Beyond the code: when AI is inclusive by design

At Crédit Agricole Italia, artificial intelligence is approached with a clear principle: innovation only has value if it is pursued responsibly and in service of people. In line with the procedural regulations of the French Group and the European Regulation on the matter, the Artificial Intelligence Act (2024), transposed in Italy in 2025, the Group has adopted an AI Policy aimed at defining the principles for the adoption and use of artificial intelligence and describing the roles played by each function. To ensure ethical and responsible AI development, the Group has implemented a technical control system throughout the entire AI software development lifecycle, providing precise oversight during the entire process, with checks on data, model selection, and model output.

The policy outlines seven fundamental requirements for developing reliable AI, including the principles of diversity, non-discrimination, and fairness. Every new solution is carefully evaluated to ensure data representativeness, accessibility, and stakeholder engagement. For this reason, the HR function, and specifically the Diversity & Inclusion team, is involved from the early stages in the design of AI-based solutions: an important acknowledgment of HR’s role and, at the same time, a significant responsibility, necessary to ensure that inclusion is a design requirement rather than an afterthought. The goal is to ensure that inclusion and diversity are considered throughout the system’s lifecycle, avoiding distortions and bias—for example, by guaranteeing data representativeness, accessibility, and consultation with informed stakeholders. This D&I-driven approach can create value in AI development, as D&I is not only about risk prevention but also about seizing market opportunities. More equitable and inclusive AI opens doors to new customer segments because, by considering new metrics and reducing bias, it can include people who would otherwise be excluded from the system. For this reason, a D&I-driven approach is not only an ethical matter but also a strategic one for companies: more inclusive AI leads to more robust and generalizable models, capable of addressing the needs and demands of as many people as possible.

Within the Le Village by Crédit Agricole ecosystem, this is a reality where solutions demonstrate how AI, inclusion, and welfare can translate into concrete impact. Le Village by CA is an open and dynamic innovation ecosystem where startups and companies collaborate to grow the businesses of the future and support companies in their innovation journey. In Italy, Le Village by CA consists of a network of five hubs—Milan, Parma, Padua, Sondrio, and Catania—capable of accelerating over 190 startups and supporting more than 80 partner companies in their innovation journey, promoting crowdfunding, charity, and corporate volunteering activities. Within the ecosystem, several startups stand out for implementing AI solutions aimed at making care systems more equitable, accessible, and capable of responding to the complexity of real lives, driving innovation in health and well-being.

Geen was created as a femtech solution in reproductive health, with a gender approach and strong attention to real needs and individual characteristics. In the initial phase, the project integrates an AI-based solution to support orientation and matchmaking toward professionals and care pathways. It soon becomes evident, however, that orientation difficulties affect care systems more broadly, revealing wider potential. Services exist, but they are fragmented, distributed among different actors, and difficult to navigate, both for individuals and the organizations providing them. Geen evolves as an AI triage layer applicable to various contexts, supporting companies, local authorities, and healthcare facilities to guide people toward the most appropriate services and care pathways, enhancing the existing offerings. AI enables the interpretation of natural language, consideration of individual and contextual characteristics, and improvement of access appropriateness, reducing bounce, dispersion, and inefficient use of resources.

Longeviva, on the other hand, was born from a common experience many people have had: giving up a gym membership despite knowing it would be beneficial, because starting is easy, but maintaining consistency over time is much harder. According to Longeviva, the solution does not lie in willpower but in the absence of systems capable of supporting real-life behavioral change. The true challenge in prevention and lifestyle management is not knowing what to do but being able to do it when daily life takes over. From this insight, Longeviva was developed as a technological infrastructure designed to fill a structural gap in the healthcare system: professional knowledge exists, but it is tied to isolated moments and traditional one-to-one interactions. Professionals have limited time, while behavioral change requires continuity, adaptation, and presence in daily life. Longeviva connects expertise, methodology, and real life in a scalable way: AI does not replace the professional or automate decisions but acts as an enabling infrastructure. In this way, human knowledge can be distributed, maintained coherently, and made accessible on a large scale, generating tangible benefits for people and greater impact. Artificial intelligence can either amplify inequalities or become a lever for inclusion, care, and well-being. The difference lies in the chosen direction: designing responsibly and innovating with a human-centered approach can help build fairer organizations and a more equitable society, where technology does not replace people but supports them.