Category: [:fr]Veille & tendances[:en]Insights & trends[:it]Notizie e tendenze[:es]Novedades y tendencias[:]

France Travail's 2026 labour needs survey contains two figures for construction that appear to contradict each other. The sector reports around 143,000 planned hires, down 16.4% year on year. At the same time, 65% of those projects are rated hard to fill, a record level. Fewer hires planned, more difficulty making them. That apparent paradox says something precise about the nature of the problem, and about where to look for room to manoeuvre. What the figures say, and what they do not First, the gap in perspective. Across the French economy, the 2026 survey counts nearly 2.3 million hiring projects, down 6.5%, with 43.8% rated difficult against 50.1% the previous year. Tension is easing nearly everywhere, except where you are: construction sits 21 points above the national average. Two reading cautions, which matter if you are building a decision on this. The survey measures declared intentions, not completed hires, and France..

Over the past two years, "AI" has found its way into nearly every planning software brochure, ours included. The trouble is that the word covers very different things: a genuine optimisation engine, a rule-based sort given a new name, or nothing at all. Let us sort through it. What automation genuinely does well today Four uses are mature, in the sense that they rest on proven techniques and produce verifiable results. Constrained allocation. This is the heart of it. Assigning people to tasks while simultaneously respecting required skills, availability, rest periods and hour caps is a combinatorial optimisation problem that machines handle better than we do. With a team of ten, an experienced planner does fine. With sixty, four overlapping qualifications and rest constraints, nobody explores that solution space in their head. The orders of magnitude explain why. Assigning thirty people to thirty slots already means more than 1032..