(Note to Readers: This article applies the excellent analyses of researcher Philippe Silberzahn to our vehicle rental and Revenue Management sector. The historical and theoretical examples are drawn from his texts, while the transposition to the WeYield ecosystem is an editorial proposition added for our community.)
General Managers and Revenue Managers, we are confronted daily with the emergence of new tools: dynamic pricing algorithms, artificial intelligence, and automated fleet management. We might believe that the integration of these technologies will automatically multiply our profitability, but the reality is quite different.
Technology as an Accelerator: The Imperative of Human Investment
The dramatic advances in artificial intelligence (AI) lead to hopes for massive productivity gains, but there is no direct and linear relationship between the use of a technology and the increase in performance. Worse still, a poorly integrated technology can harm the productivity of your agency or department.
To understand this phenomenon, we must look at the Lowell textile mills in the 19th century: by entrusting their workers with a third mechanical loom, managers thought they would increase production by 50%, but it was a failure. Why? Because the worker's true job was no longer weaving, but monitoring the machine and checking quality. By adding machines, managers created a bottleneck that was not technological, but human, because the volume of verifications exceeded what one worker could manage.
This same dynamic is reproducing today with AI, on a much vaster scale. The cost of production (pricing analyses, report generation, etc.) is collapsing, but the cost of verifying this data is drastically increasing because it still relies on human judgment. Faced with the risks of AI errors (or "hallucinations"), the temptation would be to use another AI to verify the first, but this is a trap: the models share the same blind spots and repeat the same errors with confidence.
Delegating the verification of one artificial intelligence to another AI is a very risky approach according to the sources.
The main identified risks are as follows:
- Shared Blind Spots: Since AI systems are generally trained on similar data, they possess the same blind spots and commit the same errors "with full confidence."
- A Lack of Independence and an Illusion of Reliability: The control of one model by another offers no objective supervision. Instead, it creates a "game of mirrors" where the AIs give the impression of agreeing with each other, which generates a false certainty while repeating the same errors.
The sources use a very clear metaphor to illustrate the absurdity of this approach: it amounts to "having the bar watched by its customers."
That is why, although the cost of AI-generated production decreases, the cost of its verification increases: this essential control must continue to rely on the judgment of humans specifically trained for this complex task.
What about Revenue Management in Car Rental?
The lesson for our Revenue Management departments is clear: innovation does not consist of replacing humans with machines, but of inventing the best ways to combine the two. To achieve optimal productivity, the Lowell mills had to accept a temporary drop in output and, above all, triple their investment in the training of each worker. Tomorrow, a Revenue Manager equipped with AI will produce infinitely more, on the strict condition that the company invests heavily in their training so that they develop the expertise and critical sense necessary for supervising these new systems.
Ultimately, does "Disruptive Technology" really exist?
In the race for competitiveness, we are all looking for the "disruptive technology" that will give us a decisive advantage over our competitors. However, disruptive technology, in itself, does not exist.
What truly creates a disruption in a market is not the novelty of a tool, but the way it is used and the model that is built around it. If you integrate a powerful AI algorithm but maintain the same internal organization, the same economic model, and the same hierarchy, you will only achieve a "sustaining" improvement: you will do the same thing, just a little more efficiently. Disruptive innovation can also exist without any new technology, like the low-cost airlines that revolutionized their sector with the same planes, but a completely redesigned business model.
Historically, when faced with a new technology, the greatest mistake managers make is the "stuffing temptation": forcing the novelty to fit into the existing organizational model. When the tank was invented, the French army scattered it to support its traditional infantry, underutilizing its potential. The Germans, on the contrary, rethought their entire tactic around the tank, transforming this invention into true disruptive innovation.
Conclusion
To disrupt the car rental market, adopting a new algorithmic tool will not be enough. The real innovation lies in your capacity, as leaders, to adopt an entrepreneurial stance to challenge your existing models. Build new Revenue Management processes around the technology, intensively train your teams to master it, and that is how you will create a decisive competitive advantage. At WeYield, we approach performance improvement by relying on technological solutions supported by powerful AI and coaching analysts to increase the quality of their analyses and decisions. To learn more, contact us.
Source
https://philippesilberzahn.com/2026/03/23/ia-et-productivite-la-grande-lecon-de-l-industrie-textile/
https://philippesilberzahn.com/2023/12/04/une-technologie-de-rupture-ca-n-existe-pas/
Photo credit
.jpg)
.jpg)