Skip to main content

Report Chatting Overview

Author: Ahmet Selman Güclü

Introduction​

The Report Chatting feature is an AI-powered functionality that allows users to interact with and query specific country reports directly through a chat interface. This feature is accessible through the Download Portal, where users can find country-specific reports and engage with their content through a dedicated chat button.

What is Report Chatting?​

Report Chatting enables users to:

  • Ask questions about specific country reports.
  • Get AI-generated responses based solely on the report's content.
  • Access insights and information from reports in a conversational manner.
  • Receive accurate, context-aware answers about the report's data.

How to Access​

The feature can be accessed through the Download Portal in two steps:

  1. Navigate to the "Country Reports" section in the Download Portal.
  2. Click the chat button (💬) next to any country's report to start a conversation.

The chat interface will open with a greeting message specifically mentioning the selected country, indicating that the conversation will be focused on that country's report.

Download Portal

Opening Report Chat

Report Response

Technical Overview​

The Report Chatting system is built on a sophisticated architecture that combines several key components. Here's a high-level overview of how the system works:

How It Works​

The system operates through two main workflows:

  1. Query Processing Flow (Left Side)

    • Users interact with the chat interface in the frontend
    • Their questions are sent to the backend server via API requests
    • The AI model processes these questions and searches the database for relevant information
    • Using the retrieved context, the AI generates accurate responses
    • These responses are sent back to the user through the frontend interface

  2. Data Processing Flow (Right Side)

    • PDF reports are processed through a specialized pipeline
    • The pipeline extracts, structures, and organizes the report data
    • This processed information is stored in the MongoDB database
    • The database serves as a knowledge base for the AI model to reference when answering questions

What's Next​

In the following documentation sections, we'll dive deeper into:

  • The integration between frontend and backend components
  • How report data is processed and structured
  • The detailed mechanics of the report chatting functionality