I've made quite a mess here but basically the system pipeline is: - Summarize conversation history - Summarize and identify new message intent - Route to "Reservation_Agent" / "General_Question_Agent" / "Basic_Communication_Agnet" - Finalize the output and send via Facebook Graph API Other features: - Handle batch messages (when user send multiple message at a time) - Stop answering when a human agent has replied - change user conversation state - RAG of course (room rates and FAQs) - Handle Facebook false requests (sometimes facebook sends one message multiple times but it is quite rare) - Fallback (just backup agent with different chat model in case the main one is disconnected) Now I need to deal with messages in a different language, in this case Vietnamese, which I don't know what's the best approach is. Im trying detect the language first and then route to an identical architecture with the system promptsalso in Vietnamese. Does any one know what the best solution to this is? Please your insights would be really helpful! DM if you need a similar system, we can discuss further

So we need 3 niche to try before sizing down to one specialized niche. I already got a hotel chatbot in development and testing Looking towards 1. HR Automation ( Voice Agent for Interview ) 2. Generic Appointment Setter ( since a lot of services need to make appointments, e.g. salons, medical businesses, dentist, even in the education industry ) Are those ideas valid and still in demand? Really appreciate if you guys can share your opinions and discuss about these, maybe others can benefit from it too!

This might sound really dumb, but I just knew that n8n chat node can be made public with a link or embed on websites. Even have the respond with webhook option, which I think can be used to handle batch messages? Haven't seen any youtube videos talking about that. Has anyone used that in a production environment? Put on any websites? I need more information on this!

Some of you might be overthinking everything right now… 👉 “Am I moving fast enough?” 👉 “Why hasn’t it worked yet?” 👉 “What if I fail?” Let me be real with you: It’s normal to feel this way — but you can’t let those thoughts lead. Every win takes time. Every breakthrough has a painful middle.And sometimes, the only thing you can do is keep going through it. Because guess what? ✅ The pain will pass. ✅ Your effort will compound. ✅ And you’ll look back at this version of yourself with respect — not regret. So today… 📌 Breathe. 📌 Trust the process. 📌 And give yourself some grace. You're not behind. You're just in progress.Hope you have a calm, clear Sunday. Let’s keep building. 🚀

Hi all, I recently saw a job description for an HR assistant doing repetitive tasks like resume filtering, interview scheduling, and document handling. I’m exploring how tools like n8n, Airtable, Gmail, OCR, and AI agents can automate HR workflows (e.g. parsing resumes, sending emails, tracking applicants). I’d love to learn: - What real HR problems can automation meaningfully solve? - What causes these solutions to fail or get ignored in production? If you’ve built or seen HR automations succeed (or fail), I’d love to hear your insights!