Classification of artificial intelligence tooling in architecture. The information-algorithmic dimension

Abstract

The article presents an information–algorithmic axis of classification for
architectural AI tools and shows how AI-driven systems, platforms, and workflows can be
systematically organised by two parameters: the information dimension (types and sources of data
used and produced) and the algorithmic dimension (families of AI/algorithmic methods). On the
information side, the framework distinguishes geometric and spatial data, environmental and
performance data, programmatic and regulatory information, user- and sensor-derived streams, and
higher-level domain knowledge. On the algorithmic side, it spans rule-based and parametric
approaches, evolutionary optimisation, agent-based models, classical machine learning, deep neural
networks, generative models, reinforcement learning, and large foundation models.
Using a matrix of I–A pairings, the article maps a set of contemporary tools and research
prototypes in architecture and urbanism, covering generative design platforms, performanceoriented
surrogates, code-compliance assistants, immersive evaluation workflows, structural design
agents, and city-scale GeoAI systems. The analysis reveals dense clusters such as geometry-driven
generative design and performance prediction, as well as under-explored combinations, including
context-aware reinforcement learning and programme-to-geometry generative pipelines. The
discussion highlights typical trade-offs between data requirements, algorithmic complexity,
interpretability, and the role of human oversight.
The information–algorithmic axis complements functional and lifecycle classifications and
provides a practical lens for understanding data needs and computational behaviour of architectural
AI tools. It can inform the selection and chaining of tools in digital workflows, the design of
curricula that prepare architects to work with data and algorithms, and the identification of
promising gaps for future research and development.