Configure Autonomy Policies
Set the permission mode for each AI action type — using the Autonomy dashboard, no API calls required.
Overview
Every autonomous action in Plenix is governed by a policy that defines how much autonomy it has. You configure these policies from the Autonomy dashboard — no API calls or technical setup required.
Where to configure policies
Go to Autonomy in the left sidebar. The main Autonomy page shows all known action types as cards, grouped into Diagnostics and Remediations.
Each card represents one type of action the AI or your automation rules can request. The four mode buttons on each card control what happens when that action is triggered:
| Mode | What it means |
|---|---|
| Disabled | The action is completely blocked — requests are rejected and logged |
| Suggest only | The action is queued in Approvals — a human reviews it before anything runs |
| Requires Approval | Same as Suggest — queued, waiting for explicit sign-off |
| Fully Automatic | The action runs immediately, no approval needed |
Click any mode button to change it — the update saves instantly.
Default modes
If you have not changed anything, Plenix uses these defaults:
- All 7 diagnostic playbooks (disk usage, service status, event log errors, pending updates, network test, printer queue, resource consumers) — Fully Automatic, because they are read-only and make no changes to the device
- Remediation actions (restart service, clear temp, force GPO update, reboot device) — Requires Approval until you explicitly set them otherwise
This means you can start using diagnostic playbooks and AI Triage immediately, while keeping human oversight over anything that changes a device.
Enabling AI Ticket Triage
AI Triage is on by default for Enterprise tenants — you do not need to configure anything to start receiving triage notes on new tickets. If you want Triage to run diagnostic playbooks automatically when it identifies an affected device, ensure the relevant playbook's mode is set to Fully Automatic (the default).
Setting a rate limit
Each action card has an optional rate limit setting: the maximum number of times that action can run per hour across your entire tenant. This protects against runaway automation. Click the rate limit field on a card to set a value. Once the limit is hit, further requests are rejected and logged — they do not queue in Approvals.
Viewing recent runs
Below the policy cards, the main Autonomy page shows a Recent AI Action Runs table — every action that has been requested, with its status (pending, approved, running, succeeded, failed, rejected) and a link to review if it is pending approval.
Where to go next
Was this article helpful?