Fidelity Home Group - AI Chatbot & Workflow Automation
An AI-powered chatbot integration project designed to connect customer communication channels with an existing company knowledge base.
1. About the Company
Fidelity Home Group is a Florida-based mortgage company helping individuals and families navigate the home financing process.
The company provides a range of mortgage solutions for home purchases, refinancing, and real estate investments. It combines digital tools with personalized guidance from mortgage professionals to create a more streamlined home-buying experience.
Website: https://www.fidelityhomegroup.com/
2. Project Overview
An AI-powered chatbot integration and workflow automation project designed to connect customer communication channels with the company's knowledge base and customer-specific context.
The solution integrates multiple communication platforms and uses n8n to orchestrate workflows, retrieve relevant information, and process incoming messages.
By combining a shared company knowledge base with contact-specific conversation context, the system enables AI-generated responses that are informed by both company information and previous customer interactions.
3. My Role
Automation & Integration Developer
Worked as part of a three-developer team, contributing to solution ideas, experimentation, workflow development, and integrations across existing systems.
My work focused on building automation workflows, connecting third-party platforms, and integrating AI capabilities with the company's existing communication and data infrastructure.
4. Key Features
- Multi-Channel Communication Integration: Connect Outlook email through HubSpot, HubSpot Webchat, and Chatwoot SMS to centralized AI-powered workflows.
- Workflow Orchestration: Use n8n as the central automation layer to coordinate incoming messages, data retrieval, AI processing, and system integrations.
- Shared Knowledge Base: Use Pinecone to store and retrieve company-wide knowledge, including legal and business reference documents and chatbot response guidelines.
- Knowledge Synchronization: Integrate Google Drive as a knowledge source and synchronize updates to keep the knowledge base aligned with the latest documents.
- Contact-Specific Context Management: Use Supabase Vector and metadata to maintain contact-related information and conversation context.
- Unified Contact Context: Associate incoming messages across supported communication channels with a unique contact ID, allowing the system to retrieve relevant customer history across channels.
- RAG-Based Response Generation: Retrieve relevant company knowledge from Pinecone and combine it with contact-specific context from Supabase Vector to provide contextual information to GPT-4o for response generation.
5. Technology Stack
- Workflow Automation: n8n
- CRM & Communication: HubSpot
- Messaging Integration: Chatwoot
- Knowledge Vector Database: Pinecone
- Contact Context & Vector Storage: Supabase Vector
- Knowledge Source: Google Drive
- AI Model: OpenAI GPT-4o
Project Gallery
Some project images may not be available because I no longer have access to the original materials or do not have permission to share them publicly.