Published:
How Arlo handles a stop request
Someone writes back: “Please take me off your list.” This is what happens next, and why none of it rests on a language model’s judgement.
Found by fixed phrases, not guessed
Arlo reads every reply for explicit opt-out wording in German and English: “unsubscribe”, “remove me”, “take us off”, “bitte austragen”, “keine weiteren E-Mails” and many more. The check is a fixed list of phrases, not a model’s reading. The same reply always gets the same answer, and the check runs the moment the reply comes in.
Only the person’s own words count. Your earlier email quoted below their answer, and the opt-out line in its footer, are set aside first. Otherwise every reply that quotes the footer would read as a stop.
The list errs in one direction. If the phrase is there, Arlo stops, even when the rest of the reply sounds friendly. Stopping once too often costs one prospect. Missing a real stop breaks a promise.
What stops, and where
A stop request applies at once and everywhere: every sequence to that person, from every mailbox your team connected. Planned emails to them are cancelled. The address goes on your do-not-contact list under Settings → Do not contact, marked “Permanent”: that entry can’t be removed, not by your team and not by Arlo.
Arlo checks that list twice: when it plans an email, and again right before sending it. A stop that arrives while an email waits for your approval, or after you approved it, still wins.
One honest limit: an email that is already being sent at the moment the stop arrives still goes out. Nothing after it does.
Softer answers are replies, not stops
“Not interested” or “not right now” is a reply, and Arlo treats it as one. Any reply from a company pauses Arlo’s other sequences at that company, so nobody there gets a follow-up while a colleague’s answer is still open.
When Arlo can’t tell which conversation a reply belongs to, it doesn’t guess. It asks you, under Conversations → Waiting on you.