This presentation explores how we revolutionized our cloud service help desk operations using Amazon Bedrock and RAG (Retrieval-Augmented Generation) technology.

We'll discuss the challenges faced by our help desk, including time-consuming document searches and unnecessary ticket escalations to managers.

Our solution: a serverless RAG Chatbot (Web Application) built on AWS, leveraging Amazon Bedrock and Knowledge Base feature.

This system integrates various internal documents, service specifications, past user interactions to provide context-aware responses.

We'll talk about the technical aspects of our implementation, including:

- Maintaining conversation history to enable contextual understanding

- Building a data pipeline for personal information masking

- Enhancing RAG response accuracy

We'll also address business-related challenges following implementation:

- Driving user adoption and change management within the help desk team

- Setting business goals , KPIs and metrics

Key takeaways include:

- The ease and speed of developing serverless RAG chatbots with AWS

- Best practices for applying RAG in help desk operations

- Techniques for improving RAG response accuracy