Dokani.io

Say Goodbye to Return Headaches. Automate refunds and Returns like Amazon — without the Amazon-sized budget.

Comment 3

Dokani.io - AI-Powered Customer Service Platform

Overview

Dokani.io revolutionizes customer service with AI-powered agents, intelligent triage, and comprehensive analytics. Built for e-commerce businesses struggling with overwhelming customer demands during crisis situations.

Inspiration

The inspiration for Dokani.io came from an experience that highlighted a critical gap in the e-commerce ecosystem. In mid-June, several of my close friends who run e-commerce and f-commerce businesses faced devastating losses due to cargo company mishandling, resulting in thousands of dollars in product damage. However, what truly broke their spirits wasn't just the financial loss—it was the overwhelming flood of return and refund requests that followed. These entrepreneurs, running one-person operations, found themselves completely overwhelmed by hundreds of customer grievances they simply couldn't handle efficiently. The frustration of not being able to properly address customer concerns led to long-term business damage and lost customer relationships. Witnessing their struggle, I realized there had to be a better way to manage customer service at scale. Initially, I had planned a different project for this hackathon, but when I heard their stories, the vision for Dokani.io crystallized immediately. The Bolt hackathon and its hosts provided the perfect platform and tools to solve this longstanding problem that affects countless small business owners worldwide.

What it does

  • Dokani.io is an AI-powered customer service platform that revolutionizes how businesses handle customer interactions, particularly during crisis situations like product recalls or shipping issues. Our platform combines cutting-edge AI technology with human-like communication to provide:
  • Intelligent Customer Triage: Automatically categorizes and prioritizes customer requests based on urgency and complexity
  • AI-Powered Voice & Video Calls: Enables businesses to conduct personalized customer service calls using AI agents that sound and look like real people
  • Smart Return Management: Streamlines the return process with automated documentation, evidence collection, and decision-making
  • Real-time Analytics: Provides insights into customer satisfaction, response times, and business impact
  • Multi-channel Support: Integrates with popular platforms like WhatsApp, Facebook Messenger, and Shopify The platform is designed to scale from small one-person operations to large enterprises, ensuring no customer concern goes unaddressed.

Core Features

AI Agents & MCP Servers

  • Customer Service Agent: Handles inquiries with natural language understanding
  • Triage Agent: Categorizes and prioritizes requests with fraud detection
  • Policy Agent: Enforces business policies with A/B testing capabilities
  • Conversation MCP Server: Manages multi-turn conversations and context
  • Request MCP Server: Processes returns, refunds, and order updates
  • Policy MCP Server: Version control and compliance monitoring

Advanced Capabilities

  • Voice & Video AI: Human-like AI agents using ElevenLabs and Tavus
  • Real-time Analytics: Live metrics and predictive insights
  • Multi-channel Support: WhatsApp, Facebook Messenger, Shopify integration
  • Risk Assessment: AI-powered fraud detection and customer profiling
  • Policy Management: Version-controlled policies with impact analysis

How we built it

We built Dokani.io using a modern, scalable tech stack that prioritizes performance and user experience:

  • Frontend: Next.js 14 with TypeScript and Tailwind CSS for a responsive, modern interface
  • Backend: Supabase for real-time database operations, authentication, and file storage
  • AI Integration: OpenAI GPT for natural language processing, ElevenLabs for voice synthesis, and Tavus for video AI
  • Real-time Communication: WebSocket connections and WebRTC for voice/video calls
  • Deployment: Netlify for hosting with Entri for development workflow

The architecture follows a microservices approach with Edge Functions handling complex AI operations, ensuring low latency and high reliability. We implemented a sophisticated file upload system for evidence collection and built a comprehensive analytics dashboard to track business metrics.

Challenges we ran into

The biggest challenge was integrating multiple AI services and ensuring seamless real-time communication. Coordinating between OpenAI's language models, ElevenLabs' voice synthesis, and Tavus' video AI required careful orchestration and error handling. Another significant challenge was building a robust file upload system that could handle various file types while maintaining security and performance. We had to implement proper authentication, file validation, and storage management across different cloud services. The real-time nature of the platform also presented challenges with WebSocket connections, ensuring messages were delivered reliably and handling connection drops gracefully. We also faced the complexity of managing multiple conversation threads and maintaining context across different communication channels.

Accomplishments that we're proud of

We're incredibly proud of creating a platform that genuinely solves a real-world problem affecting countless small business owners. The integration of multiple AI services into a cohesive, user-friendly platform is a significant technical achievement. We successfully built a system that can handle voice and video calls with AI agents that sound and look remarkably human-like, which opens up new possibilities for customer service automation. The real-time analytics dashboard provides valuable insights that can help businesses make data-driven decisions. Most importantly, we created a platform that scales from small operations to large enterprises, democratizing access to sophisticated customer service tools that were previously only available to big corporations.

What we learned

This project taught us the importance of building with real user problems in mind. The initial inspiration from our friends' struggles guided every technical decision we made, ensuring the platform would actually solve their pain points. We overcame the challange of integrating multiple AI services and how to create seamless user experiences across different communication channels. The complexity of real-time systems taught us the importance of robust error handling and graceful degradation. We also discovered the power of modern development tools like Bolt, which accelerated our frontend development significantly, though we learned that backend integration and edge functions still require careful attention and sometimes manual intervention.

Future Roadmap

Immediate (3-6 months)

  • Multi-language support and emotional intelligence
  • WooCommerce, Magento, BigCommerce integrations
  • Mobile apps for iOS and Android
  • Advanced predictive analytics

Medium-term (6-12 months)

  • White-label enterprise solutions
  • Computer vision for damage assessment
  • Global expansion with regional customization
  • Advanced security and compliance features

Long-term (1-2 years)

  • Healthcare, education, and government sector expansion
  • Autonomous customer service operations
  • Strategic partnerships with logistics and payment providers
  • AI ethics framework and responsible AI development

Built With

Share this project:

Updates