AI at VINCI: from field deployment to scaling up

The approach

AI at VINCI: a pragmatic approach focused on business needs and operational pain points

At VINCI, the deployment of artificial intelligence follows a pragmatic approach, focused on business needs and operational pain points. Leonard, the group’s innovation and foresight platform, helped initiate this momentum and continues to explore, structure, and scale high-potential use cases across the organization, in coordination with initiatives from VINCI’s various business units.

Bruno Daunay, Director of Leonard’s Artificial Intelligence Program, leads these initiatives, which prioritize measurable gains and scalability, with the support of partners such as Wavestone.

The interview

From identifying use cases to demonstrating value

What does Leonard do in practice, and what role does AI play?

Leonard was created in 2017 to help the Group, which is highly decentralized, coordinate on strategic topics for the future (autonomous mobility, AI, energy, climate, etc.). The platform is structured around foresight, intrapreneurship, and open innovation.

AI was initially a monitoring topic as early as 2019, before becoming a recognized performance lever for certain use cases. A dedicated program was therefore structured to enable entities to develop high-value applications and build capabilities. The approach remains pragmatic: AI is a means, not an end, and is only adopted if it generates measurable performance.

How do you identify priority projects? Are gains achieved quickly?

Topics originate from existing processes and operational pain points. The objective is to assess the benefits of an augmented solution: time savings, quality improvement, risk reduction (particularly in terms of safety), or the creation of new products.

Benefits can be quantified relatively early, but profitability takes time: development, deployment, adoption, and scaling. For a new product, it typically takes around three years to reach break-even.

Scaling up

Structuring use cases around shared needs and measurable gains

How do you manage scaling, and what are the challenges?

At VINCI, a use case does not automatically become a Group standard: it may remain local if it does not address a shared need. When scaling potential exists, the solution is structured as an offering, with a strong focus on demonstrating value to the entities, which remain the decision-makers.

The main challenge is organizational: each project must convince its internal “clients,” which makes industrialization demanding but also limits unnecessary investments.

Another challenge concerns support functions, where similar developments may exist across multiple countries. Alignment is not always straightforward, as contexts and priorities differ, making it difficult to demonstrate a global benefit.

How do you build AI awareness across teams, and what skills are required?

Awareness-building initially relies on educational initiatives to explain the possibilities and limitations of AI without overestimating it. It is then extended through operational workshops with business teams to identify concrete use cases and decide on potential investments.

Key skills rely on strong business expertise to identify the right problems, complemented by technical profiles capable of developing solutions and demonstrating their economic value.

Looking ahead

Keeping AI focused on clearly defined needs

What are the main challenges ahead, and your vision of AI’s role in companies?

The major challenge will be to rigorously assess the return on investment, particularly in support functions, to avoid disproportionate spending compared to actual gains.

In the long term, AI will be integrated into many standard tools. Companies will need to decide between what they develop internally and what they purchase on the market.

Finally, a common misconception remains: that AI does everything. In reality, it remains a tool that depends on clearly defined needs and on teams’ ability to ask the right questions.