The Workforce Hub: Usage, ROI and Rehearsals

The AI Workforce tab is where you see your whole AI team at once: what each AI member is doing, what is waiting for a person, what they cost and what they are worth. You'll find it under Settings → Artificial Intelligence → AI Workforce. You need View AI Workforce, and the AI Workforce app switched on for your organization.

It has four tabs.

Team

Every AI member, one row each:

• Status: Working, Off shift, On hold, Paused, Blocked, Starting or Retired. A paused one says who paused it: Paused by a supervisor, Organization paused or Platform kill switch. One on hold says why, for example Daily budget reached.
• Health: a warning sign appears when something it needs is missing or refused. Point at it to see what: its Session, Mailbox, Phone, Workstation or Real-time wake. "Checked every 5 minutes" means it is not woken the moment something arrives, so it notices new work within five minutes instead.
• Doing now, Runs today, Cost today, waiting Approvals and its Last run.
• Actions: Wake now (make it check in at once), Pause or Resume, and Open profile. Pause and resume need you to supervise it, or Manage AI Workforce.

The list updates by itself as AI members work. When you have none yet, Hire an AI member takes you to the hiring form.

Approvals

The actions your AI members proposed and are waiting for a person to decide. Switch between Mine, My AI team and Everyone, and between Pending, Decided and All. Click a row to review it. See How to Approve or Reject an AI Action.

Usage

What your AI members cost, from Kamo's own metering, for the period you choose with From and To:

• Total cost, then the cost per AI member (with its Runs, Input tokens and Output tokens);
• per Capability, such as text generation, speech to text or embeddings, with the quantity used;
• Cost per day, as a chart.

The costs are what your own AI provider accounts charge, worked out from the prices recorded for each model in the AI Provider Registry. Kamo does not bill model usage itself; your provider does.

Return on investment

The work your AI members completed in the period, set against what the same work would have cost a person:

• Handled: Emails, Chats, Calls and Tasks.
• AI cost and Cost per item.
• Human equivalent: the minutes that work would take a person, times a Human hourly cost.
• Savings: the difference.

The figures rest on Assumptions you can change, if you have Manage AI Workforce: Minutes per email, Minutes per chat, Minutes per call and Minutes per task (from 0.5 to 240 each), and the Human hourly cost (USD). Leave the hourly cost empty to use the blended rate from your timecards; the page says which one it used (Set by your organization, From timecards or Not set). With no hourly cost at all, it shows no savings.

Rehearsals

A rehearsal shows what an AI member would do with real work, without it doing anything. It replays the email, chat, tasks and calendar it received over the last few days through its current charter and models. Nothing is sent or saved.

Rehearsals are on the AI member's own workbench, in the AI room of its profile:

1. Under Rehearsal, choose how many Days to replay, from 1 to 14.
2. Click Start rehearsal. It needs the AI member to be Working, and only one rehearsal runs at a time.
3. Watch it run ("Running: 4 of 12"). When it is done, Past rehearsals shows a summary such as "12 items, 5 actions, 2 approvals". Open it to see each item and what the AI member would have done: acted, asked for approval, escalated, or decided there was nothing to do.

A rehearsal is the safest way to check a new charter, a change of model or a move to more autonomy before it meets a real customer. If the AI member tries to really change something during a rehearsal, it is stopped and the report says so.

Starting a rehearsal needs you to supervise the AI member, or Manage AI Workforce. The model calls a rehearsal makes are real, so they cost what any other work costs.