Slow, Human Work
Julie Averill was the CIO of Lululemon for eight years. Her guest essay in this weekend’s New York Times is the most honest thing I’ve read about enterprise AI.
Her core observation: AI is genuinely powerful technology that still needs people, and “putting it to use is slow, human work.” The tools keep improving. The gap keeps narrowing. What remains is the part that was always ours: messy data and complicated human decisions.
The numbers back her up. She cites MIT research finding that 95 percent of enterprise generative AI pilots never delivered real results, and a survey in which three quarters of executives admitted their AI strategy is more for show than substance.
There are more vendors, dashboards, and data than anyone can act on, and every new tool adds both signal and noise. Judgment is what separates them, and judgment doesn’t ship in a dashboard.
Don’t get me wrong, I love the tools (we build them too). But let’s not forget where the value sits: people who listen, study, measure, and repeat. Slow, human work.
