Continuous authentication (CA) provides ongoing monitoring of user identity to mitigate the risk of unauthorized access to corporate assets. In this work, a new framework that automates building and maintaining adaptive CA systems intended for diverse teleworking environments is presented. In the proposed approach, a number of CA methods from four main categories are employed, namely: biometric-physiological, biometric-behavioral, interactive, and contextual. The proposed framework considers various decision factors, including user convenience, privacy, data and system security, authentication efficiency, infrastructure availability, and business requirements. It utilizes client requirements as well as empirical data derived from both laboratory experiments and log files of real-world CA systems in order to calculate the most efficient combinations of CA methods represented as sets of decision rules, that are customized for specific use case. The modular architecture and integration capabilities of this framework provide a robust foundation for ongoing research and development in the domain of telework security solutions.