Project Description – BizMate
Overview: BizMate is an AI-powered business assistant built to help small and medium-sized enterprises (SMEs) manage their operations through Telegram. It handles inventory, customer orders, supplier coordination, analytics, and customer service—all integrated into a seamless messaging experience.
Key Features & Functionality:
- Business Registration & Bot Integration: Businesses can register and automatically create a Telegram bot for customer interactions.
- Inventory Management: Add, update, and monitor stock levels, track expiry dates, and receive low-stock alerts.
- Order Management: Handles both customer and supplier orders with real-time updates and fulfillment tracking.
- Customer Service AI Agent: Provides personalized responses, tracks orders, manages discounts within pricing limits, and prevents fraudulent use.
- Analytics Agent: Auto-generates business reports, visualizations (bar, pie, line, scatter), and insights using SQL + charts.
- Visit Logging & History Tracking: Every customer visit is logged for analytics and personalized service.
- Security: Session control, user verification, and self-service abuse prevention.
Technologies Used:
- Language & Frameworks: Python, SQLAlchemy, AsyncTeleBot (Telegram Bot API), Google’s Agent Development Kit (ADK)
- AI Model: Gemini 2.0 Flash
- Database: MySQL (hosted on Google Cloud), structured with well-related normalized tables
- Hosting: Google Compute Engine (GCE)
- Environment & Secrets Management:
python-dotenv - Visualizations: Auto-generated using Matplotlib & Google ADK visualization agents
Data Sources:
- All business, customer, product, and transaction data are stored in the MySQL database.
- AI agents interact with real-time data via custom SQL tools and temporary views, with no external datasets used.
Inspiration:
The project was inspired by the operational struggles faced by many small business owners in Nigeria who lack affordable, automated tools to manage daily activities. Telegram’s ubiquity and ease of access made it the perfect entry point for such businesses.
What We Learned:
- Building intelligent agents that coordinate with each other through Google ADK.
- Managing a complex relational schema and syncing it with AI workflows.
- Implementing real-time Telegram interactions with security, chat memory, and fallback logic.
- Designing intuitive interfaces for users with minimal tech skills.
Challenges Faced:
- Handling timing issues with Telegram callbacks and database writes.
- Ensuring AI agents remain grounded in the actual database logic without hallucinating responses.
- Designing a schema that supports flexibility.
- Debugging ADK agent interactions and integrating visualization outputs dynamically.
Outcomes & Impact:
BizMate has demonstrated that low-cost AI-powered tools can be made accessible to everyday businesses. It simplifies complex backend operations and turns a Telegram chat into a full business assistant. With further training and scaling, BizMate can become a go-to digital tool for SMEs in emerging markets.
Log in or sign up for Devpost to join the conversation.