From Tools to Partners: Validating Human-AI Collaborative Workflows in Architectural Design

This contribution aims to present a systematic literature review that
synthesises six prominent Human-AI collaboration models to validate a phasespanning “Digital Partner” workflow for architectural practice aligned with the staged logic of the RIBA Plan of Work (selected as reference of the design process structure). Through a comparative matrix evaluated against six interrelated criteria, specifically, 1. role of AI, 2. phase integration, 3. interaction style, 4. feedback mechanisms, 5. scalability, and 6. ethical control, this initial study confirms convergences in mixed-initiative co-refinement and preservation of human agency, while identifying critical gaps in cross-phase continuity, accountability, and operational governance. To address these gaps, the paper translates the matrix into a conservative but implementable workflow spanning briefing, concept, development, and delivery, with explicit AI roles, human authority checkpoints, and auditable handoff artefacts. The contribution is positioned as Phase 1 of a research programme that prepares follow-up empirical validation in professional practice and advanced design education. The research demonstrates that human- AI collaboration represents a fundamental re-conceptualisation of design authorship
and professional responsibility in the perspective of future more intense interactions and teamwork.