Build N8N Customer Service Agents
Budget: $750 – $1,500 AUD
I already run n8n for internal automations and now I want to push it further by wiring up AI-powered customer service agents. The goal is to design flows that let a single agent work seamlessly across our website chat widget, email inbox, and social-media DMs, using the same knowledge base and conversation memory everywhere.
Here is what I need:
• A set of n8n workflows that spin up an automated customer-service agent, connect it to our chat, email, and social channels, and switch context correctly between them.
• Skill modules for the agent to 1) answer FAQs, 2) process and confirm orders, and 3) identify, categorise, and resolve complaints—or hand them off to a human when confidence drops.
• Clear instructions on how to add new intents, update training data, and maintain the flows.
Please build and test everything inside n8n (self-hosted) using whichever LLM, vector store, or plug-ins you feel are best; just keep the stack open-source or easily swappable. I’ll consider this complete once I can see a live demo handling the three task categories above on all three channels, along with exportable n8n JSON so I can import it straight into production.
Here is what I need:
• A set of n8n workflows that spin up an automated customer-service agent, connect it to our chat, email, and social channels, and switch context correctly between them.
• Skill modules for the agent to 1) answer FAQs, 2) process and confirm orders, and 3) identify, categorise, and resolve complaints—or hand them off to a human when confidence drops.
• Clear instructions on how to add new intents, update training data, and maintain the flows.
Please build and test everything inside n8n (self-hosted) using whichever LLM, vector store, or plug-ins you feel are best; just keep the stack open-source or easily swappable. I’ll consider this complete once I can see a live demo handling the three task categories above on all three channels, along with exportable n8n JSON so I can import it straight into production.