Presence, capacity and next item: how omnichannel routing decides who works on what
Routing rules (Settings > Routing) distribute conversations, cases and tasks. This article covers the pieces that decide WHO can receive an item and in WHAT ORDER work is served.
Routing rules (Settings > Routing) distribute conversations, cases and tasks. This article covers the pieces that decide WHO can receive an item and in WHAT ORDER work is served.
Your status (presence) Every person has one status: Available, Busy, Break, Lunch, Training, Offline, plus any the administrator adds. Change it in the Work panel (bottom-left button) or in Settings > Routing > My availability. The same status is read by conversation and case routing, by the Next item button and by the voice queue: setting Lunch means no new items and no ringing.
Each status says what it does: whether it receives new items, whether it releases the items you already have (used by the "assignee unavailable" trigger), whether it returns to Available by itself (Break after 15 minutes, Lunch after 60) and whether it requires a reason. Someone without a declared status receives items as before.
Administrators see the team live in Support > Supervision: status, since when, load and time per status today. What appears there is what each person declared plus the open items the system already counts. There is no screen, keyboard or location tracking, and a status set by an administrator is recorded as such.
Weighted capacity By default the limits are counts (at most N open conversations, cases, tasks). Weighted capacity (Settings > Routing > People) lets you say that a critical case weighs 6 units and a chat weighs 2, and give each person a capacity in units. Load is the sum of the units of everything open, across channels. Routing and the Next item button stop offering items to someone who is full. Turn it on only if you want it; off, nothing changes.
Next item (pull) The Work panel and the case console have a Next item button. It serves the top of the queue you can take: SLA already breached first, then the closest deadline, then priority, then whoever has been waiting longest. It respects your status, your capacity, the queues you belong to and the skills those queues require. If nothing is for you, the button says why (status, capacity, queue membership, skill). Two people clicking at the same moment get different items.
New rule conditions Rules can now match on the customer's language (declared on the contact or company, or detected from the text; unknown language never matches), country and state, account tier and key accounts, the case product, the linked incident severity, and the SLA target (for example, first response target up to 60 minutes). Round robin and least load can be limited to people of a territory. A condition only matches when the fact is known; an item without that data goes to the next rule.
When nobody could take it Routing happens when the item arrives. If at that moment nobody could receive it (everyone out of rotation, at their limit, on a break, outside business hours, or without the skill the queue requires), the rules simply do not assign it, and until now nothing looked at that item again: it stayed with no owner until someone picked it up by hand.
Turn on **Re-check unassigned work** in Supervision (the Stuck work card). From then on, every minute the system takes the work that has had no owner for longer than you accept and asks the SAME rules again. Nothing else changes: same rules, same conditions, same skill and capacity checks, same trail. A rule that already acted on an item never acts twice, so re-checking never assigns anything twice.
Two numbers are yours to set: after how many minutes an item is re-checked (10 by default) and after how many minutes it counts as stuck (60 by default). Stuck is never earlier than re-check: the screen would be accusing work that nothing had tried to distribute yet.
Items that already have an owner are never touched here. Taking work out of someone's hands is a different decision and it has its own name: the "assignee unavailable" trigger and the "pass to the next person" action.
Acting from the supervision screen The wallboard used to say where it hurts and let you do nothing about it. Below the queues and the people there is now the list of items themselves, in the same order the work queue serves them, with who owns each one. From there you can reassign (one item or several at once), take the item over yourself, and change priority. Tasks have no priority in this product, so that option is not offered for them.
Every action from this screen is recorded in the assignment trail with the reason "supervisor", next to the automatic assignments. That is on purpose: supervising people leaves a trace of who changed what.
The stuck list also shows WHY an item is still there: no arrival rule exists for that kind of work, or the rules ran and found nobody. From the outside "nobody picked it up yet" and "nobody CAN pick it up" look identical, and only the second one needs you.