https://github.com/truongd3/ai-powered-restaurant-assistant

AI-Powered Restaurant Assistant: Create an Agent to support restaurant customers and a Table Booking system with Claude-3 on Amazon Bedrock
https://github.com/truongd3/ai-powered-restaurant-assistant

aws bedrock claude-3

Last synced: 7 months ago
JSON representation

AI-Powered Restaurant Assistant: Create an Agent to support restaurant customers and a Table Booking system with Claude-3 on Amazon Bedrock

• Host: GitHub
• URL: https://github.com/truongd3/ai-powered-restaurant-assistant
• Owner: truongd3
• Created: 2024-07-13T02:52:52.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2024-07-13T03:10:57.000Z (over 1 year ago)
• Last Synced: 2025-03-28T00:04:04.307Z (7 months ago)
• Topics: aws, bedrock, claude-3
• Language: Jupyter Notebook
• Homepage:
• Size: 411 KB
• Stars: 3
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Creating Agent with Knowledge Base and an Action Group

This project involves creating an agent with Amazon Bedrock and integrating it with a Knowledge Base and an Action Group. With this integration, the Agent will be able to respond to a user query by taking a sequence of actions, consulting the knowledge base to obtain more info, and executing tasks using the lambda function connected with an Action Group.

## Agent Architecture

Let's create a **AI-Powered Restaurant Assistant** agent that connects with a Knowledge Base containing the restaurant's different menus. This agent also connects to an action group that handles table booking. 

![Agents architecture - showing an agent responding on one end using APIs and action groups and then on the end responding to other questions with a knowledge base on a vector database](images/architecture.png)

The action group created in this project uses [function details](https://docs.aws.amazon.com/bedrock/latest/userguide/agents-action-function.html) to define the functionalities for 

`create_booking`, `get_booking_details` and `delete_booking`.

The action group execution connects with a Lambda function that interacts with an Amazon DynamoDB table.

_Learn more_: [https://catalog.us-east-1.prod.workshops.aws/workshops/4c28c535-249d-44dc-939b-9f0942078336/en-US/50-create-agent-with-knowledge-base-and-action-group](https://catalog.us-east-1.prod.workshops.aws/workshops/4c28c535-249d-44dc-939b-9f0942078336/en-US/50-create-agent-with-knowledge-base-and-action-group)