In an era where policy decisions have far-reaching social and economic consequences, Jan Niti is leveraging AI to help governments anticipate the real-world impact of their initiatives before implementation. By combining AI-driven compliance checks with advanced policy simulation, the platform evaluates legal validity, economic feasibility, and potential conflicts with existing regulations. Its multi-agent environment models stakeholder responses dynamically, representing groups such as farmers, teachers, unions, and government departments, and drawing insights from historical data, pilot projects, and unstructured sources like news reports and bureaucratic notes. Beyond predicting reactions, Jan Niti also forecasts economic outcomes, helping policymakers avoid wasted resources, fund leakage, and unintended consequences.
In an exclusive discussion with The Interview World at the India AI Impact Expo 2026, Rishika Saraswat and Saksham Pruthi, Co-founders of Jan Niti, delve into the company’s technology, methodology, AI models, data sources, virtual stakeholders, and the broader impact of their AI-driven solutions on policy making. Here are the key insights from their insightful conversation.
Q: Could you elaborate on the AI-powered solutions that Jan Niti provides?
A: We provide a solution that enables governments to anticipate the real-world impact of their policies before implementation. Our platform combines AI-powered compliance checks with advanced policy simulation. The compliance module assesses legal validity, economic feasibility, and potential conflicts with existing policies. Meanwhile, our AI-driven policy simulation creates a multi-agent environment in which virtual agents represent each stakeholder. These agents interact dynamically, allowing us to model and predict stakeholder responses to proposed policies with high fidelity.
Q: Who are the stakeholders, and how do you engage or create them virtually?
A: We deploy distinct AI agents to represent each stakeholder, such as farmers, teachers, unions, housing authorities, and education departments. Each agent is trained on historical data that captures how these stakeholders reacted to similar policies or decisions in the past. We also incorporate insights from government pilot projects, which test policies on a small scale in select districts. Then, we evaluate how these patterns scale to the broader real world, leveraging diverse unstructured data, including news reports, bureaucratic notes, and other relevant sources. By integrating all these inputs, our platform accurately predicts how stakeholders are likely to respond once a policy is implemented.
Q: What trusted data sources do you use to train your AI models?
A: Currently, we rely on historical news reports from trusted, vetted sources, carefully reviewed by our team rather than automated LLMs. We complement this with bureaucratic notes and other unstructured data that document how past decisions unfolded and the impacts they generated. For example, a policy rollout might have triggered protests in a specific region, and we capture such outcomes in our dataset.
Since historical data alone is often limited, we also generate simulated data to fill gaps. Our next step is to gain access to government ministries or agencies to run a targeted pilot project. This would allow us to work with the actual data that bureaucrats and policymakers use to predict policy impacts today. Using these insights, we can design a pilot that aligns precisely with the needs and priorities of decision-makers, demonstrating the platform’s predictive capabilities in a real-world context.
Q: Do you also forecast the economic impact of government policies?
A: Absolutely. We often observe that policies are implemented only to trigger protests or unintended consequences. Frequently, resources are allocated to projects that do not achieve their intended outcomes. Every policy rollout carries both financial costs and social impact. Building infrastructure, for example, consumes government budgets and taxpayers’ money, often without guaranteed benefits.
By using Jan Niti proactively, policymakers can evaluate in advance whether a proposed policy will deliver positive outcomes. If a policy risks generating protests or adverse effects, these issues can be anticipated and mitigated before implementation. In this way, the platform helps preserve resources, reduce public disruption, and maximize the impact of government spending.
Q: How do you address challenges such as fund leakage and project overestimation?
A: While strategic planning is a focus for later stages, we can already see the impact on operational policies, such as infrastructure deployment. For example, in a teacher transfer policy, staff may need to be relocated across regions, requiring housing, transportation, and other logistical support. Budgets were allocated for these resources, but because the policy design did not account for real-world constraints, it failed, and the resources were wasted.
Subsequently, policymakers had to redesign the policy, introducing limits on transfer distances and reallocating resources again. With a platform like Jan Niti, such inefficiencies could be anticipated and avoided, saving both money and time while ensuring policies are implemented effectively from the outset.
Q: Has this product been implemented on the ground, or is it yet to be launched?
A: Currently, we have developed a product and prepared a demonstration specifically for the Uttar Pradesh government. We are actively working to secure pilot projects with government agencies. Although our team is only two months old, we have engaged closely with bureaucrats and technology experts to refine and validate our approach. Based on these conversations, the feedback has been clear: this platform could provide significant value and support for policymaking.
