Chat assignment
A workspace-wide singleton that controls how new conversations are auto-routed to agents, the default concurrent cap per agent, and what happens to open conversations when an agent goes offline.
Configuration
Settings → Chat Assignment.
Auto-assignment
Master switch. When off, new chats land unassigned and require manual pickup from the queue.
Default concurrent cap
The maximum open conversations any agent can hold before the router skips them. Range: 1–100.
Per-agent overrides on Settings → Users override this cap for individual agents (e.g. a senior agent with a cap of 10 while the workspace default is 5).
Routing strategy
Pick one:
- Round robin — rotate evenly across eligible agents in id order. Even load over time.
- Load based — always pick the agent with the fewest open conversations. Recommended.
- Skill based — match agents to conversation tags first; falls back to load-based when no tag match exists.
When an agent goes offline
Action taken on every conversation owned by an agent who logs out or loses presence:
- Reassign — re-route every open chat to another eligible agent.
- Queue — clear the assignment so a human picks the chat back up from the queue.
- Hold — leave the assignment as-is; admin handles manually.
Forward-compat config
A small block of toggles for features not yet enforced by the runtime:
- Priority routing.
- Show queue position to the customer.
- Show estimated wait to the customer.
These are stored but have no runtime effect today. They’re labelled “Not yet enforced” inline.
Best practices
- Start with Load based and a moderate concurrent cap (5–7). Tune the cap based on observed agent saturation in the supervisor dashboard.
- Set the offline action to Reassign for high-availability teams and Queue for smaller teams where no fallback agent is realistic.
- Override concurrent cap per-agent only when you have specific seniority or capacity differences. Workspace-wide caps are easier to reason about.