The article addresses the gap between companies' declared adoption of artificial intelligence and the effective transformation of their processes. Many companies claim to use AI, but few can demonstrate with concrete numbers what changes resulted from this adoption. This phenomenon is described as one of the most relevant topics in the economics of innovation and one of the least understood by the organisations themselves.
The retail sector serves as the central example of this issue. It is common for a retail chain to acquire AI-based demand forecasting tools, present them as an innovation milestone in internal reports, but continue to decide replenishment and campaigns exactly as before, with store managers trusting their own experience more than the model outputs. The tool exists, it was paid for, and it appears in the technology budget, but it did not change the decision-making process.
The article uses a historical parallel to illustrate the problem. When electricity arrived in American factories at the
