The Missing Half OF AI
Alliance · October 13, 2026

84% haven't redesigned work

Add an AI step. Remove nothing. Wait for the results.

Deloitte's State of AI in the Enterprise 2026 asked 3,235 leaders in 24 countries. 84 percent have not redesigned jobs or workflows around AI.

That isn't laziness. Adding is nearly free now: a copilot on every seat, rolled out over a weekend. Subtracting is expensive, because someone has to decide what a person no longer does, and who agrees to it. So the addition ships and the decision never gets made.

The cleanest look at what that costs: Brynjolfsson, Li and Raymond studied 5,172 customer-support agents given an AI assistant, in an otherwise unchanged workflow. Productivity rose 15 percent on average. But the average hides the finding. The least experienced improved on both speed and quality. The most experienced gained little speed, and lost some quality.

Same tool, same workflow, and it helped one half of the room more than the other. That is not a failure of the people using it. It's a question nobody asked on their behalf: where does experience go now?

One move, the shortest on this list: for every AI step you added, name the human step you removed, and who agreed to it. If nothing was removed, you didn't redesign the work. You added to it.

Alliance isn't a tool in a seat. It's a decision about who does which half, made on purpose by someone with the authority to take something away.

Next Tuesday: chapter twelve. Twelve weeks of evidence, research and real cases, from Intent to Alliance, brought together in one picture.

💬 What did you stop doing entirely to make room for AI? One line is enough. "Nothing" is an honest answer.

#TheMissingHalf #Alliance #FutureOfWork #HumansAtCenter #WorkflowDesign

The numbers in this piece

Every figure named above, with its source. Links go to the primary study.

84%
84 percent of organizations have not redesigned jobs or workflows around AI
+15% (QJE 2025)
Support agents gain 15 percent productivity on average. The distribution is the actual finding: the least experienced improve both speed and quality, while the most experienced gain little speed and lose some quality.
← One team, not two halves