The screens
Nine of them. Eight are yours, one is ours, and all of them are read-only until you decide otherwise.
Every screen carries the same menu down the left, so you can get from any one to any other. Signing in once covers every screen in that browser tab.
Two of them answer most days on their own. Needs you is the list of things only a person can decide, and Analytics is what happened while you were not looking.
The console is in English today; the names below are as they appear.
Needs you
You or your team, daily
Everything waiting on a person, in one list: the tickets the agent handed over and the actions it wants approved. Sorted by how long the customer has been waiting rather than by what kind of row it is, because a two-day-old handover matters more than a four euro refund that arrived a minute ago.
/screens/needs-youAnalytics
You, whenever you sign in
The front page. Five numbers over a period you choose, 7, 30 or 90 days, each with how it changed against the period before. Under them: what kind of tickets came in, which of your rules answered them, and where your team recorded a different answer than the agent did. Until setup is finished it shows what is left instead of a row of zeroes.
/screens/Connect
You, once
Connect a helpdesk and a store. Connecting arms nothing: the connection is read-only until you change that yourself.
/screens/onboardingActivity
You, weekly
Every ticket the agent read, what it drafted and what it wanted to do. This is the screen a shadow week is judged on. It reports per-conversation facts and computes no headline accuracy figure, because that would be a claim nobody made.
/screens/activityApprovals
Your support team, daily
Proposed actions waiting for a person. Each row carries the money at risk, computed from the order rather than from the customer’s email, and the policy version that permitted it.
/consolePolicies
You, when the rules change
Your written rules, versioned. Save a change and the screen replays it against your own past tickets and lists the decisions that would have come out differently. That is a simulation, not a promise about next month.
/screens/policiesAutonomy
You, rarely
What the agent may do on its own, per workflow, and the limit on each. Also where the account-wide stop lives.
/screens/autonomyImports
Us
Ticket history brought in from your helpdesk and the labels reviewed on it. The one screen here that is ours rather than yours.
/screens/importsResult
You, at the end of a shadow week
The page a shadow week ends with: what the agent handled, how much of it a person changed, and how many minutes that saved. Every figure states what it was measured over, and a week too thin to answer says so rather than printing a number anyway.
/screens/reportWhat none of them can do
No screen carries out anything in your store. The approvals screen records a decision and queues it; execution is off by default and has never run, in production or in a sandbox. Every screen authorises against the server for itself, so what you can see and change is decided by the server rather than by which page you opened.