CivicSense AI

Project Story: CivicSense AI

1. Project Title

CivicSense AI — Turning Community Problems Into Action

1. One-Line Pitch

CivicSense AI is an AI-powered civic action agent that converts community complaints from voice notes, images, and text into structured issue reports, priority scores, official letters, and follow-up dashboards.

1. The Story Behind the Problem

In many communities, people face real problems every day: broken streetlights, bad roads, choked gutters, flooding areas, abandoned boreholes, sanitation issues, school challenges, and unsafe public spaces.

But most of these problems are only discussed on WhatsApp groups, radio call-ins, Facebook posts, or informal conversations. The complaints are real, but they are not properly documented. Evidence is scattered. No one clearly knows which office should handle the issue. Follow-up is weak. Many problems remain unresolved because they never become structured civic action.

This creates a big gap between citizens, community leaders, and local authorities.

CivicSense AI was created to close that gap.

Instead of allowing complaints to disappear, CivicSense AI captures them, understands them, organizes them, and turns them into action-ready reports.

1. The Problem We Are Solving

Community problems usually fail to get solved because of five major issues:

Complaints are unstructured. Evidence is not properly attached. The responsible department is unclear. There is no priority or urgency scoring. There is no memory or follow-up system.

For example, someone may report that a gutter near a school floods anytime it rains. That complaint may appear in a WhatsApp group, but it may never reach the right authority in a clear format.

CivicSense AI changes this by converting messy complaints into official, trackable civic cases.

1. The Solution

CivicSense AI allows citizens, volunteers, assembly members, NGOs, schools, and local government officers to submit community problems using:

voice notes text photos location short descriptions

The AI agent then analyzes the submission and produces:

issue category urgency level affected people location summary evidence summary responsible office suggestion official complaint letter public awareness post follow-up reminder dashboard/map status

This makes every complaint easier to understand, easier to forward, and easier to track.

1. How the AI Agent Works

CivicSense AI is designed as a multi-agent civic workflow system, which fits the Qwen Cloud hackathon’s focus on AI agents. The hackathon includes tracks such as MemoryAgent, Agent Society, Autopilot Agent, and EdgeAgent, and it is focused on building sophisticated AI agent systems using Qwen Cloud/Alibaba Cloud tools.

The system can be divided into several AI agents:

Agent 1: Complaint Understanding Agent

This agent reads or listens to the citizen’s complaint and understands the core problem.

Example input:

“The gutter near the school floods anytime it rains and children cannot cross.”

The agent identifies:

problem type: drainage/flooding affected group: school children and residents risk: high need: urgent inspection and desilting Agent 2: Evidence Agent

This agent reviews uploaded photos, text, and location details. It summarizes the evidence and checks whether more information is needed.

Example:

“Photo shows stagnant water near a school entrance. Location landmark is missing. Ask user to add nearest landmark.”

Agent 3: Priority Scoring Agent

This agent gives the issue an urgency score based on:

public safety risk number of people affected school/health/community impact repeated reports time sensitivity

Example:

Urgency Score: 8/10 Reason: Children are affected, flooding happens repeatedly, and sanitation risk is high.

Agent 4: Responsible Office Agent

This agent suggests which office or department should handle the issue.

Example:

Assembly Works Department Environmental Health Unit NADMO, for flood-risk situations Education Directorate, for school-related issues Agent 5: Civic Document Agent

This agent automatically generates:

official complaint letter petition draft incident report social media awareness post follow-up message Agent 6: Memory & Follow-Up Agent

This is where the project becomes powerful. The agent remembers previous reports from the same area and tracks whether the issue has improved, remained the same, or worsened.

This fits the MemoryAgent idea because the agent uses persistent memory to compare new complaints with past reports. The Qwen Cloud hackathon has a MemoryAgent track focused on agents that use long-term memory and context across sessions.

1. Example User Journey

A citizen opens CivicSense AI and clicks Report a Community Issue.

They record a voice note:

“There is a broken streetlight around the junction. The place becomes dark at night and people are afraid to pass there.”

The user adds a photo and location.

CivicSense AI responds:

Issue Category: Public Safety / Streetlight Urgency Level: Medium-High Affected Group: Residents, students, traders, pedestrians Suggested Responsible Office: Assembly Works Department / Electricity Unit Recommended Action: Streetlight inspection and replacement Generated Letter: Ready Follow-up Reminder: 14 days Status: Open

The system then places the issue on a map and stores it in memory.

If another person reports the same issue later, the AI does not create a duplicate. Instead, it links the new report to the existing case and increases the priority score.

1. Why This Project Matters

CivicSense AI can help communities move from complaint to action.

It can be useful for:

citizens assembly members local government offices NGOs schools community volunteers journalists disaster response teams sanitation and health officers

The project is powerful because it does not only generate text. It creates a civic workflow.

It helps answer:

What is the problem? Where is it happening? Who is affected? How urgent is it? Which office should handle it? What evidence is available? Has it been solved? What follow-up is needed?

Built With

• and-dashboards.-backend-fastapi-fastapi-is-good-for-ai-projects-because-it-is-python-based
• buttons
• clean-web-application.-vite-is-simpler-and-lighter-than-next.js-for-a-hackathon-mvp.-ui-design-tailwind-css-+-shadcn/ui-use-tailwind-for-fast-styling-and-shadcn/ui-for-clean-components-like-cards
• dialogs
• easy-to-connect-with-ai-apis
• forms
• frontend-react-+-vite-+-typescript-use-this-for-a-fast
• sperbase