Proposed: Creation of an ‘Ask V-Sekai’ Chatbot Using Vercel, OpenAI, and Neon PostgreSQL

Metadata

  • Status: Proposed
  • Deciders: V-Sekai Development Team
  • Tags: V-Sekai, Chatbot, Vercel, OpenAI, Neon, PostgreSQL, AI, Serverless

The Backdrop

V-Sekai aims to provide immediate, AI-driven answers to community queries using a chatbot integrated with Discord.

The Challenge

How to develop an ‘Ask V-Sekai’ chatbot that utilizes Vercel serverless functions for interaction, OpenAI for natural language understanding, and Neon PostgreSQL for storing and retrieving knowledge?

The Strategy

  • Design and deploy a serverless function on Vercel that interfaces with Discord to receive questions.
  • Use OpenAI’s API to process natural language queries and generate responses.
  • Store and retrieve chatbot training and interaction data using Neon’s PostgreSQL service, leveraging the pgvector extension for efficient similarity analysis.
  • Continuously improve the chatbot’s accuracy and response quality through data analytics and model fine-tuning.

The Upside

  • Quick, accurate responses to user inquiries, enhancing the V-Sekai community experience.
  • Leverage cutting-edge AI and serverless technologies for a scalable, low-overhead solution.
  • Gather valuable interaction data to refine the chatbot and community support strategies.

The Downside

  • Initial complexity in setting up the integration between Vercel, OpenAI, and Neon PostgreSQL.
  • Ongoing maintenance and potential costs associated with OpenAI API usage.

The Road Not Taken

  • Traditional FAQ systems that do not offer real-time, AI-driven interaction and learning capabilities.

The Infrequent Use Case

  • Complex queries that require human intervention and are not yet within the scope of the AI’s training.

In Core and Done by Us?

  • The development, deployment, and maintenance of the ‘Ask V-Sekai’ chatbot will be managed internally as a strategic tool for community engagement.

Further Reading