Loading
Because an unsourced draft has to be verified line by line. Checking a plausible paragraph against reality takes longer than writing three sentences you already knew.
Agent assist drafts a reply to an open ticket from your knowledge base, the customer's record and the ticket history.
Each claim in the draft cites the article or record field it came from. Where there is no source, it says so rather than filling the gap, and the agent edits and sends.
A ticket spanning a billing question and a delivery date drafts with the invoice position from Books and the milestone from Loop, each linked, so checking takes seconds.
It will not send. There is no auto-send setting for agent assist, because the value is a person with context checking before the customer sees it.
Knowledge base articles, the customer's entitlement, their history, and the current position in Flow, Books and Loop.
It weights the articles and phrasings your team actually keeps.
Yes, from the same sources, in the customer's language.
It says so and drafts nothing rather than inventing an answer.
Fourteen days, every module, no card. Or half an hour with someone who will run it on your own records and tell you where it does not help.