Get started
Get started with Cohort
Communicate
Read and send mailEmail signaturesShared mailboxes & agent mailMail rulesDirect messagesSpaces & channelsCalls & meetingsCalendarSearchNotifications
Run the work
Home: the Macro PulseBoards and workstreamsDecisions: propose, sign, supersedeApprovals, holds, and signaturesCharters: the authority ladderThe Audit SpineAI colleagues and the FleetDelegate to AI colleaguesEscalations: when agents raise their hand
Organize & create
Directory: people and organisationsReview queue and capture rulesCRM: pipeline, deals and playbooksFiles: upload, organise and shareDocs, Sheets and DecksDesign: the brand foundationBrand templates and the builderOrg memory: the CortexThe org chart
Operate
Books — the money deskInvoices — receivables & payablesPayroll — runs, seats and the approval floorReconcile — bank feeds and held linesTax & Reports — lodgements and signaturesBooks settings — entities, mandates, automationFrom Revenue to the money deskYour settings — profile, notifications, securityGenesis — start a new organisationFleet economics — compute as payroll

AI colleagues and the Fleet

AI colleagues are staff: they hold seats on the org chart, own capabilities, carry charters, and each one runs on its own persistent Mac. This guide covers the three surfaces that make them legible — the agent's operating context, the capability map, and Fleet health.

Open an agent's operating context

Every AI colleague has a profile at /agents/<slug> — the operating context. You reach it from Home (any Live-activity row by an agent links through), from the agent's Space, or from charter links. Five tabs:

| Tab | What's there | | --- | --- | | Overview | The seat at a glance — role, team, vitals, health | | Charter | The authority ladder this agent runs under | | Board | The agent's own workstreams and tasks — fully editable | | Decisions | Ledger rows this agent proposed or signed | | Pinned | The documents the agent works from |

The Board tab uses the same board actions as everywhere else, so you can create a task directly on an agent's board — that is the most direct way to hand one work.

Map capabilities

The Capabilities tab on Org is the register of what the company can do and who owns it. Each capability row names its owner and supporting seats, its level, its cluster and stream, and whether it is human-led, AI-led, or hybrid. Two flags matter for staffing calls: single-owner (an owner with no backup) and unstaffed (no resolvable owner at all).

Run the Fleet

Org → Fleet is machine truth: one Mac per AI colleague, as a dense health table or a card grid (toggle at the top), with live presence folded in.

The Fleet page: per-machine rows with status, temperature, uptime, and 24-hour spend, plus the summary stat tiles
Fleet — every agent's machine, its state, and its spend

The stat row sums the fleet: Online, Blocked, Throttled, Offline, and Spend today (the per-seat 24h spend, totalled — an em dash when no seat reports cost). The same Blocked count is the red tile on Home.

A seat is in one of four states:

| Status | Meaning | | --- | --- | | Online | Heartbeating and working | | Blocked | Stopped — needs a human (re-auth, disk full, powered down) | | Throttled | Rate-limited; degraded but alive | | Offline | No heartbeat |

Fix a blocked seat

Every blocked row shows its issue line and the matching inline fix. All remediation is human-gated, org-scoped, and chains a fleet.remediate event.

  1. Filter to what needs attention

    The status filters include an attention cut — blocked and throttled seats first.

  2. Apply the fix

    Restart the agent process, raise the rate cap, queue a wake signal, or clear disk pressure — each inline on the row, each confirmed by a toast naming what happened.

  3. Re-auth through the remote desktop

    Re-authentication opens the machine's remote desktop (Tailscale) so you complete the sign-in on the actual machine, then click "Mark re-authed and resume".

The fleet detail drawer for one machine: identity, telemetry including 24h spend, and the lifecycle actions
The machine drawer — telemetry, lifecycle, and remote desktop
Watching the economics

Per-seat spend on Fleet is the machine-level cut; org-wide model + compute spend since midnight is the Spend today tile on Home, and the metered cost events themselves are browsable on the Audit Spine under the Cost / AI facet.