An AI customer support agent that knows when to hand off.

Start with the questions you answer every week

Most support inboxes are full of the same questions: where an order is, how returns work, whether you ship to a country, how to reset something. These have written answers already, in your help center, policy pages and internal docs. An AI support agent is useful when it answers those questions from that content and nothing else.

ZynfoAI is trained on the sources you add: website pages, uploaded files and selected Google Docs or Sheets. Together they make up your AI knowledge base. You can show the source link under each answer, so customers and your team can see where a reply came from. If you’re still deciding whether an AI agent fits your support desk at all, our guide to AI chatbots for customer service covers what to hand over and what to keep.

Hand off early, not as a last resort

A good support agent doesn’t try to handle everything. Refund exceptions, billing mistakes and angry customers need a person. In ZynfoAI you decide when that happens: when a customer asks for someone, when a ticket is needed, or through a workflow you build in Automate.

Handoffs respect your working hours. Inside them, the chat is queued in the Inbox with an AI-written summary, routed round-robin, to whoever is least busy, or to the whole team. Outside them, the customer sees your offline message. Either way the teammate sees the full conversation before replying. The human handoff page shows each step.

WhatsApp has its own rules

On WhatsApp, the agent replies inside Meta’s 24-hour customer service window that opens when a customer messages you. After the window closes, businesses can only send pre-approved template messages. That fits support well: the customer writes first, and the agent answers while the window is open. You connect your number through Meta’s Embedded Signup.

Measure what’s missing

After launch, check the dashboard each week. Unanswered Questions shows what triggered your fallback message, Trending Topics shows what people ask about most, and session ratings show how customers felt. Add the missing article, resync the source, and the agent can answer it next time. Our playbook on cutting repeat tickets with an AI support chatbot walks through that weekly review.