The increasing frequency and severity of cyberattacks demand robust security solutions, especially for mission-critical environments. A key challenge is the dynamic configuration of secure services in Cloud Continuum for ad-hoc mesh networks, which traditional methods struggle to address. In this work, we propose ZETIn, a Zero Trust Infrastructure that leverages the power of Generative AI and a chatbot with LLMs to automatically configure services based on natural language inputs. The proposal ensures end-to-end security, simplifies the configuration process, reduces human error, and enhances network resilience and adaptability. Preliminary results demonstrate effective dynamic service configuration, high security standards, and improved network resilience, demonstrating that the integration of Generative AI with Zero Trust principles is a significant step forward in enhancing security and efficiency in mission-critical ad-hoc networks.