AI-Augmented Impact Frames: A Closed-Loop Architecture for Purpose-Aligned Decisions

This paper introduces a closed-loop decision architecture for aligning institutional purpose with operational decisions in AI-mediated environments. It presents the concept of impact frames as structured representations linking intent, evaluation criteria, and feedback signals. The paper shows how AI systems can both stabilise and transform interpretive structures, affecting how meaning is encoded and acted upon. It contributes a design-oriented perspective on governance architectures capable of maintaining alignment under conditions of complexity and technological mediation. This paper is part of the Coherence Programme. Further materials and companion papers are available at: https://thecoherenceprogramme.org/