• My questions are:

    1. Is passive real-time monitoring of active AI conversations on the roadmap?
    2. Is forced human takeover of an active AI conversation planned?
    3. Is an AI fallback/timeout mechanism planned when a human handoff is requested but nobody responds? like AI can comeback for a recue after 30-45 secs of no response from human end rather then customer waiting endlessly, Is there any existing way to accomplish this using MxChat actions, integrations, hooks, webhooks or APIs?
    4. Could MxChat eventually support a workflow where AI and human support dynamically hand control to each other while preserving the complete conversation history?
    5. I understand some of these capabilities may not currently exist. I’m mainly interested in knowing what is already possible, what can be achieved through existing integrations, and what is on the roadmap.
      Thanks — I really like the current MxChat/WooCommerce architecture and am trying to plan the support workflow properly before going live.
    • This topic was modified 6 days, 15 hours ago by ajitsingh83.
Viewing 1 replies (of 1 total)
  • Plugin Support m4xw3ll

    (@m4xw3ll)

    Thanks for reaching out and here are some answers!

    1. Passive real-time monitoring — Not currently available; the Transcripts dashboard is a manual-refresh view. If you refresh while someone is chatting you can see their messages come in.
    2. Forced human takeover — Not currently available; handoffs are always initiated from the conversation via triggered action or tool call, not seized by an agent. I have thought about adding this and it’s technically possible, so will consider.
    3. AI fallback after 30–45s of no agent response — Not possible today, but it’s something we expect to ship in an update within the next week or so. Today’s closest control is the live agent availability schedule, which stops the bot offering a human outside your staffed hours and keeps the AI answering instead.
    4. Dynamic AI ↔ human handoff with full history — This already works today. The AI hands off to Slack or Telegram with the conversation context plus the visitor’s name and email, your agent replies straight into the chat widget, and they hand control back with !endchat in Slack or #close in Telegram – multiple round-trips per session are supported, and the transcript stays one continuous conversation throughout.

    Thanks!

Viewing 1 replies (of 1 total)

You must be logged in to reply to this topic.