Step 1 (days 1–2): policy, limits and the always-escalate list in Setup
The first step with an Order Support FTE is to write down what the store actually does, because most stores carry their post-purchase policy in the heads of two people. In the Setup tab the owner records the return window per category, the refund limit and goodwill limit the FTE may resolve within, the exchange and replacement rules, the lost-parcel rule with its day count, the tone, the phrases never to use, and the always-escalate list: legal language, safety complaints, chargeback mentions, repeat claims from one address, and anything else the owner names.
The baseline is recorded at the same time from the help desk: first-reply time, tickets per order, and the share of refunds versus exchanges over the last ninety days. The Order Support FTE never asks for what onboarding already captured, so Setup is the policy and the limits, nothing else.
Step 2 (days 3–7): review the ticket audit in the Queue tab
The second step is the ticket audit, which the Order Support FTE runs on the last ninety days of tickets and reports in the Queue tab. Every post-purchase ticket is classified by type, where-is-my-order, return, exchange, order edit, damaged, missing, and by cause where the cause is visible: a tracking link sent before the carrier scan, a vague checkout estimate, a product with repeated damage claims. The audit shows what share of the queue each type is and how long each took to resolve.
The owner's job is to read it and correct the classification where the FTE has misread a ticket type, because those corrections become the rules it classifies by from then on. The audit also usually surfaces a policy gap: a case the team handled three different ways, which the owner now decides once.
Step 3 (days 8–21): approve drafted replies in Approvals
The third step is the first drafted replies in the Approvals tab. At Level 0 every reply is a draft: the ticket, the order, the carrier scan and the policy excerpt beside it, and the proposed reply below. Routine where-is-my-order replies dominate, each stating the last scan and the estimate. Return requests show the window check and the exchange offer the policy allows. Anything above a limit or on the always-escalate list does not appear as a draft; it appears as an escalation.
The owner or the support lead approves, edits or sends back with a note. Edits to tone become remit rules; a send-back that says the FTE misread the policy becomes a correction to the policy text itself, which is usually where the ambiguity was.
Step 4 (days 8–30, daily): review escalations from the morning brief
The fourth step runs daily: the morning brief in the Today tab lists escalations first, each with the thread, the order, the scan, the policy excerpt, the FTE's summary and a suggested resolution. The person decides and marks whether the escalation was correct. That mark is escalation accuracy, and over the month it is the number that says whether the FTE's judgment about what needs a person can be trusted.
The brief also shows yesterday's counts: replies sent on approval, escalations, exchanges chosen versus refunds by the customer's logged choice, and any limit the FTE approached but did not cross.
Step 5 (day 30): the review in Performance and the autonomy decision
The fifth step is the monthly review in the Performance tab. It compares first-reply time, resolutions inside policy and exchanges versus refunds against the baseline, shows escalation accuracy, and lists the questions answered most often with the upstream cause tagged. It proposes the fixes those causes imply: a checkout estimate, a notification timing, a returns page rewrite. It names anything that went wrong plainly.
The autonomy decision follows the record. Routine where-is-my-order replies with a long, clean approval record are usually promoted to send without a gate; returns and claims stay gated longer; the limits and the always-escalate list do not move. Any promotion can be reversed from Setup.
What does done look like after 30 days?
Done after thirty days with an Order Support FTE looks like a written policy where there used to be a habit, a queue classified by type and cause, routine order-status replies earning autonomy on evidence, escalations that arrive complete, and a cause report that tells the owner which upstream fix would remove the most tickets. It also looks like a remit that has grown by the rules the send-backs taught it.
Done does not look like a claim that support is solved. Claims still spike when a carrier lane fails; the difference is that the spike is reported with its tickets and its cause the morning it starts.