Operational framework #2
Time-to-Skill
Measuring ramp-up duration as a first-rank budget metric.
Time-to-Skill is the average duration, expressed in days or weeks, an employee needs to reach target productivity on a new work gesture. Measuring it is today as rare as it is indispensable: no organisation can estimate the return on an AI investment without knowing the human latency that separates deployment from real yield.
The principle states that as long as training is discussed in hours attended rather than productivity reached, effort and outcome are conflated. Time-to-Skill flips the perspective: it measures what training produces, not what it costs. It turns time into a currency — the only one that compares directly with the opportunity cost of an AI transformation.
In practical terms, Time-to-Skill is computed per work gesture, not per individual. Three moments are instrumented: T0 (initial exposure), T1 (first productive act without assistance) and Ttarget (target productivity, defined ex ante). The gap Ttarget minus T0 becomes the input for every planning trade-off: should we buy, train or hire to fill a given competence?
One example: an insurer that measured Time-to-Skill on the use of an underwriting copilot found that the internal median was 47 days, while the top decile reached 142 days. Rather than extending training for the long tail, the organisation redeployed part of the budget toward an HRIS overhaul, which reduced the median to 28 days and capped the top decile at 71. AI ROI turned positive eleven months earlier than planned.
Time-to-Skill is to be measured, published and committed to. It belongs in every HR steering review on the same line as turnover and cost-per-hire.