Agents & automations

Tori taking the boring parts off your plate — on a schedule, on a trigger, or on demand.

What an agent is

An agent is a small Tori-powered routine that runs on a schedule (every Monday morning) or a trigger (when a task moves to Done). Each agent has a clear role: summarize, draft, route, notify. Agents don't take action without human review unless you explicitly mark them as auto-execute — and even then, every run is logged.

Things teams automate

Weekly status summary — every Friday afternoon, Tori writes a digest of what shipped, what slipped, and what's next.
Stale-task pinger — every Monday, surface tasks that haven't moved in 7+ days.
New-member onboarding — when someone joins, send them a welcome with the right onboarding doc for their role.
Standup prep — 15 min before standup, draft each member's update from the day's task moves.

Creating an agent

From the workspace dashboard, open Agents → New. Pick a template or start blank, give the agent a clear role description (what should it do, in plain English), set the trigger (schedule or event), and choose whether outputs are draft-only or auto-executed.

Note

Agents respect your workspace plan's automation budget. Free workspaces have a small budget; Team and School plans have larger ones. See Plans.

Staying in control

Every agent run produces a log entry — who/what triggered it, what Tori produced, what action was taken. You can pause, edit, or delete an agent at any time. If something goes wrong, you have the trail to see exactly what happened.

Tip

Default new agents to draft-only. Once you've seen a week or two of output you trust, then flip auto-execute on. The friction of reviewing drafts at first is small; the cost of a mis-firing automation can be much larger.