Check out the video demonstration of this Feedback Agent on LinkedIn.
An automated Feedback Agent built in n8n that processes customer reviews submitted through a QR‑code form. The workflow reads each new entry from Google Sheets, analyzes it with Google Gemini, and routes it to the correct team automatically — turning raw feedback into structured, actionable insights.
- Automates the entire feedback‑handling pipeline
- Uses AI to classify customer messages
- Updates the source sheet with clean, normalized labels
- Routes each item to the correct team via Gmail
- Eliminates manual sorting, scanning, and triage
This agent acts like a small digital employee — always on, always consistent.
For a full breakdown of the workflow, including node‑by‑node explanations, routing logic, and diagrams, see:
👉 workflow-overview.md
This README intentionally stays high‑level to avoid repeating the detailed documentation.
- n8n for workflow automation
- Google Sheets as the data source
- Google Gemini for AI classification
- LangChain for structured output parsing
- Gmail for notifications
The Feedback Agent connects to several external apps — Google Sheets, Google Gemini, and Gmail — and each one needs a secure way to confirm your identity.
In n8n, apps authenticate through three methods (see image in repo):
- OAuth2 — log in and approve access; the service issues a temporary token
- API Keys — provide a secret key; anyone with the key can access the service
- Basic Authentication — username + password; typically used for older internal tools
For this project:
- Google Sheets uses OAuth2
- Google Gemini uses an API Key
- Gmail uses OAuth2
Once these connections are set up, the Feedback Agent can read new entries, analyze them with AI, and route each message automatically.
- All customer feedback comes in through a QR‑code form
- Every submission lands in one Google Sheet
- The agent detects new rows, analyzes the text with Google Gemini, and routes the message to the right team
This is the only workflow summary included in the README — the deeper logic, nodes, and routing rules remain in workflow-overview.md.
workflow.json— exported n8n workflowworkflow-overview.md— full technical documentationREADME.md— project summary and demo link
Small cafés and local businesses often receive valuable feedback but lack the time to process it.
This agent closes that gap by delivering:
- Faster response times
- Cleaner data
- Better visibility into customer needs
- Zero manual overhead
A practical example of how automation and AI can elevate everyday operations.