BCG’s new study, When Everyone Uses AI, Companies Risk Losing Critical Skills, names the danger distributed de-skilling — the collective erosion of human capability across an organisation, not just in one employee. In a global study of 70 C-suite and senior leaders, half are already observing it and more than 60% expect it to be a material threat within three to five years.
The unsettling finding is a paradox: the skills leaders rate most important to long-term performance — problem understanding & framing and judgement & decision-making, plus creative thinking and causal reasoning — are precisely the ones most exposed to AI. BCG frames this as a system-design problem, not a talent problem, and lays out six mitigation strategies.
Yet only one in ten companies has an organisation-wide strategy to address it, and a third haven’t explicitly discussed it — largely because the erosion is invisible on the surface: output still ships while the capability underneath quietly thins.
Read BCG’s six fixes together — governance, work redesign, performance systems, replenishment rituals, nonlinear AI use, reflective tooling — and one dependency runs under all of them: they assume you can see whether the human skill is holding or fading. Most organisations can’t. BCG names the exact gap — “certification rates go up while actual capability may not.”
That’s the layer GoMeasure supplies. We baseline and continuously measure human capability from demonstrated work — assessments, AI interviews, work simulations, 360° and telemetry — so judgement, problem-framing and reasoning become something you can watch move by team and cohort, and re-verify after training. It’s how distributed de-skilling becomes visible while it’s still reversible — and how you point an AI-off zone where the debt is actually accruing.
We don’t sell AI policy or workflow redesign — BCG’s strategies are the right shape for that. GoMeasure is the measurement layer underneath: baseline and re-verify the human, so leaders can prove capability is holding, not just hope it is.
