Evaluate AI Response Groundedness in RAG Systems

# Evaluate AI response groundedness using document retrieval with Google Sheets logging

This template enables systematic evaluation of how factually grounded AI agent responses are in relation to source documents. It is designed for developers and ML engineers who need to test retrieval quality and detect hallucinations in RAG systems. Evaluation results are automatically logged into Google Sheets for analysis and benchmarking.

## Who it´s for
- RAG system developers needing to test AI response groundedness
- Machine learning engineers evaluating retrieval performance in agents
- Teams deploying documentation-based chatbots who require reliability checks
- Analysts running model benchmarking on structured datasets
- Companies using n8n to automate AI application testing workflows

## What the automation does
- Loads a PDF document (e.g., Bitcoin whitepaper) into an in-memory vector store using OpenAI embeddings
- On input, a LangChain AI agent retrieves relevant content and generates a response
- The response and retrieved context are sent to a second LLM to evaluate factual grounding
- Evaluation results — including hallucination score, citation accuracy, and confidence — are saved to Google Sheets
- Can be triggered manually, by chat message, or via dataset ingestion

## What´s included
- Ready-to-use n8n workflow
- Trigger logic: manual execution, chat message received, evaluation dataset fetched
- Integrations with Google Sheets, OpenAI API, and PDF document source
- Basic setup and adaptation guide

## Requirements for setup
- n8n instance (cloud or self-hosted)
- OpenAI API key
- Google Sheets access (via Google Sheets API)
- Source PDF document or ability to upload a custom one

## Benefits and outcomes
- Reduced risk of AI hallucinations in production agents
- Quantitative assessment of retrieval and generation quality in RAG pipelines
- Automated logging for comparative analysis across model versions
- Faster iteration when testing new embeddings, chunking strategies, or models
- Transparent and reproducible evaluation process

## Important: template only
Important: you are purchasing a ready-made automation workflow template only. Rollout into your infrastructure, connecting specific accounts and services, 1:1 setup help, custom adjustments for non-standard stacks and any consulting support are provided as a separate paid service at an individual rate. To discuss custom work or 1:1 help, contact via chat