Set up chat for your support team
π£ Open Beta β Premium and Enterprise. Live chat with a human agent (including AI-to-human handoff) is in Open Beta. It's available to Premium and Enterprise tenants and is on by default for any new space, in both existing and new sites. Existing spaces stay off; an admin can turn it on at the space level.
This article covers the space-level setup you need to complete once before your team can accept live chat conversations with customers. For customer-experience setup (chat widgets, routing, handoff), see Set up your chat experience. For the agent guide once chat is on, see Manage live conversations with customers.
Before you start
A Premium or Enterprise tenant with Service Collection
You have admin permissions in the Customer Service Management space.
(Optional) Your support team have agent roles in the space.
Turn on chat in your space
Ensure the feature is turned on:
Go to your space and open the space settings.
Open the chat settings.
Turn on Live chat.
Once chat is on, your team can set their availability and start taking conversations through the conversation panel.
π Turning on chat at the space level doesn't expose chat to customers yet. Customers can only see and start chats once a chat widget is configured for one of your customer experiences. See Set up your chat experience.
Understand the conversation limit
The conversation limit is the maximum number of active conversations an agent can have at once. When an agent reaches the limit, new conversations are offered to other available agents until one of theirs is resolved.
The limit is set by the Maximum conversations field. It applies at the site level across all spaces β it isn't a per-space setting. A site admin sets it in Jira administration:
Go to Jira administration > Jira apps.
Open Customer Service Management.
Open Chat settings.
Set Maximum conversations to the value you want.
Choose a limit that matches how many conversations your agents can comfortably handle at once. A lower limit helps agents stay focused and respond quickly; a higher limit lets each agent take on more conversations when your team is busy.
Get your team ready
Before your team starts taking conversations:
Walk them through Manage live conversations with customers so they know how to accept, run, and resolve conversations.
Agree as a team how you'll use availability (for example, who's on, who's on a break, who's heads-down).
Agree what happens when only one agent is available. If that agent declines a conversation, it's offered back to them, because no one else is available to take it.
Routing and fallbacks
Incoming conversations are offered to your available agents in the space in a round-robin pattern, up to each agent's conversation limit. Each agent can accept or decline a conversation that's offered to them.
If only one agent is available and they decline a conversation, it's offered back to them, because no other agents are available to take it.
If no agents are available, the customer sees your fallback message straight away. If agents are available but no one accepts within 20 minutes, the conversation ends and the customer sees your fallback message. To set that fallback message, see Set up your chat experience.
Conversations in your queues
Live chat conversations are work items, so they appear in your space's existing queues alongside your other work unless you filter them out. To keep conversations separate, create a dedicated queue and filter by the conversation work item type, or exclude that work item type from your existing queues.
You can use JQL to build queues that target conversations specifically. Only agents who are assigned to or participating in a conversation can see its details.
Don't set the assignee on conversation work items through queues or automation. The assignee is set automatically by the routing system when an agent accepts a conversation. Changing it through queues or automation can cause the assignee to lose sync with the actual conversation, which leads to misleading assignment and reporting.
Service-level agreements (SLAs) for chat
Chat conversations are work items, so you can set SLAs against them using the standard SLA process. You can narrow SLAs to chat conversations using the work item type.
A time-to-first-response SLA on chat works the same way as on other work items β the timer stops when an agent replies.
Was this helpful?