The Problem

1. DAOs are riddled with human challenges: greed, incompetence, apathy and lack of participation, lack of subjects matter expertise, among others.
2. web3 solutions are increasingly fragmented, meaning that even if we get great new products, they don't have a clear Product Market Fit as the other elements they ned to succeed are in other chains. For Example: an innovative AI solution that is isolated and cannot communicate with the DAOs where it is meant to be performing transactions on.

The Solution

We deploy an AI Oracle to be able to inject advanced logic into NEAR protocol smart contracts. Our governance agent would consider proposals on gov. forums and post its throughts, reply to others who are also commenting, and decide whether to support or not the project.

Additionally, we implement Chain Signatures to be able to participate in Governance regardless of where the governance takes places. For example, our agents and AI oracle can assess a proposal from any EVM and automatically participate in that governance based on results.

Proof of Concept During this Hackathon we built a Proof of Concept that includes:

• Prompting of first the AI Agent entering the prompts manually.
• Deployed several DAOs on Ethereum
• Established chain signatures
• Successfully performed several actions on Ethereum DAOs (creating proposals, voting for proposals) while signing with the same NEAR wallet

What is Next

Time is limited and ambition is high! Next steps are:

1. Create a more complex network or agents that collaborate with each other (cf. just one agent now). We spent some time looking into frameworks such as CrewAI and Langchain for this but run out of time.
2. Trusted Execution Environment. We are aware of the shortcomings of the current oracle setup
3. Portofolio management/Private Equity: enable the agents to receive USDC, begin accumulating the tokens that they believe are underpriced, have potential for growth if governance is improved, etc.

1. AI Agent we are manually entering the prompt. Agent attributes are based on the basic agent framework from CrewAI
2. AI Oracle expands on the original Onchain-Ai (https://github.com/evgenykuzyakov/onchain-ai). The oracle receives the message request and has a bot that is constantly checking for new messages. When a message is detected it is sent to Bot to execute action.
3. Chain signatures (https://linktr.ee/chainsignatures)