Skip to content

Latest commit

 

History

History
78 lines (49 loc) · 3.58 KB

README.md

File metadata and controls

78 lines (49 loc) · 3.58 KB
Raw

Customer Support

This sample is a C# console application that uses Generative AI services like Ollama or Azure OpenAI to:

  • Generate customer support ticket summaries
  • Generate embeddings for product manuals
  • Perform semantic search over product manuals

Project structure

  • ManualIngestor.cs - Ingestion service that extracts data from product manual PDF files, chunks text into smaller segments, generates embeddings, saves them to a JSON file.
  • ProductManualService.cs - Storage service that uses an IVectorStore to save and search product manual embeddings.
  • TicketSummarizer - AI service which uses an AI model to generate summaries of customer support tickets.

Requirements

Quick Start

Get started setting up your environment using either GitHub Codespaces or Dev Containers.

Environment Codespaces DevContainer
GPU (Recommended) N/A Coming Soon
CPU Open in GitHub Codespaces Coming soon

Configuration

For this application, you can either use local language and embedding models with Ollama or models hosted on Azure OpenAI.

Ollama

  1. Download llama3.2 language model

    ollama pull llama3.2

  2. Download all-minilm embedding model

    ollama pull all-minilm

  3. In Program.cs, depending on what you want to use the model for set:

    • useOpenAIChat to false
    • useOpenAIEmbeddings to false

Azure OpenAI

  1. Deploy a chat and embedding model. For more details, see the Azure OpenAI documentation.

    • chat - A language model that supports chat. (i.e. gpt-35-turbo)
    • embedding - An embedding model. (i.e. text-embedding-3-small)

    See the Azure OpenAI documentation for a full list of models available.

    If you use deployment names other than chat and embedding, update them in Program.cs;

  2. In Program.cs, depending on what you want to use the model for set:

    • useOpenAIChat to true
    • useOpenAIEmbeddings to true

  3. Configure environment variables for your endpoint and key. For more details, see the Azure OpenAI documentation.

    • AZURE_OPENAI_ENDPOINT - Your Azure OpenAI endpoint.
    • AZURE_OPENAI_KEY - Your Azure OpenAI Key.

Using managed identity (Recommended)

By using a managed identity from Microsoft Entra, your application can easily access protected Azure OpenAI resources without having to manually provision or rotate any secrets.

  1. Assign Azure OpenAI user role to your managed identity. For more details, see the documentation.
  2. In Program.cs, set useManagedIdentity to true.