Feed

The Feed answers one question: what needs doing today? It is a worklist, not a notification column — every card carries a decision or an action, and once it’s done it disappears.

One card, one item

Every card has a state: pending → executed, skipped, auto-handled or failed. Pending is the default view; everything else is history.

Cards come out of live operations, not from a separate task manager:

Family	What’s on it
Replies	draft to approve, reply sent, positive reply to review
Follow-ups	due follow-ups, stalled conversations with no reaction
Calls	lead with a phone number and a current reason to call
Meetings	meeting booked, meeting prep
Proposals	switch copy, change cadence, sharpen playbook or targeting
Deliverability	pause suggested, report
Knowledge	distilled learning to approve
Escalation	what a human has to decide

The principle: nothing pauses silently

When the system spots a problem — poor deliverability, a disconnected account, copy that isn’t landing — it does not stop on its own. It puts a proposal in the Feed and you decide. This is deliberate: a silent automatic stop is indistinguishable from “running normally” from the outside, and that is exactly how outages nobody notices come about.

The same rule applies in the other direction: anything that goes live and takes effect externally passes through the Feed as an Approval first.

Filtering and scope

By channel (LinkedIn, email, phone), by playbook, by sender, plus free-text search. Someone responsible for only certain playbooks sees only those cards — the Feed is scoped to your responsibility, not to the whole workspace.

Individual lanes are switchable per workspace (call hints, knowledge approval) and stay entirely invisible otherwise, rather than sitting around as an empty section.

From card to receipt

Every card links to its Run — status, scope, cost, origin. Nothing to go dig up: the card says what to do, the run says what happened.