• 

  Sample output of the system

DeepDive: AI-Powered Multi-Agent Discussions for Informed Decision-Making

Inspiration

Life is full of complex decisions, from voting to major purchases. For unfamiliar topics, decision-making can be stressful. DeepDive aims to alleviate this by simulating expert discussions, providing users with deeper insights and a more enjoyable decision-making process.

What it does

DeepDive creates a framework where multiple AI agents, guided by a moderator, engage in discussions on specific topics. Users can observe these dynamic conversations to gain comprehensive insights on their areas of interest.

How we built it

• Utilized the Fetch.AI framework to host and manage multiple agents
• Implemented tool capabilities using ToolHouse
• Powered by Llama language model hosted on Groq

Challenges we overcame

• Implementing effective messaging in a decentralized multi-agent system
• Balancing agent autonomy with coherent, purposeful discussions

Key achievements

• Successfully orchestrated three AI agents communicating cohesively to provide user insights
• Created a novel approach to information gathering and decision support

Lessons learned

• Gained deep understanding of multi-agent system dynamics
• Discovered how collaborative AI can achieve outcomes greater than the sum of its parts

Future directions

• Refine the core concept for product development
• Expand topic coverage and agent specializations
• Enhance user interaction and customization options

Built With

• fetch.ai
• groq
• llama
• openai
• python
• toolhouse.ai