Quick answer
An AUX code — short for auxiliary, telephony's original word for "not ready" — is the reason an agent declares when they step away from taking calls. Genesys calls it a reason code, Five9 calls it an aux state or not-ready reason, Talkdesk calls it a status reason. Whatever the vendor vocabulary, the mechanics are identical: a short, tenant-defined list (lunch, coffee, coaching, dental) that an agent picks in their softphone, stamped on every state change, and rolled up downstream into adherence, occupancy, and payroll reporting.
Devotel Orbit ships this as a per-tenant break-reason taxonomy — the Voice → AUX Codes page in the dashboard, backed by four API endpoints over your organization's own catalog. Each code carries a label, an adherence bucket, a paid flag, and a sort position. Agents pick from the list; owners and admins edit the list; reporting buckets it.
Why break taxonomies matter
A state change without a reason is a number you can't act on. The taxonomy is what turns "agent went unavailable for 11 minutes at 10:42" into "coffee break — paid — within the floor's allocation." Three disciplines depend on it:
- Adherence. Workforce management compares scheduled state against actual state. If every step-away collapses into one generic "unavailable," the adherence report can flag time as out-of-schedule without telling you whether it was a legitimate break or genuine shrinkage. A taxonomy splits the two:
lunchandcoachingare paid, planned, and report differently than an unexplainedunavailable. - Reporting. Break minutes roll into occupancy math — the share of logged-in time an agent spends on customer work versus everything else. With named reasons, an operations lead can answer "where does non-contact time go" from a report instead of a guess. Without them, the question is unanswerable.
- Payroll discipline. The paid-vs-unpaid distinction is the one your payroll export reads. Paid coffee break, unpaid lunch — that split is encoded on the code itself, not negotiated per agent per day.
This is why Genesys, Five9, and Talkdesk all ship a reason-code editor as day-one table stakes: without it the rest of the WFM stack has nothing honest to measure.
Building a sane taxonomy
The failure mode is not a missing taxonomy — it's an unchecked one. Three rules keep a break catalog usable:
- Group by the fixed buckets, not by mood. Underneath your labels sits a small, fixed adherence vocabulary —
available,on_call,wrap_up,break,lunch,training,unavailable,logged_out. Your slugs extend it (coffeerolls intobreak,1-on-1rolls intotraining), but the report groups on the bucket. Taxonomies that invent a hundred bespoke buckets produce reports nobody reads; taxonomies that map many slugs onto few buckets produce reports a supervisor can act on. - Cap the list. If the agent's picker needs a scrollbar, the taxonomy is wrong. Ten to fifteen codes cover the floors that actually report well. Every code you add is a choice an agent has to make mid-flow and a line somebody has to explain in a quarterly review.
- Ban free-form. The moment agents can type their own reason,
breakforks intobreak,Break,brb, andon break— and your reporting splits across four strings that mean one thing. A closed, tenant-managed list is the mechanism that keeps the data honest; the editor exists so addingdentalis a thirty-second admin task, not a code change.
One naming discipline to set early: the slug is forever. The code (coffee) is what gets stamped on every historical state change — the label (Coffee break) is display. Rename a label freely; renaming a slug orphans the history, so retire the old one and create the new one instead.
Where the tenant manages it: Voice → AUX Codes
In the Devotel Orbit dashboard, the taxonomy lives at Voice → AUX Codes — a per-tenant editor over your organization's catalog. The page lists the active codes in the order you chose, with the archived set one toggle away.
- Create a code with a slug, label, adherence bucket, paid flag, and sort position. Slug collisions against an active code return a conflict; colliding with an archived code is fine.
- Edit the label, bucket, paid flag, or ordering in place. Owners and admins can edit; agents and supervisors pick from the list but can't change it.
- Archive a code to remove it from every agent picker without touching history — archived slugs stay valid on past records, and restoring one is a flip back to active.
Defaults already exist — lunch, break, and training map to their buckets and anything else falls back to unavailable — so a floor works before you customize anything. The catalog only overrides the defaults for the codes you actually define.
The same four operations are public API endpoints — list, create, update, and archive/restore — so provisioning a standard taxonomy across workspaces or driving it from an internal tool is scriptable. Every write lands on the audit trail with who, when, and what changed; the author of this control is your organization, not the platform.
Where the codes surface downstream
AUX codes are configuration, but their job is reporting. The same tenant-owned catalog feeds two downstream surfaces:
- Adherence. When an agent changes state, the reason slug they picked resolves to its adherence bucket before the default mapping is consulted — so your
coffeecode counts as a break minute, not an unexplained unavailable one. The adherence view then compares scheduled to actual using bucketed time, which is the difference between a shrinkage report and noise. - Reporting and exports. Paid/unpaid flags travel with the time records, so occupancy reporting and payroll exports read the flag directly instead of asking finance to re-classify raw timestamps. Archive a code and its historical minutes keep their recorded slug — the report past February still says
coffee.
Both paths read tenant configuration only; nothing in the AUX-code surface touches call routing or outbound dialing of any kind.
Open the editor
The full field list, API request shapes, conflict behavior, and archive/restore semantics are in the AUX codes guide. For the adherence side, the intraday adherence and requests guide shows how bucketed time rolls into the schedule-vs-actual view. The wider supervisory loop the taxonomy feeds is covered in quality leaderboards that work and the workforce engagement primer.
Open Voice → AUX Codes in your workspace, create the five to ten codes your floor actually steps away for, and the next adherence report starts answering "where does non-contact time go" instead of shrugging at it.