Originally written in French. Translated by AI — the meaning has been preserved, not the prose.
Angle
It is not the tool that redefines the Product Manager's job, it is the price of its ancillary layer. As long as collecting, reformulating, documenting and maintaining absorbs most of the day, that production stands in for the definition of the role: you are assessed on it, recruited on it, you identify with it. When its cost falls, nothing about the product or the market has changed, and yet the layer stops taking up the room — which demonstrates after the fact that it was not the job, only its most expensive part. The shift that follows goes towards the work that layer made impossible: seeing where the market is heading, naming the right problem early, turning down seductive ideas. It is not free — the freed time has to be reallocated explicitly, and delegated execution comes back in the form of framing and supervision.
Synthesis
Taken separately, these observations look like a field report on an automated competitive intelligence setup. Put end to end, they describe an economic mechanism that does not concern intelligence alone.
The starting point is a silent arbitration. A subject whose entire cost sits upstream — watching videos, reading, translating, classifying, keeping a history — loses every day-to-day arbitration, whatever conviction you may have of its importance. This is not a prioritization failure: it is the sequencing of the cost that chooses, in the PM's place, what gets dealt with.
That mechanism, repeated across all ancillary tasks, produces an unexpected cumulative effect. The compensation layer — the one that makes up for what the organization does not circulate on its own — becomes voluminous enough to define the role. Cost acts as misleading proof: what is heavy looks important, and what fills the days ends up naming the position.
The collapse of the price turns the demonstration around. Nothing has changed about the product or the market, and what looked constitutive turns out to have been merely expensive. That is also why automation is experienced as an identity threat: it takes away what made competence visible.
Two reservations remain, and they prevent reading this shift as an automatic gain. The first is that delegated work does not disappear: it comes back as framing, as review, as supervision — a short, irregular load, badly estimated because it costs almost nothing right up until the day a false output has already served as the basis for a decision. The second is that freed time does not reallocate itself; the natural slope is to give it back to the schedule.
What the whole thing brings into view: the definition of a job follows the cost of its tasks, and when that cost moves, the evaluation criterion has to move with it — otherwise you go on counting tickets while the value has shifted elsewhere.
Tensions / contradictions
The main tension bears on what the collapse of a cost demonstrates. The angle concludes that the layer was not the job; one may object that a heavy task is sometimes heavy because it is decisive, and that automating it silently degrades the result. Nothing here supplies a criterion for settling that before trying.
A second tension, between two notes in this batch: the compensation layer is presented as incidental, whereas reformulating a need or following up at the right moment sometimes produces the understanding itself. The border between the plumbing and the job is therefore not as clean as the shift assumes.
Questions
- What signal would make it possible to establish that freed time has really been reallocated to understanding the market, rather than absorbed by new ancillary tasks?
- How does an organization evaluate a role whose main contribution no longer leaves any countable trace?