Unified FAQ AI Assistant

Overview

LumiWealth, a U.S.-based leader in algorithmic trading education, wanted to improve how their users accessed information. Their existing chatbot had limited reach, it couldn’t pull answers from the many places where LumiWealth’s content lived: Discord, GitHub, their public website, and internal files in an S3 bucket.

The challenge? Centralizing that knowledge without actually centralizing it.

That’s where AI Point stepped in, with a Retrieval-Augmented Generation (RAG) solution that connects all of LumiWealth’s sources and delivers instant, accurate answers to users through a single AI assistant.

The Challenge

LumiWealth’s growing user base needed fast, accurate answers. But their information was spread across:

• Discord: Real-time community conversations and pinned answers
• GitHub: Code, issues, and documentation
• Website: Static FAQs and help center articles
• S3 Bucket: Internal documents and long-form guides

Their existing chatbot was siloed, it couldn’t tap into this wealth of knowledge, leading to missed answers, user frustration, and increased load on the support team.

Our Solution

AI Point designed and delivered a unified FAQ AI Assistant built on top of a RAG architecture, deployed securely on AWS Cloud. Here’s how it worked:

Multi-Source Retrieval

We built a pipeline to ingest, index, and regularly sync data from:

• Discord threads & pinned messages
• GitHub repos (issues, wikis, READMEs)
• LumiWealth’s website (FAQs, help articles)
• AWS S3 (PDFs, docs, spreadsheets)

Context-Aware Answering

Using a Retrieval-Augmented Generation pipeline, we:

• Embedded and stored the content in a vector database
• Used OpenAI’s GPT model to generate natural language answers
• Dynamically injected the most relevant data into the prompt

Seamless Integration

The assistant was deployed as an API and integrated with LumiWealth’s existing chatbot UI, so the user experience remained smooth, only now, it was smarter.

Secure & Scalable on AWS

We containerized the backend, deployed it on AWS, and ensured scalability and uptime with managed services and auto-scaling policies.

Results

LumiWealth’s AI assistant now acts as a single source of truth, capable of understanding and responding to complex queries, regardless of where the answer lives.

Impact:

• 60% reduction in support questions routed to human agents
• Instant answers from Discord, GitHub, and internal docs
• No manual data centralization required
• Future-proofed with easy addition of new sources or documents

Client Insight

“We wanted to take our chatbot to the next level, one that could tap into all our resources, not just canned responses. What AI Point built is exactly that. It’s like giving our users a search engine that actually understands them.”
— Robert, Founder, LumiWealth

Tech Stack

• LLMs: OpenAI GPT
• Architecture: Retrieval-Augmented Generation (RAG)
• Infrastructure: AWS (EC2, S3, ECS, CloudWatch)
• Vector Database: FAISS
• Integrations: Discord, GitHub, Web Scraper, S3
• Deployment: Containerized backend with REST API

Explore More AI Use Cases

LumiWealth’s solution is part of a growing trend, where GenAI goes beyond chatbots and powers intelligent, integrated workflows.

Check out similar projects:

• Stock News Analyzer with Multi-Agent Coordination
• Cloud Infrastructure Management for Trading Bots
• Algorithmic Trading Monitoring System

Need a Smart AI Assistant That Connects Everything?

We specialize in AI solutions that plug into your tools and data, no matter where they live.

📩 Get in touch to explore a tailored RAG-powered solution for your team: www.aipoint.io/contact

‍