YouTube Video
Project Description
#NYC Mayoral Candidate Political Position Chatbot
Project Description
Following local politics has become increasingly difficult for everyday citizens. With busy schedules, fragmented news sources, and the sheer volume of information available, most New Yorkers struggle to understand where mayoral candidates actually stand on issues that matter to them—housing, public safety, education, transportation, and more.
We created an AI-powered chatbot that summarizes the political positions of NYC mayoral candidates, making it easy for voters to get clear, factual answers about candidate stances without having to wade through campaign websites, debate transcripts, and news articles.
Core User Flow
- User asks a question about a candidate or issue (e.g., “What is Cuomo’ position on affordable housing?”)
- The agent synthesizes the information using the voice of the candidate using a knowledge base based on mayoral debate
- Users can drill deeper by asking follow-up questions or requesting comparisons across multiple candidates
Judging Criteria
Working Prototype
The chatbot runs end-to-end in a conversational interface. Users can immediately start asking questions and receive accurate, sourced responses about NYC mayoral candidates.
Innovation & Creativity
Rather than building yet another general-purpose chatbot, we focused on solving a specific civic problem: voter information access.
Real-World Impact
Lowers the barrier to political participation by making candidate research accessible and saves voters hours of research time.
Promotes informed civic engagement in local elections, which historically have low turnout due to information gaps.
Theme Alignment
This project directly demonstrates the “agent” paradigm—an AI system that autonomously searches, retrieves, and synthesizes information to accomplish a user goal (understanding candidate positions) that would otherwise require significant manual effort.
Technical Complexity & Integration
LLM workflow in Elevenlabs with candidate voice. Scraping and parsing political information using LLM.
Prior Work
None