What happened overnight?
This is an illustrative scenario, not a customer account. The people and the company are invented; every event in it maps to a real mechanism on ftes.ai, named as it happens.
The company sells a software product with a small support team that used to arrive to a full overnight queue. The Customer Support FTE, connected to the helpdesk and the help center through a dedicated scoped account, read each overnight ticket and did what its remit allows. Order-status and password-reset questions in the one category promoted last month were answered and sent within bounds, each reply citing the article or the account state it drew on. How-to questions in the still-gated categories were drafted with citations and queued for approval. Three tickets asked a question no article answers; each was escalated as a knowledge gap rather than guessed.
One overnight message contained the word complaint and a description of a failed renewal. It matched the Asks list and went nowhere near a reply.
What does the support lead see first?
The Today tab, waiting-on-you first. The complaint sits at the top as an escalation with the four-part summary: what the customer asked, what the FTE found in the account and the help center, why it did not resolve, meaning it is a complaint about billing, and a suggested next step, which is to check the renewal record before replying. Below it, the queued drafts for the gated categories, then the three knowledge gaps grouped into one proposed article, then yesterday's numbers and the spend line.
The lead reads the complaint summary, opens the renewal record, and hands the ticket to the billing specialist with the summary attached. Nobody re-asks the customer for their account details.
How do the approvals go?
In the Approvals tab the queued drafts show the ticket, the reply, and the citation. The lead approves most in seconds. Two are sent back: one because the tone is more formal than the Brand Brain's voice, which is a tone reason and will show in the send-back log as such; one because the cited article is out of date since a feature changed last week, which is a facts reason and prompts the lead to correct the article in the help center, where it will be re-read as a source. The proposed article for the recurring gap is read, edited for one sentence, and approved; it publishes to the help center through the connection and becomes a source for the next reply.
The send-back reasons are what the lead will read on Friday, sorted, to see what to change.
- Approve: cited drafts, seconds each
- Send back: one for tone, one for a stale article
- Article: approved from three gaps, published, now a source
What does the team do with the afternoon?
The hard cases. The billing specialist resolves the complaint and, because the customer's frustration was about a policy the help center states badly, proposes a wording change the lead will take to the product team. An agent who used to spend mornings on password resets spends the afternoon on an account migration a customer needs walked through, starting from the FTE's escalation summary. The lead reads the quality review the FTE drafted overnight, replies scored against the team's rubric with itemised notes, and notices a pattern about one policy that two agents describe differently; the coaching note is drafted for the lead to deliver, not sent by the system.
Nobody on the team received feedback from a machine. The lead received a pattern.
What happened when a ticket mentioned a refund?
Mid-afternoon a ticket arrived asking for a refund on a duplicate charge. Refunds are on the Asks list at every autonomy level; the FTE gathered the order and payment records, wrote the four-part summary, and escalated to the person who owns refunds, with a holding reply from the approved template sent to the customer promising a person would respond. The refund decision, the reply, and the credit were the person's. The FTE's part was to make sure the person started with the evidence rather than the inbox.
The always-escalate list did not shrink because the record was clean. It never does.
What does the record show at the end of the day?
The audit trail: every auto-sent reply in the promoted category with its citation, every approved draft with the lead's name, both send-backs with their reasons, the article's approval and publication, the complaint's and the refund's escalations with their summaries and receivers. The Performance tab: first-line resolution per category for the day, escalation quality, and the knowledge gaps closed. At the monthly review the lead will read these against the baseline and decide whether the how-to category has earned the Promotion Standard's consideration, and which category comes next.
The library's piece on how a Support FTE works describes each mechanism outside the story; the escalation design piece describes the summary standard the complaint followed.