> ## Documentation Index
> Fetch the complete documentation index at: https://docs.weeziq.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Human Handoff

WeezIQ's AI chatbot knows its limits. When a customer is frustrated, angry, or explicitly asks to speak with a person, the chatbot does not keep trying to answer on its own. Instead, it triggers a **Human Handoff** — collecting the customer's details and immediately notifying your support team via email, while sending the customer a confirmation that help is on the way.

<Info>
  Human Handoff is automatic and built into every WeezIQ chatbot. No configuration is required to enable it.
</Info>

***

## How It Works

<Steps>
  <Step title="The AI Detects the Need for Escalation">
    The AI monitors the tone and content of every message. If a customer is:

    * Visibly angry or highly frustrated
    * Making a serious, unresolved complaint
    * Explicitly asking to speak with a human or a manager

    ...the AI will decide to escalate the conversation rather than continue generating automated responses.
  </Step>

  <Step title="The AI Collects the Customer's Email">
    Before escalating, the AI needs a way for your team to follow up. If the customer's email is not already known from a lead capture form, the chatbot will politely ask:

    > *"I'd like to connect you with a member of our team. Could you share your email address so they can reach you directly?"*

    The AI will not proceed with the handoff until it has obtained the customer's email.
  </Step>

  <Step title="Two Emails Are Sent Simultaneously">
    Once the customer's email, name, and a brief issue summary are collected, the system fires two emails at the same time:

    <Tabs>
      <Tab title="Staff Notification Email">
        Your support team receives an **urgent alert email** containing:

        * Customer name and email address
        * A summary of the issue the customer described
        * The exact date and time of the escalation (UTC)
        * The domain (chatbot) where the conversation occurred
        * A **direct link to the conversation** in your WeezIQ inbox

        The subject line reads: `🚨 Urgent: Customer "[Name]" needs human assistance`

        This email is sent from `notifications@weeziq.com` on behalf of WeezIQ.
      </Tab>

      <Tab title="Customer Confirmation Email">
        The customer receives a **reassurance email** that their request has been received:

        * Addressed by name
        * Confirms their concern has been forwarded to the support team
        * States that a team member will reply to their email address shortly
        * Signed with your company or organization name

        The subject line reads: `We've received your request — [Company Name]`
      </Tab>
    </Tabs>
  </Step>

  <Step title="The AI Informs the Customer In Chat">
    After sending the emails, the AI immediately responds in the chat window to let the customer know what happened:

    > *"Your concern has been forwarded to our support team. Someone will contact you at \[email] shortly."*

    <Note>
      The AI will only escalate once per conversation. If the customer asks again, it will confirm that the team has already been notified and no duplicate alert will be sent.
    </Note>
  </Step>

  <Step title="10 AI Credits Are Consumed">
    Each successful human handoff deducts **10 AI credits** from your account — regardless of how many team members received the notification email.

    <Warning>
      **10 credits per escalation.** This is in addition to the 1 credit consumed per regular AI message earlier in the conversation. Monitor your credit usage on the Usage page if you have high escalation volumes.
    </Warning>
  </Step>
</Steps>

***

## Who Receives the Staff Email?

WeezIQ uses a smart priority system to decide which members of your team receive the escalation notification. This is based on your organization's member roles.

| Priority                        | Recipients            | CC'd                |
| ------------------------------- | --------------------- | ------------------- |
| **1. Members exist**            | All Members           | All Admins + Owners |
| **2. No members, admins exist** | All Admins            | All Owners          |
| **3. Only owners**              | All Owners            | —                   |
| **4. No organization**          | Domain owner directly | —                   |

<Info>
  This routing means support staff (Members) are always the primary recipients, with managers and owners kept in the loop via CC. If your organization has no role-based members, the domain account holder receives the alert.
</Info>

### Setting Up Your Team for Handoff

To ensure the right people receive escalation alerts:

1. Go to **Settings → Members** in your WeezIQ dashboard.
2. Invite your support agents and assign them the **Member** role.
3. Invite managers and assign them the **Admin** role.
4. Owners receive all escalations as CC by default.

***

## Credit Cost Summary

| Action                   | Credits Consumed |
| ------------------------ | ---------------- |
| Regular AI message       | 1 credit         |
| Human Handoff escalation | **10 credits**   |

<Tip>
  If you notice high escalation rates, review your Knowledge Base. Frequent handoffs often indicate gaps in the chatbot's training data — questions it cannot answer that should be covered by your documentation, FAQ, or Help Desk Q\&A pairs.
</Tip>

***

## Viewing Escalated Conversations

Every escalated conversation is accessible from the **Conversations** page in your WeezIQ dashboard. The staff email includes a **View Conversation** button that links directly to the full chat transcript in your conversation inbox, so you have full context before reaching out to the customer.

***

## Frequently Asked Questions

<AccordionGroup>
  <Accordion title="Can I disable Human Handoff?">
    Human Handoff is a core feature of every WeezIQ chatbot and cannot be disabled. It is designed to ensure customers always have a path to real support when the AI has reached its limits.
  </Accordion>

  <Accordion title="What if the customer doesn't provide their email?">
    The AI will politely ask for an email address before escalating. If the customer refuses or the conversation ends before the email is collected, the escalation is not triggered and no credits are consumed.
  </Accordion>

  <Accordion title="What if the escalation email fails to send?">
    If the staff email fails to deliver (for example, due to an email provider issue), the system logs the error internally. The customer confirmation email is sent independently — so the customer may still receive their confirmation even if the staff alert fails.

    Credits are only consumed if at least one email is delivered successfully.
  </Accordion>

  <Accordion title="Will the same customer trigger multiple escalations?">
    No. The AI is instructed to escalate only once per conversation session. If the customer continues to ask within the same session, the AI will confirm that the support team has already been notified.
  </Accordion>

  <Accordion title="Can I customize the escalation email templates?">
    The email templates are currently standardized by WeezIQ. The staff email includes your domain name and the customer's information. The customer email is signed with your company or organization name as set in your WeezIQ account.
  </Accordion>
</AccordionGroup>