Live Chat in CloudRadial Chat: How It Works

Live Chat is what happens when a client needs to talk to a person. Whether they are coming from a Dialog conversation that collected their initial information or a ChatAI session that attempted troubleshooting, the escalation to a live agent follows the same pattern: the client requests help, an agent picks up the chat, and the conversation continues with full context from everything that came before.

The key difference from traditional support channels: the agent already knows what is going on. The conversation transcript, every question asked, every answer given, and every troubleshooting step attempted, is available the moment the agent joins. There is no "can you describe the problem again?"

• The escalation flow
• The group chat: three perspectives
• Agent commands during Live Chat
• Handling multiple chats
• Cross-platform bridging
• What gets logged

The escalation flow

Step 1: Client reaches Live Chat

How a client reaches Live Chat depends on the Responder. With Dialog (Starter), the client is automatically offered Live Chat after their ticket is created, provided it is during your configured Live Chat hours. With ChatAI (Pro), the client can ask directly ("Can I talk to someone?") or ChatAI may suggest it when it cannot resolve the issue.

Step 2: Client enters the queue

The client is placed in the Live Chat queue and sees a message confirming they are waiting. They can continue working while they wait, since chat is asynchronous.

Step 3: Chat posts a request card

Chat posts a Chat Request Card into the designated Teams channel and creates a conversation for the agent to join. The client does not see anything yet. How the chat gets assigned depends on your Live Chat Strategy, configured under Settings on the Live Chat Settings tab:

• Do not auto-assign: the card posts to the channel and any available agent can take the chat by clicking Join Live Chat on the card. This is the default.
• Round Robin: Chat automatically assigns the chat to the next available agent in rotation.
• All Operators: all available agents are added to the group chat.

Step 4: Agent joins the chat

When an agent clicks Join Live Chat (or is auto-assigned), Chat creates a group conversation with three participants: the client, the agent, and Chat. The client now sees that an agent has joined and can start communicating directly. Chat replies to the original card in the Service Channel to confirm the agent has taken the chat. By default, the ticket in your PSA is automatically assigned to the agent who picks up the chat first; this can be toggled off under Settings on the Live Chat Settings tab.

Step 5: Conversation and resolution

The agent works with the client to resolve the issue. Within the group chat, Chat displays a TicketCard that is visible only to the agent. The TicketCard shows the ticket details and provides buttons to update status, change priority, and add notes, all without leaving the conversation. The agent can also @-mention Chat to execute commands, which are invisible to the client. Because the Live Chat conversation is a standard Teams group chat, the agent can add other team members directly into the conversation if they need help. When the issue is resolved, the agent closes the ticket and the conversation ends with a complete record in the PSA.

The group chat: three perspectives

This is the core of how Live Chat works in CloudRadial Chat, and it is different from most chat systems.

• What the client sees: a conversation with their support team. They can send text, screenshots, and file attachments. They see the agent's messages. They do not see anything the agent does with Chat behind the scenes.
• What the agent sees: everything. The full conversation the client had with the Responder before escalation, plus the agent's own conversation with the client, and Chat's responses to any commands. If ChatAI already asked "have you tried restarting?" and the client said yes, the agent sees that and moves on.
• What Chat does: handles the plumbing. It relays messages between platforms when the client and agent are on different tools, executes commands (updating tickets, fetching information, managing the queue), and stays available for the agent to use throughout the conversation.

Agent commands during Live Chat

Agents interact with Chat through commands that are invisible to the client. In Teams, common commands are available as buttons on Adaptive Cards, so agents do not need to memorize syntax. See the agent commands reference article for the full list.

Handling multiple chats

Agents can handle more than one Live Chat at the same time. While waiting for a computer to restart in one conversation, the agent can work on another. Each chat is a separate group conversation in their Teams or Slack environment, and they can switch between them just like any other conversation. This makes efficient use of the dead time that is inherent in support: processes running, downloads completing, users locating information. On a phone call, the agent sits and waits. In chat, they stay productive.

Escalation within the team

Because Live Chat runs as a standard Teams group chat, adding another agent is as simple as adding a participant. The new agent joins with full visibility into everything that has happened: the client's original description, Chat's interaction, and the first agent's troubleshooting. There is no separate handoff conversation.

Cross-platform bridging

Live Chat works across platforms. The client and agent do not need to be on the same tool: client on UCP to agent on Teams, client on Teams to agent on Slack, client on SMS to agent on Teams, client on Web widget to agent on Slack. Chat handles the translation. Messages, formatting, and attachments are relayed between platforms.

What gets logged

Everything. The complete conversation (client messages, agent messages, Chat interactions, and command usage) is captured in the Chat admin portal as a full transcript. Chat writes the transcript to the ticket in your PSA twice: once when the initial ticket is created (capturing the Responder conversation), and again when the Live Chat ends (capturing the full agent interaction). This matters for quality assurance, training, accountability, and continuity: if a client calls back about the same issue, any agent can read the original conversation.