The support channel stopped waiting on one person

The bottleneck was a person, not a process

Every growing team has one: the person who has been there since the beginning and holds the pricing exceptions, the refund edge cases, and the reason the enterprise plan is worded the way it is. In this team, roughly sixty questions a week routed through that one head. The queue was invisible because it lived in DMs, and the cost was invisible because answering feels like helping.

The fix was not another wiki. The team had a wiki; the wiki was where answers went to age. The fix was an employee whose job is the answering itself: one that reads the same sources, watches the same channels, and puts a citation under every claim so trust does not have to be taken on faith.

Cited or refused, and why that matters

The employee answers only what the org brain can support. When a rep asked about a discount policy that had never been written down, it said so, plainly, in the channel. That refusal did more for adoption than any correct answer: the team learned the employee does not guess. The unwritten policy was written down the same afternoon, and the next person to ask got a cited answer in seconds.

Corrections behave the same way. When the answer about the enterprise SLA was outdated, the account manager corrected it in-thread; the correction became an attributed memory, visible in the ledger, and every future answer reflects it. Nobody re-teaches the same fact twice.

What changed for the human

The person who used to be the queue now reads a weekly digest: what was asked, what was answered, what was refused, and which refusals turned into new brain entries. Their job moved from answering to editing. That is the pattern we see across teams: the employee does not replace the expert, it stops the expert being an inbox.