AI Decision Support

CLARA: County-Level AI for Resilience and Assistance

County-Level AI for Resilience and Assistance is evolving from an accessible research assistant into a scalable, agentic decision-support platform for data, models, simulations, and stakeholder planning.

CLARA: County-Level AI for Resilience and Assistance project image

Why CLARA

CLARA was created to reduce the distance between advanced infrastructure resilience research and the people who need to use it. County stakeholders, utilities, emergency managers, educators, and researchers often need to combine reports, geospatial layers, infrastructure models, hazard data, graph representations, simulations, and resilience metrics. CLARA turns those disconnected research products into an interface that can retrieve, explain, orchestrate tools, and support scenario-driven questions.

Evolution

The project has progressed in stages. CLARA established the proof of concept: a human-centered LLM/RAG assistant for ARISE research products and county-level infrastructure questions. CLARA 2.0, now under development, expands the architecture from stored models toward automated county-scale model generation, national GIS data integration, multi-domain co-simulation, and LLM-based tool orchestration. CLARA 3.0 is proposed as a future extension that would connect CLARA-style agentic interfaces with climate-informed hazard prediction, uncertainty-aware optimization, economic consequence analysis, and resilience planning workflows.

The long-term vision is not a chatbot sitting on top of documents. CLARA is an orchestration layer that connects evidence, models, simulations, and decisions.

My Role

I co-created CLARA and lead the technical development across agentic AI workflows, data and GIS integration, synthetic infrastructure model generation, HFG-TK coupling, CISCS co-simulation integration, resilience metric interpretation, and stakeholder-facing decision workflows.

CLARA Evolution

Three stages of the same research platform, moving from achieved proof of concept to current development and proposed future capabilities.

CLARA concept image
Achieved; public release planned soon

CLARA

The first version established CLARA as a county-level AI assistant for making ARISE research outputs more accessible. It integrated documents, datasets, infrastructure modeling concepts, HFG-based analysis, weather context, resilience metrics, and co-simulation outputs through a conversational interface.

  • LLM/RAG interface for querying ARISE research products and county-level resilience questions.
  • Proof-of-concept support for document-based retrieval and what-if scenario evaluation.
  • Designed to lower technical barriers for county stakeholders, students, educators, and researchers.
CLARA 2.0 concept image
Under development

CLARA 2.0

CLARA 2.0 moves from a static proof of concept toward a scalable decision-support framework. It uses national GIS datasets and synthetic infrastructure model generators to dynamically build county-scale multi-domain models, connect them with HFG-TK, and orchestrate co-simulation workflows through natural language.

  • Automated model generation for infrastructure domains when detailed local models are unavailable.
  • GIS-enabled pipelines using national datasets, asset layers, weather and hazard inputs, and infrastructure generators.
  • LLM-based orchestration of HFG construction, multi-domain simulation, diagnostics, resilience metrics, and stakeholder queries.
CLARA 3.0 concept image
Proposed; awaiting funding

CLARA 3.0

CLARA 3.0 is the proposed next step within the broader PRAIRIE vision. The CLARA-specific portion would extend the platform toward climate- and hazard-informed agentic decision intelligence, connecting Earth-system prediction, infrastructure consequence modeling, uncertainty-aware optimization, scenario reduction, and stakeholder-facing interfaces.

  • Agentic decision-support workflows for long-term resilience planning and short-term operational response.
  • Integration pathway for probabilistic hazard information, cascading infrastructure impacts, and socio-economic consequence metrics.
  • Future-facing interface for utilities, emergency managers, infrastructure operators, planners, and community stakeholders.
Previous ProjectHFG-TK
Next ProjectCISCS