A multi-agent system is designed by creating a supervisor agent named 'bedrock-supervisor-agent' with enabled multi-agent collaboration. The agent assesses user questions and directs them to the appropriate agents, specifically for inquiries related to accommodations and restaurants. If a user's question falls outside these topics, a standardized response indicates that assistance is unavailable. Additionally, technical setup involves creating an alias for versioning, uploading specific CSV files, configuring the Lambda function with Python code, and ensuring proper permissions and timeout settings for optimal function performance.
Your job is to assess the user's question and pass it to the appropriate agent for further handling. If the user's question is not about accommodation or restaurants, reply back with I can't help you, I can only help with restaurants and accommodation.
Give the agent a name. I choose "bedrock-supervisor-agent." Enable multi-agent collaboration, then click on Create.
Upload the CSV files named "airbnb.csv & hotel.csv" into the same S3 bucket we created earlier, and these two files you will find in the GitHub repo.
Now go to your Lambda function and paste this Python code as given below...in the Layers section, click on Add a layer. Select Pandas in AWS layers and the default version of the Pandas library.
Collection
[
|
...
]