A routine gives a chief a repeatable instruction and schedule. Use it for bounded work such as reviewing a saved brief or preparing a report draft, and keep actions that need approval explicit.
A routine should say which chief owns the work, what input to read and what output to prepare. A weekly draft report is easier to review than a general instruction to manage all operations. Specify the frequency and any assumptions the chief should not make. The routine can use only the context and access actually available when it runs.
A scheduled browser task may need a signed-in session. If that session expires or a website changes, the routine may need your intervention. A saved instruction is not a guarantee that every external workflow will always succeed. Review failed runs and update the routine when the source data or website changes.
Routine capacity is a storage limit, while executions use applicable task, computer and compute allowances. An active entitlement and remaining allowance are required. Review the output and stop or edit routines that are no longer useful. Start with drafting rather than unattended publication when an action requires your approval.
Create a weekly routine for our research chief that reviews our saved business brief and drafts a short competitor update with accessible source links. Do not send or publish it. Ask me to confirm the schedule.
Sign in for a written first result before payment setup. Computer work needs an active trial or paid plan and remaining allowance.
No. Your plan has routine capacity and metered usage limits for the work performed.
Use the routine controls in your workspace or ask the assigned chief to update or remove the saved routine.
Brief an AI research chief to compare competitors and examine public sources. Get a scoped report with links, assumptions and gaps you can review.
Explore this workflow ↗Keep a business brief, saved conversations and shared files for your AI chiefs. Update changing facts and give each task the context it needs.
Explore this workflow ↗
Start with one message and route work to an existing AI specialist. Use @mentions, keep a clear handoff and build your team around actual tasks.
Explore this workflow ↗
