# Your Console

> One console covering your managed client workspaces: fleet health, cross-client activity, workflows, users, AI activity, and delegation grants.

The console brings your managed client workspaces into one view. It shows which automations are failing, what changed and where, AI generation activity, and who has access to what.

It appears in the sidebar as your **Consultant Dashboard** or your **MSP Dashboard**, according to how your practice is set up. See [Working With Clients](/msp/overview#choose-the-path-that-fits-your-practice).

The **Open Admin Console** button in the workspace switcher opens the same console, or go straight to `/app/console`.

> The console shows only the clients you hold an active delegation grant for,
> and only what that grant's tier permits. A client who revokes consent
> disappears from it immediately. See [Governance &
> Impersonation](/msp/governance).

## Tabs

### Overview

Your landing page. It shows fleet health across all managed workspaces: total workflows, how many are active, and user counts. Alongside sits the client roster and an activity rail. Start an impersonation session or open a client's workspace from here.

### Activity

A cross-client feed of workflow, execution, governance, impersonation, deployment and membership events. Filter to a single client to trace when something changed or failed.

Pending alerts appear above the feed, so what needs attention sits ahead of what merely happened.

### Workflows

Every workflow across every managed workspace, in one list. Use it to find a specific automation without switching workspaces, and to see at a glance which clients have something running and which do not.

### Fleet

Execution health across the estate, with a trend chart over a selectable window.

> **Individual runs follow the client workspace's retention policy.** The
> default is 30 days after a run finishes, and the policy can differ by
> workspace. Investigate a reported failure while its run is still available if
> you need to open the underlying detail. See [Data
> Retention](/reference/system-limits#data-retention).

### AI

AI generation activity with status, workflow, user, model, time and response details. Summary metrics show generation volume, failures, clarification rate and tool use, helping you spot patterns that need investigation.

> This tab shows activity rather than spend. See [Billing across
> clients](#billing-across-clients) for how costs work today.

### Users

Users across your managed workspaces. Useful when onboarding, offboarding, or answering "who has access to this client?"

### Governance

Delegation grants and client invites: who has granted you access, at what tier and scope, what is still pending, and what has been revoked. Manage your client roster here.

## Billing across clients

Billing today works per workspace. Consolidated billing across your clients is in
development. Talk to your account team about where it stands and what
arrangement suits your practice.

How it works right now:

- **Each tenant is billed separately.** A client's plan sets their limits, and those limits apply inside their workspace. See [Billing & Usage](/manage/billing/overview) and [System Limits](/reference/system-limits).
- **Credits are per workspace.** When one exhausts its allowance, its workflows stop with an **Out of Credits** error. Yours running out does not affect theirs, and the reverse.
- **Verify recovery after restoring credits.** Confirm the client workflow is **Live**, then send one controlled webhook request or inspect the next scheduled occurrence. Check that a new run appears, and decide separately whether any missed work still needs replacing.
- **Check the plan before you push.** A workflow that loops 500 items runs fine on Pro and fails on Free, so confirm the target workspace's plan covers what the workflow does.

## What's Next?

When the console shows you a client's workflow failing:

- **[Troubleshooting](/reference/troubleshooting)**: the common failures, by symptom.
- **[Error Handling](/build/core-concepts/error-handling)**: retries and error paths, so the next failure handles itself.
- **[Governance & Impersonation](/msp/governance#impersonation)**: going into the client's workspace to fix it.