Most automated outreach tools are built to keep going. More messages, more follow-ups, more touches — because more activity looks like more effort. That's exactly backwards, and it's the thing we designed against from day one.
The rule
The instant anyone replies to an outreach message — a rejection, a question, an offer to help, anything — every future scheduled message for that job stops immediately. Not at the end of the day, not after a review queue. Within one processing cycle of the reply arriving.
This isn't a feature toggle. It's the core safety mechanism the entire product is organized around.
Why this matters more than it might seem to
Picture the alternative: someone tells you, politely, that the role isn't a fit — and your outreach agent keeps messaging two more people at that company on the schedule it already had. That's not just an inefficiency. It's the exact behavior that makes automated outreach feel disrespectful to the people on the other end, and it's the fastest way to burn a company's goodwill toward you before you've even had a real conversation.
A hard stop protects two things at once:
- Your reputation with that company. Nobody wants to be "the person whose bot kept emailing us after we said no."
- The person who replied. If they took the time to respond — even with bad news — the least you owe them is not getting three more messages from your account that week.
How we make sure it actually works
A reply gets classified almost immediately — is this a rejection, a positive response, a genuine question, or just an auto-reply that shouldn't count as a real answer? Getting this right matters enormously, because a missed rejection means outreach keeps running exactly when it shouldn't. We deliberately tune this to over-stop rather than under-stop: if a message is ambiguous between "no" and "maybe," we treat it as a stop. A slightly early stop costs you nothing. A missed rejection costs you a company's patience.
What it doesn't do
A hard stop isn't a guarantee that nothing else ever reaches that company — if two different people are in an outreach sequence and one hasn't replied yet, only the specific thread tied to the response is affected in the way you'd expect: everything scheduled after that point, for that job, stops. What's already been sent can't be un-sent, and we're upfront about that limitation rather than pretending otherwise.
The goal isn't a perfect system. It's a system that fails safe — where the default, the instant something changes, is to stop and let you decide what's next, not to keep running on autopilot.