, not in auto-generated output metrics.
- Behavioral adoption over technical capability The bottleneck in most AI rollouts is people's relationship to the tool, not the tool's underlying capability.