Advancing expert modeling through richer data integration and next-generation behavioral simulation
Building on our current AI-augmented modeling capabilities (see Our Approach), our future work centers on behavioral simulation, data integration, and complex systems analysis—advancing how we represent decisions, networks, and markets in policy models.
The emergence of large language models creates unprecedented opportunities for behavioral simulation, but also requires rigorous validation to ensure outputs remain grounded in empirical data and sound theory. Our advantage lies in combining these capabilities with 20+ years of domain expertise to deliver both innovation and trustworthiness.
We focus on three strategic directions: (1) integrating AI-powered generative agents into agent-based models for richer behavioral realism, (2) synthesizing large-scale data through AI-assisted methods, and (3) extending proven modeling platforms into new domains and crisis scenarios.
These projects are being prepared for presentation to federal, state, and local funding programs, including SBIR/STTR, NSF, and foundation opportunities. We also welcome inquiries from funders interested in related or complementary efforts aligned with the tools and methods presented here.
An integrated compartmental modeling framework that runs orders of magnitude faster than traditional microsimulation models, while retaining rich system-level detail. HERMES simultaneously tracks coverage decisions, hospital finances, disease progression, and healthcare costs through explicit feedback loops to predict how Medicaid cuts cascade through healthcare systems long before systemic failure occurs.
Learn More →A survey-informed AI platform that creates synthetic respondent personas from longitudinal survey data to emulate individual and population responses across diverse contexts. By training language models on real survey responses, COMPASS enables rapid exploration of counterfactual questions and complex policy or social scenarios—providing timely behavioral insights when traditional data collection cannot keep pace.
Learn More →
A data-driven platform that generates privacy-preserving synthetic populations by integrating multiple overlapping datasets. It produces individual-level populations at flexible scales, preserving key statistical relationships and eliminating re-identification risk. The framework fuses individual attributes with social networks, combining egocentric survey data with large-scale network information on population structure, connectivity, and clustering.
Learn More →A platform that brings together multiple surveys, often collected separately, to uncover patterns and relationships across different topics. It analyzes responses, question wording, and context to reveal connections—such as how views on health, immigration, or politics may relate—and suggests new survey questions to help behavioral scientists explore these links further.
Learn More →
This framework models the supply, demand, and wages of healthcare and K–12 education workforces, incorporating behavioral factors, training pipelines, and career transitions. It can simulate how acute stressors—such as pandemics, work fatigue, or mental strain—affect workforce participation, retention, and wage expectations, providing evidence-based tools to anticipate and mitigate shortages before they become critical.
Learn More →We explore behavioral responses in tax compliance, financial decision-making, and regulatory environments using agent-based, network, and AI-driven methods. This includes studying how social influence, enforcement, adaptive expectations, and systemic feedback shape individual and population-level responses under changing economic and policy conditions.
Analyzing how individuals and teams interact within organizations, mapping collaboration networks, communication patterns, and workflow dynamics. We also study inter-organizational interactions, supply chains, and the propagation of information or behaviors across corporate networks, using data-driven simulations to uncover actionable insights.
Investigating how crises, systemic shocks, or rapid policy changes reshape incentives, risk perceptions, and behavioral norms across social and economic domains. Our methods integrate multi-level modeling, network analysis, and AI simulations to understand emergent phenomena and anticipate system responses under uncertainty.
Extending AI and simulation approaches to biomedical domains, including clinical trial analysis, longitudinal health data, and genetic or molecular pathways. We simulate and generate synthetic populations, explore counterfactual scenarios, and apply AI-driven models to inform research, policy, and decision-making in complex biological systems.
While these examples illustrate current focus areas, COMPASS is designed to be flexible and broadly applicable. We welcome diverse projects across simulation modeling, AI-driven behavioral and population analysis, complex systems, financial networks, organizational studies, and health research—exploring any domain where rapid, data-informed scenario analysis can provide actionable insights.
We continue to partner with public agencies, foundations, and private organizations to apply these methods to high-stakes policy challenges. To explore collaborations or discuss how these capabilities can support your mission, please contact us.
