Community health agents play a crucial role in the prevention of chronic noncommunicable diseases (NCDs), with growing global prevalence in children. This situation requires effective preventive strategies in the public health system. This project thus proposes interaction by community health agents with machine learning models and a large language model (LLM) for the identification of risks and preventive actions in the first 1,000 days of life. The LLM tool is intended for use in home visits in primary care. The idea is to provide the workers with an advanced, user-friendly tool using artificial intelligence associated with the best scientific evidence to identify pregnant women and infants who are susceptible to risk conditions, supporting the communication of results and solidly based recommendations. The expectation is that the project will empower the community health agents’ work and expand the impact of preventive health actions, representing an innovative approach in the context of comprehensive healthcare in early childhood.