ConnectedAutomation
Make the recurring update a defined process.
Use an external scheduler and connected agent to summarize new work and update a team page on a chosen schedule.
A complete worked example
Make the weekly update a repeatable publishing job.
The team has an external scheduler and agent runner. It wants a scoped weekly summary of its request queue at one stable report URL.
Fictional project and sample data. This preview does not submit, publish, or change access.
Restricted team digest · illustrative result
This week in the studio
Monday snapshot from the configured run. Three deliverables moved forward and two decisions need an owner.
- Completed
- 3 deliverables
- Decisions needed
- 2 open questions
- Next milestone
- Client review on Thursday
What needs attention
Confirm the campaign image selection and assign an owner for the review invitation. The detailed request queue remains available to authorized teammates.
How it gets here
The external scheduler invokes the authorized agent or processing service. That runner gathers the selected records, writes this dated page, verifies success, and saves its checkpoint.
What the recipient can access
The selected status summary and authorized source references.
Kept out of this handoff
Other teams' work, private client material, and runner credentials.
How it runs
- Define the digestChoose the records, reporting window, audience, and destination page.
- Configure the runnerProvide the schedule and narrowly scoped credentials in an external system.
- Make retries safeRecord successful progress and keep errors visible without duplicating results.
- You getA regularly refreshed digest with an explicit time range and source references.
A prompt to start with
Copy this into your AI tool after connecting it to Fermium. Replace the bracketed details and let it ask for the context it needs.
Prepare a recurring digest workflow for [table] using [configured external runner]. Summarize [scope] on [schedule], update [existing page] for [audience], and advance the saved cursor only after a successful run. Document setup and failure handling; do not imply scheduling exists until the runner is configured.
What this workflow needs
Needs an external scheduler, an available agent or processing service, and appropriate credentials and plan. A saved Agent workflow alone does not run in the background.
Can we start manually and add the schedule later?
Yes. Run the report on request first, then configure and verify the external runner when the routine is useful.