RHEA-DAPT: A transformative AI DSS for supporting adaptation pathways co-development

The increasing impacts of climate change require urgent, systemic and innovative responses to address growing risks to human and natural systems. In this scenario of complexity and uncertainty, the challenge is no longer merely to generate new data, but to transform existing knowledge into collective capacities to imagine, design and implement adaptation processes.

The central question guiding our research is: how can we co-create future adaptation pathways in a world where uncertainty has become the new normal?

To address this challenge, RHEA-DAPT has been developed, a Decision Support System (DSS) based on a Retrieval-Augmented Generation (RAG) architecture, conceived as a shared cognitive infrastructure for co-creating a knowledge base for transformative adaptation planning. Developed within the INTERREG AcquaGuard project, it supports climate change adaptation and resilience in flood-prone regions, including Karlovac County (Croatia) and the Veneto Region (Italy), case study regions in the project.

Methodological consistency is ensured through its alignment with the Regional Resilience Journey (RRJ) and the Regional Adaptation Support Tool (RAST), in line with the EU Mission on Adaptation. Grounded in these frameworks, RHEA-DAPT is built on principles of knowledge democratization, collective intelligence, and eXplainable AI (XAI) to enable transparent, interpretable, and collaborative decision-making.

Its multi-level architecture integrates diverse sources such as climate glossaries, regulatory frameworks, policies, territorial plans, project reports, and Nature-based Solutions (NbS) portfolios. The RAG approach reduces the need for dedicated LLM training, lowering computational costs and environmental footprints. By combining retrieval with generative models, it mitigates hallucinations and improves contextual relevance across regions.

Applied to AcquaGuard case studies and co-designed with their local actors, RHEA-DAPT demonstrates how the integration of scientific knowledge, policy and territorial expertise can generate inclusive and transformative adaptation pathways.

RHEA-DAPT embodies a new decision-making paradigm: not a prescriptive model, but a knowledge navigator that helps local actors navigate uncertainty, scenarios and possible alternatives. In this perspective, AI is not an autonomous decision-maker but a cognitive and relational facilitator, capable of supporting collective learning processes. The key question becomes not whether AI is intelligent, but how we can use it intelligently to foster new connections, stimulate critical thinking and strengthen communities capacity for co-creation.

In this uncertain future, even the idea of the future itself changes in nature: no longer a horizon of prediction, but a space of strategic foresight where envisioning what may come through scenario planning and analysis becomes the act that may transform our current choices.

In this perspective, RHEA-DAPT moves to an infinity loop, a dynamic reactivation of the adaptive cycle in climate change adaptation. Through iterative phases of reorganization, exploration, and transformation, adaptation becomes a continuous process of learning and renewal, enabling territories to achieve their climate  resilience while boosting innovative and transformative actions over time.