Revolutionizing Polkadot’s On-Chain Governance with AI

‍

Polkadot, a pioneering blockchain platform, prides itself on its decentralized governance system. The token holders are who decide which proposals are enacted, placing the network's evolution in the hands of the community.

‍

For decentralized governance to thrive, it’s crucial that the community remains engaged, well-informed, and capable of understanding active proposals and the governance system. However, given the busy lives of community members, achieving this level of engagement is challenging.

‍

We saw an opportunity to harness recent advancements in Large Language Models (LLMs) to address this challenge. LLMs excel in understanding natural language, generating text, summarizing content, and performing sentiment analysis. By integrating these capabilities, we developed an AI assistant, that answers users questions about: 

• The governance system and Polkadot in general
• Existing proposals, their objective, milestones, budget, team, etc
• Proposal summaries
• Comments and reactions to proposals, facilitating a deeper understanding of community sentiments.

‍

Behind the scenes the solution consists of 2 main components:

1. An ETL Process that gathers data from Polkadot documentation and proposals, preprocesses it and stores it in optimized databases for efficient retrieval. It also generates proposal summaries and stores them to avoid the need to generate them in real time. 
2. An AI agent developed using langchain. An AI Agent is a program that uses an LLM to determine the actions it has to take and the tools it has to use to accomplish a specific task. The agent was provided access to the following databases as tools to enable it to accomplish its objectives:

• The polkadot documentation vector database
• The proposal vector database
• The summaries sql database
• The proposal comments and reactions sql database

‍

Our solution transformed the way Polkadot’s community interacts with the governance system. The AI-powered chat interface made it easier for users to stay informed, understand complex proposals, and engage in meaningful discussions. This not only increased community participation but also enhanced the overall efficiency and transparency of the governance process.

‍