The fleet console’s robot, telemetry, and alert screens tell you what the fleet is doing. The operator roster tells you who is allowed to drive it. It is the teleoperation analogue of a contact-centre agent directory plus a skills admin: who can drive what, at what level, and until when. In ClutchCall, a job offer for a robot only reaches an operator who holds the certification that robot requires. Everything on this page exists to control that one matching decision.

​
The operator record

The roster screen reads the whole roster with api.operators(). Each row is one operator record:

​
How an operator gets on the roster

The roster is self-populating. You do not create a blank operator record and fill it in. When the roster is empty, the screen says so explicitly:
No operators yet. A row appears the first time someone opens the console.
So the sequence is: a person gets console access, they open the console, and a row for them appears in the roster. From that point on the row is the thing you administer — you set a region on it, and certifications attach to it. A new row carries no certifications. Until one is granted, that operator holds nothing, and no robot that requires a certification will be offered to them.

​
Presence states

Presence is reported on the record, not set from this screen. The roster renders it as a colour-coded chip so you can read availability down the column: attached and ready share the accent colour because both mean the operator is on shift. aux is called out in the warning colour because it is the state that silently removes someone from the available pool while they still look logged in. offline is faint because an offline operator is not expected to take work. When auxReason is set it is appended to the chip text after a · separator, so the roster shows why someone is unavailable without a second click.

​
Skills, levels, and certification expiry

A certification is a triple:
  • skill — what class of work it authorises.
  • level — the depth of authorisation within that skill. An operator may hold the same skill at more than one level; each skill · level pair is a separate certification chip with its own expiry.
  • expires — the date the certification runs out.
The roster shows every certification an operator holds as a chip reading skill · level, and adds one derived column, Nearest expiry: the expires value of the certification that lapses soonest, found by sorting the operator’s certifications by expiry and taking the first. An operator with no certifications shows — in both columns. Nearest expiry is a review column. It answers “who lapses next” at a glance so you can renew before someone drops off the roster. This screen does not renew or extend a certification, and it does not report what dispatch does at the moment one lapses — it reports the date.

​
How a job offer is matched to a certified operator

Dispatch is certification-driven. A robot declares a required certification. The automatic call distributor offers the job only to operators holding that certification. That is the whole matching rule as far as this screen is concerned, and it has two practical consequences:
  1. Certifications are the grant of access. An operator who is ready but does not hold the robot’s required certification is not offered that robot.
  2. Revoking is how you take someone off a robot. There is no separate “unassign from robot” control, because there is no assignment — there is only the certification that makes an operator eligible.

​
Revoking a certification

Click any certification chip on an operator’s row to revoke it. The chip reads skill · level ✕ and its tooltip is Revoke this certification. The console confirms first, naming the operator and the exact skill · level being removed, and stating the effect:
They stop being dispatched to robots requiring it, immediately.
Confirming calls api.revokeCert(operatorId, skill, level) and reloads the roster. Revocation is scoped to that one skill · level pair — other levels of the same skill, and all other skills, are untouched. Note what this screen does not do: there is no control here for granting a certification. The roster is where you review and revoke. Granting happens outside this screen.

​
Region

Region is a free-text grouping label on the operator record. Use the Region button on a row to edit it:
  • The prompt is pre-filled with the operator’s current region.
  • Cancelling the prompt makes no change at all.
  • An empty answer is a deliberate clear — submitting nothing sets the region to empty, and the column falls back to —.
  • The value is trimmed, then length-checked before any request is sent. A region longer than 64 characters is rejected locally with a Region length message; the API is not called.
A valid value calls api.updateOperator({ userId, region }) and reloads the roster.

​
Errors and in-flight writes

Every write on this screen is per-row. While a region edit or a revocation is in flight for an operator, that row’s Region button and certification chips are disabled, so you cannot stack two writes on the same record. Failures — from the client-side region length check and from the API alike — land in the same error strip at the bottom of the screen, showing the error message. The strip is cleared when the next write starts, so what you see is always the outcome of the most recent attempt.