Tasky is a complete and innovative AI-powered customer service management platform, designed to serve businesses of all sizes by simplifying communication processes and customer service automation.

With intelligent AI agents, the system automates conversations via WhatsApp using voice recognition and natural language processing. Agents are configurable by category (customer service, sales, marketing) and can access personalized knowledge bases for accurate and contextualized responses.

Tasky offers advanced workflows with visual automations that enable the creation of complex flows with intelligent classification, human approvals, external integrations, and execution of custom tools. The system includes handoff management for intelligent transfer to human agents when necessary, with configurable prioritization and urgency.

The platform integrates with more than 15 popular services, including Google Calendar, Gmail, Google Meet, Google Sheets, Slack, Notion, Stripe, Cal.com, HubSpot, Shopify, PayPal, Intercom, Pipedrive, WooCommerce, and Twilio, allowing agents to execute actions in external systems directly during conversations.

Tasky also offers complete contact management, conversation history, automatically scheduled follow-ups, real-time notifications, and a subscription system with usage control and limits. Built for digital transformation, Tasky represents the future of automated customer service, with scalable architecture and modern technologies.

This platform was built with modern technologies, following best practices for scalability and maintainability.

Backend Architecture

The backend follows SOLID principles and a modular architecture, ensuring clean code and easy extensibility. We implemented rigorous Test-Driven Development (TDD) to maintain high code quality and reliability.

Core Technologies:

TypeScript for static typing and enhanced development safety
PostgreSQL for robust data persistence
Prisma as the ORM for type-safe database access
Redis for efficient caching and performance optimization
AI SDK and OpenAI for LLM-powered text processing
NestJS as the web framework for building RESTful APIs
Docker and Docker Compose for containerization and orchestration
Jest as the testing framework to ensure code quality and coverage

Key Features:

Conversational AI agents for chat and voice calls
Visual workflows with advanced automations and conditional logic
Knowledge base with RAG (Retrieval-Augmented Generation) for contextualized responses
Intelligent handoff system with prioritization and escalation
Integrations with more than 15 platforms via Model Context Protocol (MCP)
Custom tools (HTTP requests, JavaScript execution, webhooks)
Contact management and complete conversation history
Automatically scheduled follow-ups
Real-time notification system (email, in-app and push)

Frontend Architecture

The frontend uses a modern and scalable architecture, facilitating platform evolution over time and team collaboration through clear separation of concerns.

Tech Stack:

React for building interactive user interfaces
Vite for fast development and optimized builds
Tailwind CSS v4 for modern and responsive styling
TanStack Query (React Query) for powerful async data management
Playwright for end-to-end testing

This architectural approach allowed the platform to scale effectively while maintaining code quality and developer productivity.