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 is a consumer brand with one brand manager who used to write every post. The Social Growth FTE, connected to the brand's accounts through dedicated scoped credentials, drafted the week's posts from the content calendar the manager approved at the last review: for each idea, a separate native draft per platform, in the Brand Brain's voice, with every product claim checked by the fact gate. One draft was held because it stated a material fact about a product that no approved source supported. It also read the overnight comments and proposed replies to the routine ones, thanks, questions about availability, within the reply rules.
One comment named a member of staff and described a bad experience in a store. That matched the Asks list and became an escalation, not a reply.
What does the brand manager see?
The morning brief, first person: drafts waiting per platform, replies waiting, one held post with the reason, one escalation, yesterday's published posts, and spend. The escalation sits first: the comment, what the FTE found, why it did not reply, which is that it names an individual and describes a complaint, and a suggested next step, which is a personal reply from the manager and a note to the store. The manager replies personally and forwards the note. The FTE will not touch that thread again.
Then the drafts. The manager approves most, sends one back because the platform-specific version sounds like the other platform's, a tone reason, and reads the held post.
How is the held post released?
The held post's claim is true; the manager knows it from the product team. The fix is not to approve the post anyway; it is to add the product specification document to the approved sources in Setup, after which the fact gate re-checks the claim and clears it. The manager then approves the post. The Brand Brain gains the fact with today's date and the manager's name as provenance, so the next post about the product will not be held for the same reason. The whole exchange took a few minutes and left a record that shows the claim was sourced before it went out.
The brand safety piece in the library explains why the gate is the mechanism, not the manager's memory.
- Held: unsourced product claim
- Fixed: source added, gate re-checked, approved
- Memory: fact recorded with provenance
How do the replies go?
The proposed replies show the comment, the reply, and the rule they were drafted under: availability questions answered from the approved facts, thanks acknowledged in the voice. The manager approves them in a minute. Replies in the routine class are still gated at Level 0; the manager has considered promoting them under the Promotion Standard and decided to wait one more cycle because the send-back reasons still include tone. A comment asking about a price was not proposed as a reply; prices are on the Never list, and it went to the manager with a suggested holding line from the approved template.
The reply rules do not loosen because the account is busy. They are reviewed quarterly and after incidents.
What does the manager do with the afternoon?
Reads the Performance tab: posts published per platform, engagement where the platform's connection reports it, reply approval rate with reasons, escalations by reason, and the unmeasured list, which includes reach on a platform whose connection does not report it and any attribution to sales, stated plainly. Plans next month's calendar from the FTE's proposals, declining two and approving the rest. Then does the work only a person does: the campaign concept with the agency for the seasonal launch, which the FTE will execute as posts once the concept is approved.
The agency's retainer now covers the concept. The daily posting is the FTE's.
What does the record show at the end of the day?
The audit trail: each draft per platform with its approval, the tone send-back, the held post with the fact gate's reason and the source change that released it, the replies approved, the price comment routed, and the escalation with its summary and the manager's personal reply noted. The Performance tab's numbers per platform with the unmeasured items labelled. At the monthly review the manager will read the reply send-back reasons and decide whether routine replies have earned the Promotion Standard's consideration.
The library's piece on how a Social Growth FTE works describes each mechanism outside the story; the brand safety piece describes the reply rules and the escalate list.