As digitalization reshapes all sectors, intelligent tools increasingly transform architectural design practices. Artificial Intelligence Systems (AIS) extend human cognition by processing large-scale data, performing complex calculations, and generating predictions. Yet, their adoption remains limited due to a lack of understanding, perceived complexity, and insufficient training. This research introduces an adaptive periodic table to classify intelligent tools relevant to architectural design. Inspired by Mendeleev’s model, it organizes learning algorithms according to three criteria: their computational complexity (eight periods, from constant to factorial), their learning methods (six families defined by supervised, unsupervised, or reinforcement learning, shallow or deep), and their role within AI processes (four blocks: analysis, model development, decision-making, causal inference). Methodologically, the table is constructed through a structured procedure that combines a transdisciplinary review of AI techniques, the definition of unified classification criteria, and the systematic mapping of algorithms that are already deployed or deployable in architectural workflows. Each algorithm is further specified through labels indicating types of inputs/outputs, accuracy, training time, and ability to capture linear or nonlinear relations. By clarifying algorithmic functions and applications, the model provides a framework to compare and situate intelligent tools in design contexts. It fosters a better understanding of their interoperability and potential synergies, emphasizing that AI complements rather than replaces human cognition. This periodic table thus fulfils both pedagogical and professional functions: it supports architects and students in the use of AI by providing an intuitive understanding of its functionalities and integration, while serving as an operational reference that enables practitioners to select, combine, and sequence AI algorithms throughout the design process.