Policies affect millions of people.
Test them before you roll them out.

Policies affect millions of people. Test them before you roll them out.

Prior Foundry helps policy teams research evidence, simulate impacts on realistic populations, and draft decision-ready documents - with full transparency and auditability.

Book a demo

Partnering with policy institutions, research organizations, and government agencies

From evidence to action in three steps

Turn complex policy questions into tested, documented decisions—faster and more rigorously than traditional approaches.

01

Scope

Map the decision space: define the options, review the evidence, and identify key constraints shaping the problem.

02

Simulate & Validate

Test options on synthetic populations from large-scale survey data. Compare outcomes across demographics and validate with real humans.

03

Decide

Optimize across trade-offs like efficacy, equity, cost, and feasibility to generate decision-ready outputs grounded in evidence and simulation.

Questions we help answer every day

From evidence to action in three steps

From government agencies to corporate policy teams, organizations use Prior Foundry to test policy impacts before implementation—across dozens of sectors and domains.

Political Campaign

"Across three versions of our candidate's economic message, which produces the highest persuasion lift among independents, and which alienates our base?"

Legislative Office

"If this childcare bill passes as written, how will take-up differ across rural, suburban, and urban households — and what design changes would close that gap?"

Academic Researcher

"Before we run this field experiment, what is the predicted treatment effect across subgroups?

Insurance Company

"Which incentives are most effective at encouraging homeowners we insure to invest in wildfire mitigation measures?"

Public Sector Consultant

"We're recommending this benefits enrollment redesign on cost grounds — but before we finalize the report, what are the behavioral effects we haven't accounted for?"

Public Health Department

"Which outreach strategy produces the highest cancer screening uptake among residents who distrust the healthcare system?"

Political Campaign

"Across three versions of our candidate's economic message, which produces the highest persuasion lift among independents, and which alienates our base?"

Legislative Office

"If this childcare bill passes as written, how will take-up differ across rural, suburban, and urban households — and what design changes would close that gap?"

Academic Researcher

"Before we run this field experiment, what is the predicted treatment effect across subgroups?

Insurance Company

"Which incentives are most effective at encouraging homeowners we insure to invest in wildfire mitigation measures?"

Public Sector Consultant

"We're recommending this benefits enrollment redesign on cost grounds — but before we finalize the report, what are the behavioral effects we haven't accounted for?"

Public Health Department

"Which outreach strategy produces the highest cancer screening uptake among residents who distrust the healthcare system?"

Drag around

Built entirely on
survey-based agents

From evidence to action in three steps

Agent profiles are constructed from high-quality probability surveys, ensuring representative modeling across diverse populations and contexts.

Representative populations

Built from General Social Survey, World Values Survey, European Social Survey, and other validated sources

Custom integration available

Integrate your organization's survey data for tailored agent profiles and scenario analysis

Cross-cultural accuracy

Survey-based agents outperform demographic-only approaches, especially in diverse contexts

Peer-reviewed methods you can trust

Our simulation approach draws on published research in computational behavioral science, with robustness testing, permutation tests, and sensitivity analysis built in.

Built by researchers who set the standard

Our team combines academic leadership in computational behavioral science with real-world policy experience at Stanford HAI, Sciences Po, the OECD, and Cambridge. We don't just use AI methods - we publish them.

Featured Research

Read the full version

Designed for sensitive policy contexts

Policy work demands rigour, transparency, and control. Every part of Prior Foundry is designed with that in mind.

Human oversight

We support your judgment, not replace it. Assumptions are explicit and editable. You control every step of the process.

Full auditability

Data Privacy

Equitable modeling

Designed for sensitive policy contexts

Policy work demands rigour, transparency, and control. Every part of Prior Foundry is designed with that in mind.

Human oversight

We support your judgment, not replace it. Assumptions are explicit and editable. You control every step of the process.

Full auditability

Data Privacy

Equitable modeling

Who we are

We combine academic leadership in computational behavioral science with real-world policy evaluation and large-scale engineering experience.

Shirin Abrishami Kashani

Co-founder

PhD candidate in Political Science at Stanford University. Former policy analyst at the OECD. Bridges the gap between research methodology and operational policy needs in public administrations.

Dr. Jonne Kamphorst

Co-founder

Assistant Professor, Sciences Po Paris. Previously postdoctoral scholar at Stanford Human-Centered Artificial Intelligence Lab. Research on LLMs for simulating human behavior.

Keshav Sivakumar

Co-founder

Computer Scientist trained at Cambridge & Stanford with experience building large-scale systems in industry. Responsible for the platform architecture and simulation infrastructure.

Prof. Dr. Robb Willer

Advisor

Professor at Stanford University. Director of the Politics and Social Change Lab, and the AI for Public Benefit Lab. Leading research on AI tools for public sector use.

Tell us about your policy domain

We design around real customer needs and maintain significant flexibility to tailor functionality to different operational contexts and requirements.

Book a demo