AWS Builder Center

Building a Live Knowledge-Sync Club Member Portal with a Document-Grounded AI Chatbot

Building a Live Knowledge-Sync Club Member Portal with a Document-Grounded AI Chatbot

Learn how an AI-powered AWS Knowledge Assistant uses Retrieval-Augmented Generation (RAG) to deliver accurate, document-based answers. This solution enables organizations to manage knowledge efficiently while ensuring reliable and up-to-date responses.

🔷Introduction As organizations grow, finding accurate information quickly becomes increasingly challenging. Traditional chatbots often generate responses based on pre-trained knowledge, which can lead to outdated or incorrect answers. To address this challenge, Retrieval-Augmented Generation (RAG) combines the power of Large Language Models (LLMs) with organizational knowledge sources.
➡️ What We Built
- Admin document add/update and publish - Automatic re-indexing after publishing - Live chatbot knowledge synchronization - Updated member document list - Source citations for chatbot answers - `event-day-briefing.md` integration - Three-question smoke test - `POST /ask` API on port `8080`
✅How It Works
➡️Main flow of the system
**Admin publishes → Re-index → Knowledge updated → Member asks → Relevant content retrieved → Answer + source returned**
✅ Document-Grounded Chatbot
The chatbot answers only from club documents. It processes the question, searches indexed sections using relevance scoring, and returns a focused answer with its source.
If information cannot be found, the system uses the configured AWS Student Builder contact fallback instead of guessing.
✅ Live Knowledge Sync
When an admin publishes a new or updated document, the content is re-indexed. Subsequent chatbot queries use the updated index, while the member dashboard displays the updated document list without a full page reload.
✅Event-Day Briefing
The required `event-day-briefing.md` document was added to the knowledge base. It contains event information such as schedules, rooms, judging times, and deadlines. Chatbot answers from this document include its filename as the source.
✅ Smoke Test
`09-smoke-test-questions.md` is used to validate the chatbot's retrieval, answers, and sources. The three required questions are demonstrated as PASS or FAIL.
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article