The owner digest turns recorded facts into short updates: exception batches during the day and a Monday digest of at most 150 words. Read them on the Dashboard's Digest page; Reload refreshes the recorded view and Sources expands the evidence behind each update.
What you see
Each project gets a line, failing projects first. It shows the latest recorded health or a dash and the reason it is unknown. A poll more than an hour old is labelled "not polled since", so yesterday's result cannot look current.
A new Branch health episode, or a silent workflow's first failure, appears once in an exception batch with a link to its run or workflow. A batch also names projects whose exceptions it cannot establish because their records, history or latest poll are incomplete. An unavailable blocker copy does not remove the original evidence: Sources retains the Branch health record or event and omits the unavailable copy.
To fit 150 words, the digest can shorten project details, count projects with unknown health, and count open asks instead of listing them. If that is still too long, it refuses to form the digest rather than cut it off. It says "Nothing needs you" only with current green health for every project, complete ask coverage, and no open asks or recorded owner-required items.
Limits and status
Status: Accepted local implementation: independent Astra reviews and combined checks passed. No live polling, scheduling or delivery. Acceptance and evidence.
Honest limits:
- Nothing runs on a schedule or sends an update. The proposed UK times are 09:00, 13:00 and 18:00 for batches, and Monday 07:00 for the digest; formation respects quiet hours from 22:00 to 07:00. These times are not a signed commission.
- The ask ledger is not connected, so asks remain unknown by default. When supplied ask records are incomplete, their count says "at least". Delight/frustration notes can be recorded by the module, but the dashboard only reads them. Reported minutes and estimated reading time are not measurements.
- Verification uses recorded and synthetic fixtures. No live delivery, applied rule or customer outcome is established. The module uses no AI model and has no brand-plate position.