Mark Bot

AI-powered enterprise knowledge assistant integrating Confluence, Google Drive, Microsoft Teams, and vector search technologies to provide intelligent access to organizational knowledge.

Overview

Mark Bot is an intelligent chatbot platform designed to help teams quickly find information across multiple knowledge sources.

The system collects documents from enterprise systems such as Confluence and Google Drive, indexes them into vector databases, and provides natural language answers through conversational interfaces including Microsoft Teams.

By combining semantic search, document embeddings, and chatbot integrations, Mark Bot enables employees to locate information faster and reduce time spent searching across multiple platforms.

---

Features

Knowledge Base Integration

• Confluence document ingestion
• Confluence table extraction
• Google Drive integration
• Automated content synchronization

AI-Powered Search

• Semantic document retrieval
• Vector similarity search
• Context-aware responses
• Knowledge grounding from enterprise content

Chat Interfaces

• Microsoft Teams bot integration
• Conversational question answering
• Enterprise knowledge assistant

Data Storage

• ChromaDB vector database
• Document store
• Table vector store
• Persistent knowledge indexing

Deployment Ready

• Infrastructure templates
• Deployment configurations
• Environment-based configuration
• Production deployment support

---

Architecture

                   +----------------+
                   | Confluence     |
                   +-------+--------+
                           |
                           v
+--------------+    +-------------+    +---------------+
| Google Drive | -> | Collectors  | -> | Document Store|
+--------------+    +-------------+    +---------------+
                           |
                           v
                   +----------------+
                   | Chroma Vector  |
                   | Database       |
                   +-------+--------+
                           |
                           v
                   +----------------+
                   | Retrieval Layer|
                   +-------+--------+
                           |
                           v
                   +----------------+
                   | Mark Bot       |
                   +-------+--------+
                           |
                           v
                   +----------------+
                   | Microsoft Teams|
                   +----------------+

---

Project Structure

mark-bot/
│
├── app.py
├── config.py
│
├── bots/
│   └── Chat bot implementations
│
├── data_collectors/
│   └── Data ingestion services
│
├── confluence_db/
│   └── Confluence document storage
│
├── confluence_table_db/
│   └── Confluence table storage
│
├── google_drive/
│   └── Google Drive integrations
│
├── chroma_db/
│   └── Vector database storage
│
├── table_docstore/
│   └── Document storage
│
├── table_vectorstore/
│   └── Vector embeddings
│
├── appManifest/
│   └── Microsoft Teams app manifest
│
├── deploymentTemplates/
│   └── Deployment resources
│
├── infra/
│   └── Infrastructure configuration
│
├── resources/
│   └── Static resources
│
├── assets/
│   └── Images and assets
│
└── Images/
    └── Documentation screenshots

---

Technology Stack

Backend

• Python

AI & Search

• ChromaDB
• Vector Embeddings
• Semantic Search
• Retrieval-Augmented Generation (RAG)

Integrations

• Microsoft Teams
• Confluence
• Google Drive

Infrastructure

• Deployment Templates
• Infrastructure-as-Code configurations

---

Installation

Clone Repository

git clone https://github.com/Illania/mark-bot.git
cd mark-bot

Create Virtual Environment

python -m venv venv

Activate:

source venv/bin/activate

Windows:

venv\Scripts\activate

Install Dependencies

pip install -r requirements.txt

---

Configuration

Create a .env file:

CONFLUENCE_URL=
CONFLUENCE_USERNAME=
CONFLUENCE_API_TOKEN=

GOOGLE_DRIVE_CREDENTIALS=

TEAMS_APP_ID=
TEAMS_APP_PASSWORD=

CHROMA_DB_PATH=

Update configuration values according to your environment.

---

Running the Application

python app.py

---

Workflow

1. Collect content from Confluence and Google Drive.
2. Generate embeddings for documents.
3. Store embeddings in ChromaDB.
4. User asks a question through Microsoft Teams.
5. Relevant documents are retrieved using semantic search.
6. Mark Bot generates a contextual answer.

---

Use Cases

• Internal knowledge assistant
• Enterprise search
• Confluence Q&A
• Documentation discovery
• Team onboarding support
• Operational knowledge retrieval

---

Future Improvements

• Multi-LLM support
• Slack integration
• SharePoint integration
• Advanced analytics dashboard
• Fine-grained permissions
• Feedback-based ranking

---

License

This project is licensed under the MIT License.

---

Author

Anna Gulich

Software Engineer | Machine Learning Engineer

GitHub: https://github.com/Illania