CLASSIFICATION OF ARTIFICIAL INTELLIGENCE TOOLING IN ARCHITECTURE. PROCESS-CYCLE DIMENSION

Read full text Summarise with VIRAT
Authors: Наталія Сергіївна Вергунова
Cited by: 0
References: 12

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The article presents a process–cycle axis of classification for architectural artificial intelligence (AI) tooling and shows how AI-driven systems, platforms, and software can be systematically organized by the parameters «Process Stage» and «ML Lifecycle». «Process Stage» aligns tools with the Royal Institute of British Architects (RIBA) Plan of Work 2020 stages (Strategic definition; preparation and briefing; concept design; spatial coordination; technical design; manufacturing and construction; handover; use), clarifying where a tool brings the highest value and which phase-typical deliverables it must connect to. «ML Lifecycle» captures which machine learning (ML) steps are practically active, where ML is understood as a data-driven AI approach that learns patterns from examples to generate predictions or outputs without explicitly programmed rules for each case. The lifecycle steps include data collection, data preparation, model development, model deployment, and monitoring and feedback. Mapping these parameters in a classification matrix and relating them to ISO 19650 information-management artefacts provides a standards-aware view of AI use across the built-environment workflow, making it easier to connect AI outputs to controlled deliverables and managed information exchanges. The proposed classification covers a heterogeneous set of tools and allows each example to be positioned by both its workflow placement and its ML work profile, distinguishing inference-first products from solutions that require continuous data loops and monitoring. The analysis indicates that early stages (0–1) are dominated by advisory analytics that strengthen feasibility and briefing, while design stages (2–4) are typically characterized by deployment of provider-trained models embedded into design environments, with project-specific work concentrated in calibration, checking, and governance. Construction (stage 5) and operation/use (stage 7) are associated with continuous monitoring and feedback cycles, near-real-time data capture, and iterative adjustment, whereas handover (stage 6) emerges as a distinct AI/ML niche linked to «as-built» generation and information completeness for transition to operation. Overall, the process–cycle axis supports clearer decisions about integration points, roles, and assurance expectations, and it highlights where lifecycle tasks such as monitoring and drift response are often under-specified in procurement and execution planning. By coupling tool selection to ISO 19650 deliverables, AI can be treated not as an add-on, but as a governed component of lifecycle information management.

Fields of Study & Subjects

BIM and Construction Integration Architecture and Computational Design Artificial Intelligence Applications

Publication details

Journal / source
Municipal economy of cities
Pages
125–125
Publication date
23 March 2026
Language
EN