The problem Nova: Chama solves

Nova is a decentralized AI agent that revolutionizes traditional group finance (Chama) by leveraging blockchain and AI.

Nova tries to solve several problems in traditional group finance(Chama) systems.

These problems include:

• Lack of transparency in fund management
• Manual tracking of contributions and distributions
• Trust-based systems prone to mismanagement
• Complex and time-consuming decision-making processes
• Limited access to financial services

How does Nova make it better

• Pool funds securely.
• Make contributions directly through their Ethereum wallets
• Get real-time visibility of their contributions and pool status
• Participate in fair, weighted voting on community decisions
• Access loans with standardized terms and automated processing
• Receive transparent interest distributions
• AI-powered risk assessment for loans
• Automated interest calculations and distributions
• Secure end-to-end encrypted communication via XMTP

What are some of practical use cases of Nova:

a) Community Savings Groups:

• Members can contribute regularly to a shared pool
• Automated tracking of individual contributions
• Fair distribution of interest based on contribution size
• Transparent voting on how to use pooled funds

b) Micro-lending Circles:

• Members can request loans with 20% contribution as collateral
• Standardized 10% interest rate for all loans
• Automated loan processing and payment tracking
• Fair access to credit based on contribution history

c) Investment Clubs

• Community can propose and vote on investment opportunities
• Transparent tracking of investment performance
• Fair profit sharing based on contribution size

Challenges we ran into

Some of the challenges we encountered included:

1. Human-in-the-Loop Decision Making Challenge

Problem: Integrating human decision-making into an automated AI agent system
Solution: Integrated Lang-graph human interaction model loop and Implemented a weighted voting system where voting power is proportional to contribution size.

2. AI Agent memory persistence in between messages
   Solution: Temporary story agent state and memory outside and injected when new messages came in to prevent agent from forgetting.

3. Group Setup and AI agent interaction

• Took advantage of XMTP Group Conversations setup to intergrate the connected agent to the correct conversation with more that one user in the group setup