Cursor 2-Day AI Hackathon — Sentio

---

Sentio

AI-powered Change Management Assistant for tracking stakeholder sentiment and driving organizational transformation

Sentio helps change managers monitor stakeholder feedback, generate actionable recommendations, and gain insights throughout organizational change initiatives. The platform uses AI agents to analyze stakeholder sentiment, create targeted surveys, and provide data-driven guidance.

Tech Stack

• Frontend: React, TypeScript, Tailwind CSS, Vite
• Backend: Python, FastAPI, LangGraph
• Database: SQLite
• AI/ML: OpenAI GPT-4o, Tavily (web search)
• Hosting: Heroku (Docker)

How to Run

Prerequisites

• Node.js 18+
• Python 3.11+
• OpenAI API key
• Tavily API key (optional, for web search)

Setup

# Clone the repo
git clone https://github.com/your-team/sentio.git
cd sentio

# Set up environment variables
cp .env.example .env
# Add your API keys to .env:
# OPENAI_API_KEY=sk-...
# TAVILY_API_KEY=tvly-...
# LOCALE=en

Run Backend

cd backend
python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
pip install -e .

# Seed with demo data (optional)
python -m app.seed --scenario six_month

# Start the server
uvicorn app.main:app --reload

Run Frontend

cd frontend
npm install
npm run dev

The app will be available at http://localhost:5173

Features

• Stakeholder Management: Track stakeholder groups with Mendelow matrix positioning
• Impulse Tracking: Record and visualize stakeholder sentiment over time
• AI Recommendations: Generate actionable recommendations based on stakeholder feedback
• Survey Generation: Create targeted surveys with AI-generated questions
• Insights Dashboard: AI-powered insights and trend analysis
• Multi-language Support: Full i18n support (English/German)

Details

Architecture

The application uses a multi-agent architecture powered by LangGraph:

• Chat Agent: General change management assistant
• Survey Agent: Generates targeted stakeholder surveys
• Recommendations Agent: Creates actionable recommendations
• Insights Agent: Analyzes data patterns and generates insights
• Orchestrator: Routes requests to appropriate agents

Demo Scenarios

Seed the database with realistic demo data:

# Fresh project
python -m app.seed --scenario new_project

# 3 months of data
python -m app.seed --scenario three_month

# 6 months of data
python -m app.seed --scenario six_month

# 10 months of data
python -m app.seed --scenario ten_month

Scope Cuts

Due to the 2-day hackathon timeframe, the following features were intentionally out of scope:

Authentication & Authorization

• No user authentication system
• No role-based access control
• No multi-tenant support

Enterprise Integrations

• No connectors for employee directories (Active Directory, LDAP)
• No SharePoint/OneDrive integration for company documents
• No HRIS system integrations
• No SSO/SAML support

Survey Distribution

• Currently exports surveys as Markdown files only
• No built-in survey hosting or response collection

Document Processing

• Knowledge base supports text files only
• No PDF, Word, or PowerPoint parsing

CI/CD & Testing

• No automated test suite
• No CI/CD pipeline
• Manual deployment only

Future Roadmap

Survey Channels (Planned)

• Web Survey Page: Hosted survey links with anonymous responses
• AI Phone Calls: Automated voice surveys for broader reach
• Email Campaigns: Survey distribution via email with tracking
• Chatbot Integration: Embed surveys in existing chat platforms (Slack, Teams)

Enterprise Features (Planned)

• User authentication with SSO
• Integration with HR systems for automatic stakeholder imports
• SharePoint connector for document ingestion
• Advanced analytics and reporting dashboards