OneHorizons Sense: Civic Signal

• 

  OneHorizons Sense Logo Butterfly

1. Inspiration - The Situation

Governments often learn about societal problems only after they have escalated into crises involving the police, emergency services, or social care.

At the same time, citizens constantly observe what is happening around them, but these observations remain fragmented across social media, messaging apps, and conversations. Valuable local knowledge is lost because it is unstructured and difficult to interpret at scale.

We wanted to explore whether GPT-5.6 could transform messy human observations into structured civic intelligence that helps municipalities understand emerging issues earlier.

1. What it does - The Complicating Factors

OneHorizon Sense: Civic Signal allows citizens to describe situations they observe in natural language.

The platform uses GPT-5.6 to transform unstructured observations into:

• What's happening
• Why it matters
• Core policy question
• Related policy domains
• Relevant Sustainable Development Goals
• Confidence score
• Fully traceable source input

The vision is to aggregate thousands of observations into meaningful signals that municipalities can use to identify trends before they become crises.

1. How we built it - Architecture

The application was built using:

• GPT-5.6
• OpenAI Codex
• Next.js
• React
• TypeScript
• Supabase
• Vercel

Codex accelerated implementation by helping us build and iterate on the application while GPT-5.6 powers the contextual reasoning engine.

1. Challenges we ran into - Developing

The hardest challenge was not generating summaries.

It was preserving nuance while producing structured information that governments could actually use.

We also wanted every conclusion to remain transparent by allowing users to inspect the original input.

1. Accomplishments that we're proud of - Sense

Our biggest achievement is demonstrating that GPT-5.6 can transform complex societal observations into structured policy intelligence rather than simply generating summaries.

The application produces consistent reasoning, identifies underlying policy questions, maps societal themes, and connects observations to the Sustainable Development Goals.

1. What we learned - Lifelong learning

Large language models become significantly more valuable when they structure human reasoning instead of replacing it.

The combination of AI reasoning and human validation has the potential to create a new feedback mechanism between citizens and governments.

1. What's next for OneHorizon Sense: Civic Signal - The Horizon

Our next milestone is collective intelligence.

Instead of analysing individual observations, municipalities will be able to identify emerging societal trends through validated citizen input.

The long-term vision is an AI-powered civic participation platform that helps governments detect issues earlier, prioritize interventions, and create a more evidence-based dialogue with society.

Built With

• chatgpt
• github
• openai
• supabase
• vercel
• vscode