Omnichannel FAQ Support Chatbot
Budget: $10 – $30 USD
The task is to build and deploy an AI-driven chatbot that will shoulder the bulk of my customer support traffic. It must live simultaneously on the website, our iOS/Android apps, and the main social channels so users get a single, seamless support experience no matter where they reach out.
Core requirements
• Real-time live-chat handoff: when the bot detects uncertainty or a user asks for a human, the conversation should slip smoothly into our existing live-chat queue (currently Intercom, but a lightweight webhook approach is fine).
• FAQ intelligence: I will supply a structured FAQ set; the bot needs to recognise paraphrased versions of those questions and serve the correct answer instantly. Updating or expanding this knowledge base should be possible through a simple admin panel or CSV upload.
• Natural language processing: intent detection, entity extraction, typo tolerance, and multilingual readiness are expected so users can phrase issues however they like. Leveraging Dialogflow, Rasa, Botpress, or an LLM stack such as OpenAI is acceptable—pick the stack you’re fastest with, as long as the model can run cost-effectively in production.
• Cross-platform deployment: embed script for the website, mobile SDK hooks for our native apps, and Messenger/WhatsApp (or other social API) connectors. Brand colours and tone guidelines will be provided for a unified look and feel.
• Usage analytics: dashboard that shows volume, resolution rate, fallback frequency, and most-asked questions so the knowledge base can keep improving.
Deliverables
1. Working chatbot codebase with documented setup instructions
2. Clean UI widgets/snippets for web, iOS/Android, and social channels
3. Admin panel or import script for FAQ management
4. Short handoff guide covering maintenance, retraining, and escalation logic
Once the bot passes functional tests—90 %+ correct FAQ match rate on a supplied test sheet, flawless live-chat escalation, and deployment on all three platform types—the project is complete.
Core requirements
• Real-time live-chat handoff: when the bot detects uncertainty or a user asks for a human, the conversation should slip smoothly into our existing live-chat queue (currently Intercom, but a lightweight webhook approach is fine).
• FAQ intelligence: I will supply a structured FAQ set; the bot needs to recognise paraphrased versions of those questions and serve the correct answer instantly. Updating or expanding this knowledge base should be possible through a simple admin panel or CSV upload.
• Natural language processing: intent detection, entity extraction, typo tolerance, and multilingual readiness are expected so users can phrase issues however they like. Leveraging Dialogflow, Rasa, Botpress, or an LLM stack such as OpenAI is acceptable—pick the stack you’re fastest with, as long as the model can run cost-effectively in production.
• Cross-platform deployment: embed script for the website, mobile SDK hooks for our native apps, and Messenger/WhatsApp (or other social API) connectors. Brand colours and tone guidelines will be provided for a unified look and feel.
• Usage analytics: dashboard that shows volume, resolution rate, fallback frequency, and most-asked questions so the knowledge base can keep improving.
Deliverables
1. Working chatbot codebase with documented setup instructions
2. Clean UI widgets/snippets for web, iOS/Android, and social channels
3. Admin panel or import script for FAQ management
4. Short handoff guide covering maintenance, retraining, and escalation logic
Once the bot passes functional tests—90 %+ correct FAQ match rate on a supplied test sheet, flawless live-chat escalation, and deployment on all three platform types—the project is complete.