Inteligência artificial para gestão de desempenho de motoristas de empresas do transporte rodoviário de cargas

Abstract

Road freight transport plays a structural role in the Brazilian economy, frequently operating under conditions of high complexity and risk, especially in the niche of thermolabile transport of pharmaceuticals and high-value-added inputs. Given the accelerated growth of fleets and the exhaustion of current monitoring systems—predominantly analog, static, reactive, and punitive in nature—this Applied Project aims to propose an advanced model for driver behavior management, structured under the concept of "Hybrid Allocation" and supported by Explainable Artificial Intelligence (XAI). The research sought to answer how the transition from this retrospective surveillance paradigm to a predictive algorithmic model can guarantee operational efficiency, cold chain security, and the continuous engagement of employees in the focus company of the study: EMF Logística. Methodologically, the work is characterized as a descriptive and exploratory case study, of an in-depth qualitative nature. Primary data collection occurred through semi-structured interviews with strategic and operational leaders of the organization, and the findings were contrasted through a multiple benchmarking process with reference companies in the mobility and transport sector (Expresso Nepomuceno, Cidade BH, and Localiza), which present different levels of maturity in the treatment of the human and technological factors. The diagnosis evidenced a mismatch between the complexity required by the operation and the fragility of the current behavioral control. As an intervention, the project developed the complete planning for the implementation of a predictive system that cross-references, in real-time, variables of the driver's profile, load specificities, and the topographic severity of the route. This architecture makes it possible to isolate unavoidable mechanical wear and tear from the professional's actual driving performance, replacing the culture of punishment with an educational ecosystem based on gamification and reward. It is concluded that the adoption of Artificial Intelligence in the Hybrid Allocation model is technically and financially viable, configuring itself as an imperative cultural transformation that mitigates labor liabilities, reduces accident rates, and elevates the dignity and efficiency of the road driver.