What does hiring look like?
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 is a small software business whose support lead has chosen, under the First-Function Matrix, to hand the how-to category of the support queue to a Customer Support FTE. The lead opens the workforce view, chooses the role, and the first-run wizard opens. The company name, the website, the workspace profile, and the billing contact are already filled, each with a provenance chip: from signup, from onboarding, from checkout. The wizard asks nothing it already knows. It asks for what is new to this function: the remit, the always-escalate list with receivers, and the connections to the helpdesk and the help center.
The lead has read the playbook on writing a first remit and has the role's remit template open.
How is the remit written?
In three lists, from the template, in about an hour with the two agents who work the queue. Owns: replies in the how-to category from the help center with a citation; triage and routing; proposed articles from recurring gaps; escalation summaries. Asks: refunds and credits to the billing specialist; account access to the lead; complaints and legal language to the lead; distressed customers to the lead at any hour; tickets with no source, to the lead as a gap. Never: issuing refunds, changing account permissions, publishing articles unattended, answering beyond the cited source, and the platform's never-automated list. The lead reads it to the agents in five minutes. Every task they can think of sits on exactly one list.
The wizard stores it in the Setup tab, where it will be revised from the send-back reasons at each review.
- Owns: cited how-to replies, triage, article proposals, summaries
- Asks: money, access, complaints, distress, gaps, each with a receiver
- Never: refunds, permissions, unattended publishing, unsourced answers
How are the connections made?
The person who owns access to the helpdesk creates a dedicated account for the role with permission to read tickets and draft replies, and no permission to close, refund, or change a customer's record; the lead enters its token in the wizard, and it is encrypted and never shown again. The same for the help center, with permission to read articles and create drafts, not to publish. The lead attempts a forbidden action through the connection as the security checklist suggests, sees it refused at the tool, and records the confirmation with the date. Both connections appear in Setup with a disconnect control that would log its use.
The lead's engineering colleague was involved for the time it took to confirm two scopes.
What does the Brand Brain look like on day one?
Pre-filled from the website and from what the platform learned at onboarding: the product's name and description, the voice as the site presents it, the policies the site states, each with a provenance chip. The lead reads it in the Memory tab, corrects two things, the tone is warmer than the site suggests, and one policy has changed since the page was written, and confirms the rest. The corrections carry today's date and the lead's name. Facts the role will need but the memory lacks are marked as gaps, and the lead fills two from the internal policy document, which becomes an approved source.
The myth piece on learning your business on its own describes why this step is a reading, not a training.
What is the Day-One plan?
The wizard ends with the FTE's proposed targets for the how-to category, first-line resolution and first response time against the helpdesk's baseline, which the lead approves, and a Day-One plan in the first person: what it will do on its first morning, which tickets it will read, what it will propose and what it will escalate, and when it starts. The lead approves the plan. The Today tab is no longer empty; it shows the plan, a getting-started checklist, and purposeful empty states on the other tabs explaining what each will hold once there is work.
Every class is at Level 0. Nothing will happen without the lead until the record says otherwise.
What happens before the afternoon ends?
The FTE reads the open how-to tickets and proposes its first replies, each with a citation to a help center article, in the Approvals tab. The lead reads the first one: the ticket, the reply, the article it came from. The citation is correct and the tone matches the corrected Brand Brain. Approving it takes seconds. The second is sent back with a reason: the article cited is the right one, but the reply should mention a step the article buries; the reason is recorded as a tone item. The third ticket asked something no article answers, and it appears as a gap escalated to the lead, not as a guess.
The lead logs the day's hours on the baseline sheet, closes the workspace, and the FTE's schedule runs from tomorrow morning. The library's go-live checklist lists what the lead confirmed today; the thirty-day support playbook lists what comes next.