internal-docs.bot / query-engine
SIMULATED RAG

Welcome to the ACME Corp Knowledge Search. I can answer questions about our WFH policies, expenses, security, and more.

Scanning Vectors...
Try:

RAG Logic Pipeline

01

Vector Search Simulation

Instead of checking the entire knowledge base, we extract keywords from the query and retrieve relevant document chunks from the backend DATA store.

02

Contextual Prompting

The retrieved text is injected into a "Safe Assistant" system prompt. This ensures the AI doesn't hallucinate info outside the company's official stance.

03

Source Attribution

Each answer is returned with metadata tracking the sources used, allowing for transparency and "Human in the Loop" verification.

Typical Notion/n8n Flow

Webhook: User asks in Slack/Tally
Lookup: Notion Database Search
Filter: Node filters by relevance score
AI: OpenAI/Anthropic generates answer

This demo uses a local Python simulation of the above flow to provide instantaneous feedback while demonstrating the logical sequence.