Governing complexity in sustainable innovation ecosystems: An empirical analysis for AI-enabled enterprise architecture

The integration of Artificial Intelligence (AI) into Enterprise Architecture (EA) offers transformative potential for advancing Sustainable Innovation Ecosystems (SIEs). However, AI's velocity often clashes with traditional, rigid governance. Using a sequential mixed-methods design – a systematic literature review followed by interviews with ten industry experts – this study maps AI's strategic impact across the EA-IE landscape. Our findings reveal that AI's value is primarily concentrated at the Business and Data layers, where it reconfigures ecosystem actors and processes. We uncover an EA paradox: while practitioners eschew formal frameworks as too rigid, they rely on architectural principles as dynamic capabilities to manage structural inertia. Theoretically, we advance from descriptive findings to causal mechanisms by proposing three testable propositions that explain how organizations can balance architectural stability with the agility required for AI adoption. Practically, the study offers a diagnostic tool for aligning AI adoption with organizational resilience. We conclude by outlining a roadmap for future research that utilizes dynamic multi-agent simulations to explore adaptive governance strategies for AI-enabled SIEs. This study contributes a localized, empirically grounded perspective on governing AI within complex adaptive systems, reframing EA as a flexible, human-centric coordination tool for innovation. • Explores AI integration in Enterprise Architecture for Sustainable Innovation. • Maps AI strategic focus to the Business and Data layers using expert heuristics. • Identifies an EA Paradox where principles are favored over rigid frameworks. • Proposes three testable propositions for governing AI in complex adaptive systems. • Provides a BDI-based multi-agent simulation roadmap for future ecosystem research.