AI: Creating the Conditions to Scale

Mission

For a major French civil engineering division, part of a leading European construction and concessions group, Artificial Intelligence had become a strategic question within its 2026–2030 roadmap.

With close to 30,000 employees and more than €6bn in annual revenues, the division delivers complex infrastructure projects across transport, energy and maritime sectors, from ports and tunnels to major engineering structures. Its performance is built on technical rigour, reliable processes and disciplined execution across a decentralised, project-driven organisation.

In such an environment, AI introduced a distinctive tension. Its potential lay in the ability of teams to identify useful applications close to the reality of projects, engineering methods and day-to-day operational challenges. Yet any emerging use of AI had to meet the standards of a business where reliability, confidentiality, system integration and alignment with Group principles are essential.

The Executive Committee therefore wanted AI to contribute to operational excellence and project productivity. Not as a disconnected innovation topic, but as a practical lever for the business. The challenge was to create the conditions for relevant uses to emerge from the field, while providing enough structure to discuss, prioritize and govern them coherently.

The mandate was to define a common framework for AI opportunities within civil engineering activities, connect local exploration with Group-level coordination, and establish the foundations for secure and scalable development.

Solution

A senior European working group was mobilised to define the principles and governance required to strike a demanding balance: allowing useful AI applications to emerge from the field, while providing the structure needed to develop them safely, coherently and at scale.

Bringing together business, support functions and AI change expertise, the group began by establishing the role AI could play in the division’s operational excellence agenda. Rather than treating it as a separate innovation or IT topic, the initiative positioned AI as a practical transformation lever: one capable of improving productivity, supporting project teams, accelerating access to knowledge and strengthening decision-making.

Working sessions with operational and functional stakeholders then explored where concrete AI uses could emerge from the reality of the business: situations where teams lose time, where knowledge is difficult to retrieve, where information remains fragmented, and where AI could bring practical value. They also drew an essential line between the activities AI could support and the professional judgement that must remain with engineers, project managers and functional teams.

In parallel, the initiative defined the conditions for governed and scalable development: principles for data usage and confidentiality, integration with existing information systems, consistency with Group guidelines and connection with Group AI coordination.

This was a change challenge, but not a conventional one. With AI, the destination cannot be fully prescribed at the outset. Uses emerge progressively, technology evolves rapidly, and relevance becomes clearer as teams test, challenge, learn and adapt. The work therefore focused not only on where AI could be applied, but on how the organisation could continue to learn as promising uses matured.

The initiative created a bridge between bottom-up exploration and Group-level steering: enough openness to capture useful applications from the field, enough structure to guide their development safely and consistently.

The work helped give emerging AI opportunities a governed and pragmatic framework for development, connected to operational reality and designed to learn as it scales.