De todo en Comidas Criollas y Boricuas
$250-750 USD
Posted 1 minute ago
•
Ends in 6 days
$250-750 USD
Paid on delivery
I need an end-to-end n8n workflow that turns my existing knowledge base into an AI assistant capable of handling Website chat, WhatsApp and Email conversations. The flow should: • Pull answers from two sources—a public help-center website (crawl or API) and a curated Q&A list I will supply in CSV/JSON. • Use an LLM (OpenAI, Claude or similar) to generate responses, then route them through the correct channel connector: web chat widget, WhatsApp Business API and standard SMTP/IMAP for email. • Detect confidence; when the answer is incomplete or low-confidence, automatically pass the full transcript to a live human agent in our help-desk inbox (Freshdesk/Zendesk webhook is fine). The hand-off must be seamless, with context preserved. Acceptance criteria 1. A single n8n workflow file (.json) importable into my self-hosted instance. 2. Working integrations for the three channels, tested with sample conversations. 3. Environment-variable driven credentials so nothing sensitive is hard-coded. 4. Clear README explaining setup, required API keys and how to update the knowledge sources. Once these points are met and verified on my server, the job is complete.
Project ID: 40828033
About the project
57 proposals
Open for bidding
Remote project
Active 3 secs ago
Place your bid
Bid amount Email addressBenefits of bidding on Freelancer
Set your budget and timeframe
It's free to sign up and bid on jobs
57 freelancers are bidding on average $371 USD for this job


Building AI-powered, cross-platform automations is part of my agency’s expanded expertise, and n8n, OpenAI, JavaScript, PHP and chatbot development directly match your requirements. I can create an importable n8n workflow that connects your help-center content and CSV/JSON Q&A knowledge base to an AI assistant serving website chat, WhatsApp Business and email. My experience with CRM-connected workflows and full cross-platform business automation supports the key details of this project: environment-variable-based credentials, channel-specific routing, structured AI responses and seamless escalation. I can also implement confidence-based handoff so incomplete or uncertain answers send the complete conversation context to a Freshdesk or Zendesk webhook, while keeping the workflow maintainable for future knowledge-source updates. I will provide the workflow JSON together with a clear README covering setup, API keys, environment variables and knowledge-base maintenance, with sample conversations used to verify the three channels on your self-hosted n8n instance. Choose me to build a reliable assistant for your comidas criollas y boricuas business.


Hi, I'm Kausar. You need an n8n workflow to turn your knowledge base into an AI assistant for chat, WhatsApp, and email interactions. I'd start by mapping the data sources and setting up the API calls to pull answers from your help center and the curated Q&A list. I’ll build the workflow to integrate OpenAI for response generation, ensuring seamless routing through the specified channels. I'll also implement the confidence detection feature to hand off low-confidence queries to a human agent, maintaining context throughout. Before handover, I’ll provide a clear README with setup instructions and environment-variable management. What specific Q&A topics should be prioritized in the workflow?


Hi there, You need an n8n workflow that transforms your knowledge base into a capable AI assistant for chat, WhatsApp, and email. This project aligns well with my experience in API development and chatbot integration. How I would approach it: - Start by connecting to your public help center and the Q&A list to pull data for responses - Integrate an LLM to generate replies and set up routing through web chat, WhatsApp Business API, and email - Implement confidence detection to pass low-confidence responses to a live agent with context preservation - Create the workflow file, set up environment variables for security, and document the process in a clear README What existing help center platform are you using, and how frequently do you update the Q&A list? This information will guide the integration. I look forward to ensuring that this workflow meets your expectations.


The trap here is the LLM answering confidently when your knowledge base is silent; those replies go to your agent, never the customer. Day 3 you get the web chat path live on your instance with your Q&A file to try; by day 7 all three channels and the hand-off are verified on your server, README included. We run our own support assistant on WhatsApp and web, so the hand-off edge cases are familiar. 1. Freshdesk or Zendesk, which one will you actually be using? 2. Is your WhatsApp Business number already verified with a message template approved? Send the help-center URL and we can start today.


Hi there, I see you need an end-to-end n8n workflow for integrating various communication channels with an AI assistant. I can help you create a workflow that pulls answers from your public help center API and your curated CSV/JSON Q&A list. I’ll ensure the responses are generated using an LLM and seamlessly routed through the web chat widget, WhatsApp Business API, and email channels. I’ll also set up the confidence detection feature to manage low-confidence responses by routing these to a live agent in Freshdesk or Zendesk. All integrations will be properly tested and I'll provide a clear README for easy setup with environment-variable driven credentials. Let’s make sure everything meets your acceptance criteria. I look forward to discussing the details further. Regards, Ali


Hi, I am Nitish from Webcast Technology. I’m excited to craft a complete n8n workflow that turns your knowledge base into an AI assistant for web chat, WhatsApp Business and email, pulling answers from both a public help‑center crawl/API and your supplied CSV/JSON Q&A file, routing responses through the appropriate channel, and handing off low‑confidence cases to Freshdesk or Zendesk with full context preserved. In a recent project I built a WooCommerce food‑ordering storefront where I integrated multiple payment gateways, set up real‑time order notifications, and delivered a detailed admin guide so the client could manage menus and promotions without developer help. Quick question — will you be providing API credentials for the WhatsApp Business API and Freshdesk/Zendesk webhook now, or should I set up test credentials for the initial proof of concept? Looking forward to delivering a ready‑to‑import .json workflow, fully tested across all three channels, with environment‑variable driven secrets and a clear README so you can maintain it yourself. Nitish


Hi there, I'm excited about the opportunity to build your n8n-powered AI assistant for De todo en Comidas Criollas y Boricuas. I understand you need a robust workflow that connects your knowledge base to multiple communication channels—web chat, WhatsApp, and email—allowing seamless, AI-driven support with smooth human escalation when needed. With extensive experience in n8n automation, API integration, and conversational AI, I’ve delivered similar projects integrating OpenAI/Claude for natural responses, and have set up multi-channel support flows leveraging WhatsApp Business API, email (SMTP/IMAP), and web chat widgets. I’m also proficient in PHP, JavaScript, and XML, ensuring flexibility when custom scripting or parsing is required. My approach will involve: - Aggregating answers from your help-center site (via crawling or API) and curated Q&A lists, ensuring the LLM has comprehensive context. - Designing a unified n8n workflow that dynamically routes conversations through the appropriate channel connector. - Implementing confidence detection to trigger seamless live-agent handoffs to your helpdesk (Freshdesk/Zendesk) with full transcript preservation. - Setting up secure, environment-based credentials and providing detailed documentation for setup, API key management, and knowledge updates. I’ll ensure the solution is clean, well-documented, and easy for your team to maintain or expand. I look forward to collaborating and helping you deliver responsive, intelligent support to your customers. Best Regards


Hi, I've built several n8n workflows handling multi-channel customer support with knowledge base integration, so I understand the routing logic and confidence thresholds you'll need to get right. I'll design a single workflow that pulls from your help-center crawl and CSV/JSON knowledge sources, passes user input to your chosen LLM, and routes responses through the web chat widget, WhatsApp Business API, and email simultaneously. When confidence is low, the workflow will capture the full conversation context and webhook it directly into Freshdesk or Zendesk for your team to pick up without losing information. All credentials will live in environment variables—nothing hardcoded—and I'll test each channel with sample conversations before handoff to make sure the routing and fallback logic work reliably under real conditions. I'll include a README covering API key setup, knowledge source refresh steps, and how to import the workflow into your self-hosted instance. I can start immediately and stay available for any adjustments or questions once it's running on your server.


Building complete, documented automation systems with Node.js, REST APIs and business-data integrations is my core focus, making me well suited to connect your knowledge base, AI responses and communication channels in n8n. My JavaScript and XML skills support custom workflow logic, structured CSV/JSON knowledge handling and API-based integrations. I also specialize in scalable, easy-to-operate business workflow automation, so I can deliver an importable n8n JSON workflow with environment-driven credentials and clear setup documentation. The design will accommodate website chat, WhatsApp Business API, SMTP/IMAP email, LLM response generation and help-desk handoff while preserving the conversation context for human agents. I will provide a practical, documented automation system that can be tested on your self-hosted server and maintained as your knowledge sources evolve. Choose me to build this AI assistant workflow with the reliability and organization your support operation needs.


Hello, I can help you build this n8n workflow end-to-end so your knowledge base can respond naturally across website chat, WhatsApp, and email while keeping a clean human handoff when needed. I’ve recently built a similar support automation for a service business where we connected a website chat widget, WhatsApp, and email into one workflow, pulled answers from a mixed knowledge base, and routed low-confidence messages to a live agent with full context preserved. The part clients usually appreciate most is when the assistant sounds helpful without making up answers, and that’s exactly how I’d structure this. A few things I’d like to confirm before I start: 1. Which platform will host the live agent inbox — Freshdesk or Zendesk? 2. Do you already have the WhatsApp Business API connected, or should I design it for your provider? My suggestion would be to include a simple confidence threshold and fallback logic so the system stays reliable as your FAQ grows. If you’d like, I can discuss the best setup with you on a quick call. Best regards, Mehreen


I can build the complete self-hosted n8n workflow for your AI assistant across website chat, WhatsApp Business, and email. The implementation will combine your public help center with the curated CSV/JSON Q&A source, use your chosen LLM for grounded response generation, and apply confidence-based escalation. When an answer is uncertain or incomplete, the full conversation context will be forwarded to Freshdesk, Zendesk, or your preferred help-desk webhook without losing transcript details. You will receive: • One importable n8n .json workflow • Tested sample conversations across all three channels • Environment-variable-based configuration with no hard-coded secrets • A clear README covering setup, API credentials, channel configuration, and knowledge-source updates • Verification support on your server, including troubleshooting any deployment-specific issues I’ll structure the workflow for maintainability, predictable error handling, and straightforward future updates to prompts, sources, connectors, and escalation rules. The result will be a practical handoff-ready automation rather than a disconnected proof of concept. I look forward to delivering a thoroughly tested workflow aligned with your acceptance criteria. Best regards


Your requirement for an end-to-end n8n workflow that integrates your knowledge base into an AI assistant presents a significant opportunity to streamline communication across multiple platforms. The goal is to automate responses via web chat, WhatsApp, and email while ensuring a seamless hand-off to human agents when necessary. The immediate challenge lies in developing the n8n workflow to handle this integration intelligently, pulling responses from both your help center and a curated Q&A list. Additionally, ensuring confidence detection for responses is crucial. Plan: 1. Create the n8n workflow file to integrate the specified channels. 2. Pull data from the help center and the curated Q&A. 3. Implement confidence detection to enable human agent hand-off when needed. 4. Develop a README for setup and API key configuration. What kind of existing tools or APIs are you currently using for the help center and communication channels? I suggest we start with a detailed overview of your existing systems to better align the workflow. If this aligns with your needs, I can immediately draft a detailed plan and propose the first steps.


Hello, As a result of a detailed review of your project requirements, I fully understand the scope and expectations for building an n8n-powered AI assistant that handles web chat, WhatsApp, and email with human hand-off. I have experience with n8n, OpenAI integration, API connectors, and multi-channel routing — and I’ve built similar workflows that pull from knowledge bases, apply confidence scoring, and route to agents with full context. I have a couple of quick questions. • Is your help-center content publicly crawlable, or will you provide an API endpoint for it? • For WhatsApp, do you already have a Meta Business account and approved number, or should the workflow be tested with a sandbox? I would be glad to discuss further details and am ready to start immediately. Looking forward to hearing from you. Best regards, Carlos.


SUPER MATCH "ONE AI ASSISTANT ACROSS WEBSITE CHAT, WHATSAPP, AND EMAIL" — The key is keeping answers consistent across all three channels while preserving conversation context when human support is needed. I would build a modular n8n workflow that combines your public help center with the supplied CSV/JSON Q&A data, retrieves relevant information, and uses an LLM to generate grounded responses. Each channel would have its own connector, while shared logic handles confidence checks, fallback responses, and escalation with the full transcript to your help desk. I would deliver the importable workflow JSON, environment-based credentials, channel integration tests, and a README covering setup and knowledge updates. Which help-desk platform and WhatsApp Business API provider do you plan to use?


HOLA, "n8n workflow" - you need one AI assistant that answers consistently across website chat, WhatsApp, and email while knowing when to hand conversations to a human. I'll build the complete workflow around your help-center content and curated CSV/JSON knowledge base, using an LLM to retrieve relevant answers, assess confidence, and route replies through the correct channel. When confidence is low, the full conversation history and customer context will be forwarded to your Freshdesk or Zendesk inbox instead of sending an unreliable answer. I'll keep credentials in environment variables and deliver the importable JSON workflow, tested integrations, and a practical README for setup and future knowledge-base updates. Do you already have WhatsApp Business API access and a Freshdesk or Zendesk account configured? Gracias por su consideración. Espero poder colaborar con usted.


Hello I build n8n workflows with OpenAI, API development, PHP, JavaScript and XML, and turning a knowledge base into a multi-channel chatbot for website chat, WhatsApp and email is exactly my lane. The hard part is the handoff: how should the assistant judge its own confidence, so it escalates only truly incomplete answers to your help desk instead of sending it every question it finds difficult? Message me with your current help-center setup and a sample of your Q&A list, and I will map out the approach or show you a working demo.


I need an n8n workflow that pulls from your knowledge base and feeds OpenAI, not just a basic chatbot wrapper. I would build it with a vector store node for retrieval, a function node to format context, then route responses through your preferred channel. Can start today, working version in 4 days. These numbers are based on the post as written, we will refine both after a quick scope talk. Want me to send a plan?


Hi, Before I touch any code I'd want to see your help-center site structure (is it crawlable or do you have an API), a sample of the CSV/JSON Q&A file, and which helpdesk you're actually using for the handoff, Freshdesk or Zendesk. I'd expect the website crawl to need some cleanup since help docs are rarely formatted consistently, and WhatsApp Business API approval can take longer than the build itself, so that's worth starting early. Once I have those, the build is straightforward: n8n handles the routing, a vector store holds your knowledge base for fast lookups, and the LLM node generates replies with a confidence check before anything gets sent. Low-confidence or incomplete answers get pushed straight to your helpdesk inbox with the full conversation attached, so your team isn't starting cold. I've built this kind of multi-channel handoff before, and keeping context intact across the switch is the part I pay closest attention to. This runs about 2 days once I've seen the knowledge base. Ready to start when you are. Best, Emrah


Your project caught my attention because a unified AI support assistant can save significant time while keeping customer conversations consistent across every channel. I'd build a modular n8n workflow connecting your help centre, curated CSV/JSON knowledge base, and an LLM such as OpenAI or Claude. I'd configure website chat, WhatsApp Business API, and email integrations, with knowledge retrieval, confidence checks, conversation history, and automatic human handoff through Freshdesk or Zendesk. All credentials would be managed through environment variables rather than hard-coded values


The key risk is context loss during hand-off, so I’d design the workflow around a single message-processing entry point that stores every incoming message and generated reply with its confidence score before routing. My first step would be building the fallback path: when confidence drops below your threshold, the n8n workflow immediately forwards the transcript and user metadata to Freshdesk/Zendesk. This matters because seamless hand-off is the hardest acceptance criterion, and having it early makes chat, WhatsApp, and email connectors easy to add later. I’ve built similar retrieval-augmented assistants in n8n using OpenAI and Claude, with webhook-triggered flows pulling from crawled sites and JSON knowledge bases. For the crawl, I’d use the HTTP Request node with a configurable sitemap URL, then store normalized text in a searchable format. Would you like to schedule a quick call to discuss your current n8n setup? Best regards
内容来自 Freelancer News,版权归原作者所有。
阅读原文 ↗