The productivity pressure paradox names the pattern in which a mandate that raises expectations without granting time to build skill undermines the gains it seeks. AI Writes Faster Than Humans Can Review: A Longitudinal Study of an Enterprise “2×” Mandate attributes the term to Miller and colleagues, citing it as one of the mechanisms supplied by qualitative work on developer AI-tool use, and restates it in its own practical guidance: demanding the target before the skill accrues is what produces the paradox.
Usage
The term is invoked in that paper to explain why the timing of an AI Mandate matters as much as its content. Its own measurements give the paradox an empirical shape without setting out to test it: the throughput gain the authors observe is not a step that arrives on announcement but a return that grows with accumulated use, rising from roughly 1.5× at three and six months to about 2× by nine months on tool, with the mandate leaving only a small residual once adoption and accumulated use are accounted for. On that reading, an organization that judges a mandate on a short horizon is measuring the period in which the skill has not yet accrued.
The recommendation drawn from it is to target the intensity of use — enablement, friction removal, protected time on tool — rather than adoption alone, and to judge the mandate over a horizon long enough for the gain to accumulate.
The wider literature the paper places this in holds that the payoff from a general-purpose technology depends on complementary investment: large IT investments are described as producing measurable gains only alongside reorganized workflows, new skills and decentralized decisions. The paper characterizes the evidence for the paradox specifically as largely qualitative or drawn from adjacent domains, and identifies as open the question of whether the developer-side accumulating channel can be separated quantitatively from the model-side one.