Politicon

Inspiration

I grew up confused by politics. Not because I wasn't paying attention, but because the adults around me were just as confused, just louder about it. My family was united by blood but divided by belief. Holidays turned into debates, relationships grew tense, and somewhere in the middle of all of it, I realized that nobody was actually talking about the same thing. They were talking past each other, shaped by whatever channel they watched or whatever their parents believed. I was too. I didn't form my own political views so much as inherit them, conditioned to lean a certain direction before I even understood what that direction meant.

That stayed with me. And the older I got, the more I noticed that the real casualty of political division wasn't any one election or policy. It was clarity. People weren't uninformed because they were lazy. They were uninformed because the information was never built for them. Legislation is written in legal language. Economic impact is buried in congressional reports. And by the time it filters down to the average person, it has already been translated by someone with an agenda.

Politicon was built to short-circuit that entire process. Not to tell people what to believe, but to hand them the one thing that politics has always kept just out of reach: the math. Because when you stop asking "which side is right" and start asking "what does this actually do to my paycheck," the conversation changes. It stops being red versus blue. It becomes us versus the problems we all actually share.

What it does

Politicon is a personalized civic intelligence platform that translates complex legislation and economic policy into clear, individual financial impacts. By combining your specific financial profile (income, location, dependents, assets) with real-time data on federal and state bills, Politicon uses AI to calculate exactly how proposed laws will affect your tax bill, cost of living, and net worth. It moves the conversation from partisan rhetoric to tangible numbers, helping users understand why certain policies matter to their specific lives.

How we built it

Politicon is built as a full-stack web application using Next.js 14 with the App Router, TypeScript, and Tailwind CSS for the frontend. Animations are handled through Framer Motion and GSAP with ScrollTrigger, with Lenis providing smooth scroll physics throughout the experience.

On the backend, Supabase powers authentication, user profile storage, and row-level security so that every user's data stays private and scoped to them alone. All AI analysis runs through the Claude API by Anthropic, called exclusively through server-side Next.js API routes so no keys are ever exposed client-side. Each API call is injected with the user's full financial profile as system context, which is what makes the output personalized rather than generic.

Economic modeling is grounded in hardcoded US data: all 50 state tax structures, 2025 federal brackets, cost-of-living indices by region, and sector employment data. This gives the AI a factual base to reason from rather than generating estimates from nothing. The design language was built to feel like a premium civic product, not a government tool. Liquid glass panels, ambient gradient meshes, scroll-driven narrative sections, and a typography system built on editorial fonts were all deliberate choices to make policy feel as approachable as a consumer app.

Challenges we ran into

Monetization without compromising the mission: The core promise of Politicon is that financial clarity should be accessible to everyone. But building and running an AI-powered platform at scale has real costs, and a product without a revenue model is not a product, it is a prototype. The challenge was identifying which features create enough value that users would pay for them without locking the most important functionality behind a paywall. The answer we landed on: the basic personalized analysis stays free, while premium features like the AI advisor with unlimited queries, policy tracking with real-time alerts, and the exportable impact report sit behind a subscription. The public impact explorer stays fully open to drive organic traffic and trust.

AI accuracy and the ethics of overpromising: Policy impact modeling is genuinely hard. Real economists disagree on projected effects of legislation all the time. An AI model working from a user's profile can produce a directionally accurate estimate, but it cannot produce a guarantee. The temptation when building something like this is to present outputs with false confidence because confidence converts. We pushed back on that. Every analysis includes a methodology note and a clear disclaimer that projections are estimates based on available data. Transparency is not a weakness in a product like this. It is the product.

Scope versus execution: The vision for Politicon is large. The hackathon timeline is not. Deciding what to build fully versus what to scaffold intelligently was a constant negotiation, and shipping something coherent and functional mattered more than shipping everything half-finished.

Accomplishments that we're proud of

Personalized Precision: Successfully integrated hardcoded US economic data (50-state tax structures, 2025 brackets) with dynamic AI reasoning, allowing for granular, user-specific projections rather than generic averages.
Trust Through Transparency: Built a system that explicitly displays methodology and confidence intervals, setting a new standard for ethical AI in financial advisory tools.
Seamless User Experience: Created a premium, editorial-grade interface that makes dense legislative data feel approachable and intuitive, bridging the gap between government complexity and consumer app usability.
Secure Architecture: Implemented robust Row-Level Security (RLS) via Supabase and server-side API routes, ensuring sensitive financial data remains private and secure by design.

What we learned

Transparency is a feature, not a disclaimer. The instinct when building AI-powered tools is to project confidence, but in a domain as consequential as personal finance and public policy, acknowledging the limits of what the model can and cannot know is what actually builds trust. We learned that being honest about accuracy, methodology, and potential error does not undermine the product. It is what makes the product credible.

We also learned that the best product ideas do not come from market research. They come from something personal that never stopped bothering you. Politicon exists because political division was something I lived inside of growing up, not something I read about. That specificity is what gave the idea its direction.

And practically: scope is a design decision. What you choose not to build in a hackathon is just as important as what you do. I've always believed that the best ideas didn't come from geniuses, but came from the environment itself. The way you interact with your environment and take advantage of your resources dictates your success.

What's next for Politicon

The immediate next step is expanding the policy database and improving the AI's ability to reason about state-specific legislation, not just federal bills. Most policies that affect people's daily finances happen at the state and local level, and that layer is currently underrepresented.

After that, the email digest feature, where users receive a weekly personalized summary of new policies that match their profile, is the highest-priority growth lever. It creates a reason to come back without requiring the user to remember to.

Longer term, the vision is a civic financial layer that sits underneath how people think about elections, not just individual policies. When you can see the cumulative dollar impact of one candidate's platform versus another's, scoped to your actual life, the conversation around voting changes. That is the version of Politicon worth building toward.

The divide that inspired this project will not be fixed by an app. But it can be narrowed, one honest dollar figure at a time.