According to a column by ITmedia NEWS, an increasing number of workplaces feel that their workload is not being reduced despite implementing AI. It points out that the time supposedly saved may be being absorbed by other tasks.
Factors Preventing Realization of Workload Reduction via AI Implementation Lie in Management Costs and Flawed Measurement Methods
This article is a translation. Read the Japanese original
In terms of measuring reduction effects, the failure to survey baseline man-hours prior to implementation is being viewed as a problem. As a result, reports are often limited to minor achievements captured by chance, and the actual benefits fail to justify the development costs.
Another factor is that the intended use of the saved time has not been defined. Without rules to allocate the reduced time toward reducing overtime or other purposes, new tasks will inevitably intervene.
Discrepancies between management and frontline staff regarding the metrics for measuring effectiveness have also been cited. While management tends to focus on costs and headcount, frontline workers view the reduction of task time as the primary achievement, creating a situation where results are not evaluated correctly.
Furthermore, there are cases where shifting to in-house development due to the cancellation of SaaS subscriptions is actually increasing maintenance man-hours. It is reported that new tasks arising from AI implementation—such as establishing usage guidelines, managing logs, and overseeing pay-as-you-go billing—are offsetting the time saved.