Verso strumenti decisionali per l’equilibrio vita‑lavoro dei lavoratori maturi: un modello integrato basato su reti neurali artificiali

In recent decades, population ageing has become a defining shift across Europe, forcing a rethink of welfare and pension systems at both national and EU levels. While data unequivocally document the trend, public perception lags. Many resist its practical consequence – the postponement of retirement – which keeps people in the labour market longer. Ageing also intersects with health and care dynamics. Longer life does not always mean healthier later life; functional decline increases vulnerability and care needs. At the same time, older people face rising family responsibilities, from grandchild care to support for very old parents, intensifying their social and domestic workload. These pressures make it crucial to improve older workers’ work-life balance (WLB) and to understand how WLB shapes retirement timing. Yet research often treats these domains separately and seldom yields actionable tools. We propose an integrated evidence-to-design approach that links WLB determinants with retirement push/pull dynamics among Italian seniors. Using an original 2024-2025 survey of Italian workers aged 55+, measuring working conditions, WLB, retirement intentions, and bonding/bridging social capital, we: (i) quantify the determinants of WLB and the push/pull factors influencing retirement; (ii) develop a pilot AI module that profiles WLB-dissatisfaction classes and ranks class-specific drivers; and (iii) outline a web-based tool that translates evidence into class-specific policy portfolios. Findings indicate that policy should shift from one-size-fits-all to targeted bundles. The proposed tool helps managers and policymakers both raise older workers’ satisfaction and allocate resources efficiently. Next steps include embedding retirement push/pull directly into a multi-task ANN, expanding stratification by industry and region, and performing external and temporal validation.