The automation project has been successful. The software handles document review, the error rate has decreased, and approximately one-fifth of working hours have been freed up in the affected area. In the steering committee, a single question is asked: How many positions can be eliminated? It is the expected question, and it is justified. However, the second question, which would be at least as valuable, remains unasked: What could these people achieve that has been left undone for years because no one ever had time for it? The decision of what automation ultimately brings to the company depends on these two questions. <\/p>\n\n
\nAutomation answers how work will be done in the future. It does not answer what happens to the freed-up time. This decision is made by management, and it shapes every automation thereafter. <\/strong><\/p>\n<\/blockquote>\n\n
A company whose management I advised<\/a> during this phase immediately cut positions after the first major automation project. The savings appeared in the next quarterly report, and the calculation seemed to work out. Two years later, the second project started and made little progress. Expertise that automation would have needed suddenly became difficult to obtain, process details remained unmentioned, and testing phases dragged on. No one openly refused to cooperate. The workforce had simply learned from the first project: those who train the machine well make themselves dispensable. The second automation paid the price of the first. <\/p>\n\n
This pattern is not solely a matter of corporate culture. It follows a logic that has repeated itself for two centuries. Three levers turn it into a controllable decision. <\/p>\n\n
The Lesson from Two Centuries<\/h2>\n\n
In his seminal work “The Technology Trap,” Carl Benedikt Frey examined the history of automation from industrialization to the present day and revealed an uncomfortable regularity: whether people support or resist technological change depends not on the technology itself, but on whether they share in its benefits. The weavers who, as so-called Luddites, stormed machines in the early 19th century were not enemies of progress. They bore its costs without participating in its benefits and resisted this distribution. Where change was shared with those affected, it accelerated. <\/p>\n\n
For a company, this history translates into a sober context: every automation sends a signal about what participation means. If it reliably leads to job cuts, the organization learns to slow down change, quietly and effectively, as in the example above. If it leads to new, more valuable tasks, the organization learns to accelerate change. The decision about freed-up capacity is therefore not merely a personnel issue. It is the investment decision that co-determines the pace of all future projects, including those reshoring efforts that AI is making possible again<\/a>. <\/p>\n\n