Case study  ·  Chat & Support

A chatbot for a fast food restaurant that answers from the latest menu

It keeps itself up to date from files in Google Drive and uses them to answer menu questions, help with orders and give restaurant details.

The problem

Customers ask the same things about the menu, prices, opening hours and orders. Keeping a chatbot's answers right every time the menu changes is the hard part.

How it works

  1. When a new file is added to the restaurant's Google Drive folder, such as an updated menu, the workflow starts.
  2. It downloads the file.
  3. Google Gemini turns the content into embeddings the chatbot can search.
  4. The data is stored in Pinecone, so answers can be found quickly.
  5. When a customer asks a question, the chatbot searches that data and replies.
Menu sheet with item name, category, description, price and calories for burgers and wraps
Part of the menu the chatbot learns from.
Pinecone index settings showing cosine metric, 768 dimensions, AWS us-east-1, serverless and a record count of 123, with the host URL blurred
The Pinecone index holding the menu data. The host URL is blurred.

The chatbot helps customers order, customise their meals and answer common questions such as store hours and promotions.

Chat window where a customer asks about opening hours and the chatbot replies with the hours, with the session ID and restaurant name blurred
A customer asking about opening hours.
n8n chat agent workflow: chat trigger, AI agent with a Google Gemini chat model, simple memory and a vector store tool backed by Pinecone and Gemini embeddings
The chatbot in n8n.

What a person still handles

  • Updating the menu file. The chatbot picks up the change from Drive.
  • Preparing and delivering the orders.

Built with

  • n8n
  • Google Drive
  • Google Gemini
  • Pinecone
  • RAG

n8n, Google Drive as the source of menu files, Google Gemini for embeddings and replies, and Pinecone as the vector store.

Your build starts with the arithmetic.

Bring a week of real examples to the free audit. We come back with the three worth doing first, what each is worth, and a fixed price. If a cheaper tool already covers it, we say so.

Book the free audit →

All case studies