Reply to Customers in JIVIQ
When a conversation needs human assistance, your team can continue the conversation directly with the visitor through JIVIQ. Sending a reply automatically pauses the AI on that thread so your team can assist without interruption.
Human Replies Summary
Reply to Take Over
The moment a teammate sends a message, JIVIQ automatically pauses the AI on that conversation thread.
Never Talks Over
The AI never talks over a human teammate. The status reads "AI · Paused — you're handling".
Transparent Identity
Visitors see the teammate's initials and first name, while AI responses display a sparkles icon and visible AI tag.
Return to AI
Teammates can manually return conversations to AI, or in Autopilot Workspaces, let conversations hand back to AI after a configured number of quiet hours.
When Should a Team Member Reply?
In JIVIQ, automated AI responses and human assistance work together seamlessly. While your AI handles routine questions grounded in your Knowledge Base, there are specific situations where a human team member should step in and reply directly.
Visitor Requests a Person
When a visitor explicitly asks to speak with a human teammate, the AI acknowledges the request, steps aside, and makes the conversation available for your team to reply.
Knowledge Base Cannot Answer
If an inquiry falls outside your verified documentation or website sources, the AI flags the conversation for your team instead of guessing or leaving the visitor unanswered.
Team Decides Assistance Is Needed
Teammates reviewing conversations can proactively step in at any time to assist with complex inquiries, high-priority discussions, or tailored technical questions.
How Human Replies Work
Replying in JIVIQ does not require complicated multi-step procedures or special takeover toggles. The transition from AI to human handling is built directly around the act of sending a reply.
1. Available to Team
Conversations needing human reply appear in your Chats list, marked with unread counts or escalated indicators.
2. Review Context
Open the thread to review previous messages, the visitor's question, and any available customer information.
3. Teammate Replies
Type your message and send it. The moment you reply, human takeover takes effect and the AI pauses.
4. Direct Continuation
The conversation continues directly between your team and the visitor without AI interruption.
bolt "Reply. That's the Whole Move."
As documented in the official Human Takeover Documentation, you do not need to click a separate confirmation screen or claim lock before answering. Sending your reply immediately takes ownership of the conversation, pausing the AI so that it will never talk over you.
What the Team Can See Before Replying
Before sending a message, teammates have access to essential conversation and visitor context in the JIVIQ admin interface, allowing for personalized and informed answers.
Full Conversation History
You can read the complete timeline of what the visitor asked, what information the AI provided, and what prompted the handoff. Visitors never need to repeat information they already explained.
Customer Identity Information (When Available)
If the visitor submitted an email address during the conversation, it appears directly in the conversation header. Note: Not every conversation contains an email address; anonymous visitors can chat freely without providing identity details.
Visitor Presence Status
Inside a conversation, the header displays presence details: Online, Idle (after roughly 45 quiet seconds), or Last seen (once gone about five minutes). Presence is approximate and indicates whether a live reply is likely to reach the visitor right away.
How Visitors Know Who Is Replying
JIVIQ maintains clear transparency between automated AI responses and human responses. Visitors are never left guessing who they are talking to, and the AI never pretends to be a person.
Automated Assistant Messages
- Carries a ✨ icon
- Displays the assistant's name (e.g., "Assistant")
- Shows a visible AI tag — even if given a human-sounding name
- The AI never claims to be human
Teammate Messages
- Shows the teammate's initials (e.g., "AT") and first name
- Displays no AI tag
- Sender line above the message switches from the AI's name to the teammate's name and initials
- Clear human transparency for the visitor
What Happens to AI After a Human Reply?
The fundamental rule of JIVIQ human takeover is absolute: the AI never talks over a human teammate.
pause_circle Immediate AI Pause on Reply
The moment a teammate sends a message, JIVIQ automatically pauses the AI on that conversation. The AI pill above the reply box updates to read: "AI · Paused — you're handling".
Hand Back to AI After (Hours)
In Autopilot Workspaces, this setting determines how many quiet hours must elapse before the AI may resume answering on the visitor's next message. Setting it to 0 means the AI never resumes automatically (your team retains ownership indefinitely). The maximum documented setting is 168 hours (one week).
15-Minute Unattended Safeguard
In an Autopilot Workspace, if a teammate takes over a conversation and then goes quiet for 15 minutes while the visitor is still waiting, the conversation returns to the team queue and the team is alerted ("Nobody followed up — open it to help this visitor") so a takeover cannot silently strand anyone.
AI handles visitor questions grounded in Knowledge Base.
AI pauses immediately; human teammate controls the conversation.
Returning the Conversation to AI
When human assistance is finished, teammates can hand the conversation back to the AI assistant so routine inquiries are once again handled automatically.
Manual "Return to AI" Action
In the conversation view, teammates can click the AI pill or the ⋯ menu and select Return to AI. This immediately releases the human takeover pause, enabling the assistant to resume.
Quiet Hours Hand-Back
If your Workspace has configured Hand back to AI after (hours), the assistant will automatically resume on the visitor's next message once the specified number of quiet hours has elapsed without new activity.
Replying Under Different AI Modes
How human replies interact with JIVIQ depends on your Workspace's configured AI responder mode: Autopilot, Copilot, or Off.
Autopilot Mode
AI normally answers visitors directly. When human assistance is needed, a teammate replies and the AI pauses automatically on that thread.
Copilot Mode
The AI never messages a visitor on its own. It prepares suggested drafts marked "waiting on you", allowing teammates to Send, Edit, or Discard each reply.
Off Mode
The AI neither replies nor drafts. Visitors write and your human team handles all conversations directly from the admin app.
tune Per-Conversation AI Controls
While the Workspace setting defines the default behavior, individual conversations can be customized directly using the AI pill above the reply box. Teammates can select AI answers this conversation, AI drafts only — this conversation, or AI off for this conversation. The individual conversation's setting always takes precedence over the Workspace default.
Best Practices for Human Replies
Providing effective human assistance through JIVIQ builds trust and solves customer issues quickly. Keep these practical practices in mind when taking over conversations:
1. Review Conversation History First
Read the prior messages exchanged between the visitor and the AI before typing. Understanding what has already been discussed prevents the visitor from needing to repeat themselves.
2. Use Available Customer Context
Check the customer context pane for identity information, captured email, or the visitor's presence status (Online, Idle, or Last seen) to tailor your response appropriately.
3. Answer the Visitor's Actual Question
Focus directly on resolving the specific topic or question that prompted human assistance, rather than providing generic introductions.
4. Avoid Repeating Information Already Given
If the AI previously cited documentation that was partially relevant, build upon that knowledge rather than re-sending the same generic text.
5. Clarify that Human Assistance Is Active
While the sender line clearly shows your first name and initials, greeting the customer naturally reinforces that a teammate is personally handling their inquiry.
Troubleshooting Replies
If your team encounters unexpected behavior when replying or taking over conversations, review these documented solutions:
AI Continues Replying After Takeover Expected
Cause: Viewing or opening a conversation does not pause the AI. The AI only pauses when a teammate actually sends a message, or when the AI pill is explicitly adjusted.
Fix: Send a reply to the customer, or click the AI pill above the reply box to manually pause or adjust AI for that conversation.
Conversation AI Setting Overrides Workspace
Cause: An individual thread's AI pill setting takes precedence over the Workspace default mode. If a conversation was set to "AI answers this conversation", it will continue answering regardless of Workspace settings.
Fix: Click the AI pill above the reply box in the conversation view and adjust it to "AI · Paused" or "AI off for this conversation".
Cannot Find Conversations Needing Human Reply
Cause: Conversations requiring human attention may be mixed into the standard chat list if filters are not applied.
Fix: On the Chats page in the admin app, look for conversations displayed in bold with unread count bubbles, or click the Escalated filter pill to view conversations handed off with hollow amber rings.
AI Resumed Automatically After Human Reply
Cause: In Autopilot Workspaces with "Hand back to AI after (hours)" configured, the AI will resume answering once the specified quiet hours have elapsed.
Fix: If you want human takeover to persist permanently until manually returned, set "Hand back to AI after (hours)" to 0 under Settings → Workspaces.
Frequently Asked Questions
Find answers to common questions about replying to customers, AI takeover mechanics, and conversation management:
Teammates open the conversation in the Chats view of the JIVIQ admin app, review the message history and available customer context, and compose their reply. Sending a message directly continues the conversation with the visitor and automatically pauses the AI on that thread.
No. In JIVIQ, sending a reply is the core action. The moment a teammate sends a message into the conversation, human takeover takes effect immediately and the AI pauses on that conversation.
Yes. Inside a conversation, the header shows presence details: Online, then Idle after roughly 45 quiet seconds, then 'Last seen' once they have been gone about five minutes. Presence is approximate and indicates whether a live reply is likely to reach the visitor right away.
AI replies display a ✨ icon, the assistant's name, and a visible AI tag. Human replies display the teammate's initials and first name without an AI tag.
Yes. When a teammate replies, the sender line above the message changes from the assistant's name to the teammate's first name and initials, making it obvious to the visitor that a person has joined.
No. JIVIQ maintains strict transparency. Even if you give your assistant a human-sounding name, AI responses always carry the ✨ icon and the visible AI tag, and the AI never claims to be human.
The AI automatically pauses on that specific conversation thread. The AI pill above the reply box switches to 'AI · Paused — you're handling', ensuring the AI will never talk over a human teammate.
The Claim button in the conversation header assigns the conversation to you so that other teammates can see who is actively handling the inquiry. Workspace administrators can also assign conversations to others.
In an Autopilot Workspace, if a teammate takes over and the conversation remains quiet for 15 minutes while the visitor is still waiting, the conversation returns to the team queue and the team is alerted ('Nobody followed up — open it to help this visitor') so a takeover cannot silently strand anyone.
Yes. A teammate can select 'Return to AI' in the AI pill or the ⋯ menu when human assistance is complete and they want the assistant to resume handling subsequent questions.
In Autopilot Workspaces, this setting controls how many quiet hours must elapse before the assistant resumes answering on the visitor's next message. Setting it to 0 means the AI will never resume automatically (your team keeps it), and the maximum setting is 168 hours (one week).
In Copilot, the AI never sends replies directly to visitors. When a visitor writes, the AI prepares a suggested draft marked 'waiting on you' with options to Send, Edit, or Discard, ensuring every message sent to the visitor is reviewed and approved by a human.
Human takeover takes effect when a teammate actually sends a message into the thread or adjusts the conversation's AI pill. Simply opening or viewing the conversation does not pause the AI. Additionally, verify that the conversation's AI pill is not set to 'AI answers this conversation'.
Yes. Team members with write access can adjust the AI pill above the reply composer for any individual conversation (such as AI answers, drafts only, or AI off), which overrides the Workspace default for that specific thread.
In the Chats view of the JIVIQ admin app, conversations awaiting team replies appear in bold with an unread message count bubble. Handed-off conversations also display a hollow amber ring and can be viewed using the 'Escalated' filter pill.