support_analyser

The Support Analysis Tool is an AI-driven solution designed to automate the analysis of support tickets, enhancing the efficiency and responsiveness of Global Technical Support. Utilizing agentic AI and RAG (Retrieval-Augmented Generation) methodologies, it processes support tickets in batches, analyzes them using a knowledge base (specifically a document repository), and generates the following insights for each ticket:

• Hierarchical Categorization: Relevant product areas and sub-categories of the issue.
• Root cause: underlying cause of the issue if the following format:
  • Primary_Cause_Type
  • Cause_Justification
  • Documentation issue/UI/UX issue/pricing issue/other
  • Details, if it is an other product issue.
• Recommendation: Tailored recommendation to address the issue based on the knowladge base.

Pre-requisites

• A .csv file with support tickets
• A folder with the knowledge articles in .mdx format. (Utilized our public documentation articles for this purpose).
• Your NR API key
• Your github PAT
• Your pine cone API key for vector embedding storage

Installation Instructions

1. Clone the repository to your local machine.

2. On your local machine, open the cloned folder in a code editor.

3. (Optional) To include more knowledge articles beyond the provided public documentation, upload your folder containing .mdx format articles to the support_analyser repository.

4. In the terminal, perform the following steps:

   a. Create a virtual environment:

   ```
   python3 -m venv venv
   source venv/bin/activate
   
   ```
   

   b. Install dependencies:

   ```
   pip install -r requirements.txt
   
   ```
   

   c. Inside your repo, create a .env file, and enter the following:

    ```
    NC_KEY=<Your NR API key>
    TEAM_NAME=<Your team name>
    GITHUB_TOKEN=<Your github PAT>
    PINECONE_API_KEY=<Your pine cone API key for vector embedding storage>
    PINECONE_INDEX=ai-hackathon-index
   
    ```
   

   d. Run the following command to vectorise .mdx files in the knowladge base:

   ```
   python3 java-agent-embed.py --index
   
   ```
   

   e. Run the application:

    ```
    streamlit run app.py
   
    ```
   

The Support Ticket Analyzer tool launches in your browser, fully operational on your local host.

How to use

In the Support Analyser tool UI:

• Upload the .csv file containing support tickets.
• Click Analyze. The process may take a few minutes to complete. After a successful analysis, the results will be displayed in a tabular format.
• You can download the analysis results in a .csv file.