Build "Chatbot & AI" E-Commerce Platform
Budget: $3,000 – $5,000 USD
Detailed Job Description: Senior Dev Team for "Chatbot & AI" SaaS Platform (Phase 1)
1. Project Vision: The "AI-First" E-commerce OS
We are not just building another chatbot; we are developing Chatbot & AI, a multi-tenant, AI-centric SaaS platform designed to act as an "Operating System" for e-commerce businesses. The core philosophy is to be ERP, CRM, and BPM agnostic while providing a deeply integrated customer service experience.
2. Phase 1 (MVP) Scope & Deliverables
The selected team will execute the first phase of our roadmap, focusing on a robust, multi-tenant architecture and omnichannel communication.
A. Omnichannel Messaging Gateway
• Channels: Full integration with WhatsApp (via Gupshup API), Facebook, Instagram, and a custom WebChat.
• Session Management: Implementation of "Session Continuity" across channels. If a user starts on WebChat and moves to WhatsApp, the context must persist.
• Proactive Notifications: The system must support outbound messaging (status updates, abandoned carts) triggered by e-commerce events.
B. Advanced AI Intelligence (The Brain)
• Training Data: We have a pre-curated set of 1500 Q&A pairs to fine-tune the intent recognition and RAG system.
• Sentiment Analysis: Real-time analysis of customer mood. "Angry" customers must be automatically prioritized for human takeover.
• Handover Summary: When a human agent takes over a chat, the AI must provide a 3-sentence summary of the previous conversation so the agent doesn't waste time reading the history.
C. Multi-Tenant Admin & Dashboard
• SaaS Architecture: A centralized web-based admin portal for multiple companies to manage their own instances.
• Agent Interface: Real-time conversation monitoring and the ability to "intervene" and take over chats via SignalR. Users should be able to connect to a live customer representative at any time. Conversations should be automatically assigned to live customer representatives. User levels and authorizations should be created, such as manager, representative, etc.
• Tenant Isolation: Strict data encapsulation at the database level (Cosmos DB) to ensure no data leakage between clients.
D. Agnostic Integration Layer
• Connectors: Implementation of an "Agnostic Connector Layer" using Azure Functions to communicate with TSoft and Shopify APIs.
• Live Operations: Tracking orders by number, stock inquiry by SKU, and automated return/exchange requests.
________________________________________
3. Technical Requirements (Must-Have)
Category Requirement
Backend .NET 8+ Core (Clean Architecture, CQRS)
Cloud Azure (App Service, Key Vault, Service Bus, Functions)
AI Azure OpenAI (GPT-4o/o1), Turkish NLP Expertise
Database Azure Cosmos DB (Optimized Partitioning for Multi-tenancy)
Security PII Masking (Masking TCKN, Phone, etc. before sending to LLM)
________________________________________
4.
Project Milestones
1. Week 1: Foundation, Multi-tenant DB design, and Gupshup Webhook Ingress.
2. Week 2-3: AI Core (1500 QA set integration), Sentiment Analysis, and FAQ RAG.
3. Week 4-5: E-commerce Connectors (TSoft/Shopify), Admin Dashboard, and Handover Logic.
________________________________________
5. Special Instructions for Applicants
• Code Ownership: All source code and IP will belong to us upon final payment.
• Documentation: All technical documentation must be in English.
• Communication: Daily progress updates via professional project management tools.
1. Project Vision: The "AI-First" E-commerce OS
We are not just building another chatbot; we are developing Chatbot & AI, a multi-tenant, AI-centric SaaS platform designed to act as an "Operating System" for e-commerce businesses. The core philosophy is to be ERP, CRM, and BPM agnostic while providing a deeply integrated customer service experience.
2. Phase 1 (MVP) Scope & Deliverables
The selected team will execute the first phase of our roadmap, focusing on a robust, multi-tenant architecture and omnichannel communication.
A. Omnichannel Messaging Gateway
• Channels: Full integration with WhatsApp (via Gupshup API), Facebook, Instagram, and a custom WebChat.
• Session Management: Implementation of "Session Continuity" across channels. If a user starts on WebChat and moves to WhatsApp, the context must persist.
• Proactive Notifications: The system must support outbound messaging (status updates, abandoned carts) triggered by e-commerce events.
B. Advanced AI Intelligence (The Brain)
• Training Data: We have a pre-curated set of 1500 Q&A pairs to fine-tune the intent recognition and RAG system.
• Sentiment Analysis: Real-time analysis of customer mood. "Angry" customers must be automatically prioritized for human takeover.
• Handover Summary: When a human agent takes over a chat, the AI must provide a 3-sentence summary of the previous conversation so the agent doesn't waste time reading the history.
C. Multi-Tenant Admin & Dashboard
• SaaS Architecture: A centralized web-based admin portal for multiple companies to manage their own instances.
• Agent Interface: Real-time conversation monitoring and the ability to "intervene" and take over chats via SignalR. Users should be able to connect to a live customer representative at any time. Conversations should be automatically assigned to live customer representatives. User levels and authorizations should be created, such as manager, representative, etc.
• Tenant Isolation: Strict data encapsulation at the database level (Cosmos DB) to ensure no data leakage between clients.
D. Agnostic Integration Layer
• Connectors: Implementation of an "Agnostic Connector Layer" using Azure Functions to communicate with TSoft and Shopify APIs.
• Live Operations: Tracking orders by number, stock inquiry by SKU, and automated return/exchange requests.
________________________________________
3. Technical Requirements (Must-Have)
Category Requirement
Backend .NET 8+ Core (Clean Architecture, CQRS)
Cloud Azure (App Service, Key Vault, Service Bus, Functions)
AI Azure OpenAI (GPT-4o/o1), Turkish NLP Expertise
Database Azure Cosmos DB (Optimized Partitioning for Multi-tenancy)
Security PII Masking (Masking TCKN, Phone, etc. before sending to LLM)
________________________________________
4.
Project Milestones
1. Week 1: Foundation, Multi-tenant DB design, and Gupshup Webhook Ingress.
2. Week 2-3: AI Core (1500 QA set integration), Sentiment Analysis, and FAQ RAG.
3. Week 4-5: E-commerce Connectors (TSoft/Shopify), Admin Dashboard, and Handover Logic.
________________________________________
5. Special Instructions for Applicants
• Code Ownership: All source code and IP will belong to us upon final payment.
• Documentation: All technical documentation must be in English.
• Communication: Daily progress updates via professional project management tools.