> ## Documentation Index > Fetch the complete documentation index at: https://docs.getswan.com/llms.txt > Use this file to discover all available pages before exploring further. # Replies and stop conditions > What finishes, stops, skips, or errors out a contact's progress through a sequence. A contact doesn't always run straight through every step of a sequence. Here's what changes its course, and why. ## A reply finishes the sequence If a contact replies — on email or on LinkedIn, on any step — Swan detects it through the connected sending account and immediately marks that contact **finished**. No further steps run, including on the other channel: a LinkedIn reply finishes email steps too, and vice versa, since a reply on any channel means the person is engaged. Swan doesn't auto-reply. It notifies the assigned team member and stops so a human can take over the conversation. This applies even to a short or non-committal reply (an emoji reaction, for example, still counts as a reply). Out-of-office and similar automated replies are filtered out and don't trigger this. A reply can also fire your [message received trigger](../triggers/message-received) — a personal trigger created automatically when you connect a sending account, which you can edit or turn off — so Swan can react to the reply beyond just stopping the sequence. Without one, replies still stop the sequence as described here; Swan just doesn't act on them. ## Manual contact stops the sequence If the sender manually messages the contact outside of Swan — for example, replying to them directly from their own inbox before Swan's next scheduled step — Swan detects that outgoing message and marks the contact **stopped**, with a reason indicating it was already contacted manually. This prevents Swan from sending a duplicate or conflicting message on top of what you just sent yourself. ## Connection requests no longer block the sequence Sending a LinkedIn connection request doesn't pause a contact anymore. As soon as the invite is sent, the sequence moves on to whatever comes next — Swan doesn't wait around to see if it gets accepted. The wait only happens if the sequence reaches a **LinkedIn message** step and the contact still isn't connected yet: * **Already connected by then** — the message sends right away. * **Not yet connected** — the contact waits at that message step for the invite to be accepted, for up to 2 weeks from when the invite was actually sent (LinkedIn itself expires invites around then). * **Accepted within that window** — the message sends. * **Rejected, or not accepted within the 2 weeks** — the LinkedIn message step (and any further LinkedIn steps) are skipped for this contact. If the sequence also includes email steps, those keep running normally. ## Per-channel failures Email and LinkedIn are tracked separately for each contact. If something goes wrong on one channel — a rejected or expired connection request, LinkedIn refusing to let Swan invite someone, a missing LinkedIn URL — Swan skips the rest of that contact's steps on that channel only. Steps on the other channel keep running. A contact only lands in an **error** state once every channel its sequence actually uses has failed. ## Recently invited on LinkedIn recovers automatically If LinkedIn blocks Swan from sending a connection request because the contact was already invited recently, Swan doesn't just give up. It first checks whether it already has a pending invitation to the same person from another sequence and reuses that, then checks the sender's own sent-invitations list on LinkedIn for a matching invite. Only if neither of those turns anything up does Swan mark the LinkedIn channel failed for that contact — and even then, email steps in the same sequence keep going. ## Missing contact info skips incompatible steps If a contact is missing a work email or a LinkedIn URL, or the assigned sender doesn't have the matching channel connected, Swan can't run steps that need it. Rather than erroring out, it skips ahead past incompatible steps to the next one it can actually run — for example, skipping a LinkedIn message step for a contact with no LinkedIn URL and going straight to an email step. ## Disconnected senders pause their contacts If a sender's email or LinkedIn connection is disconnected, every contact currently assigned to it is paused (stopped, with a reason noting the sender was removed). Reconnecting that same account automatically resumes those contacts from wherever they left off — you don't need to re-enroll them. ## Errors Some issues aren't recoverable automatically and mark the contact as **errored** instead of pausing or finishing it — this happens once every channel the sequence uses has failed for that contact (for example, an invalid LinkedIn URL on a LinkedIn-only sequence, or both a rejected connection request and a bounced email address on a sequence that uses both channels). These need a person to fix the underlying issue before the contact can be manually restarted. ## Common questions **If someone replies to step 2 of a sequence, does step 3 still send?** No — a reply on any step, on any channel, finishes the sequence for that contact. No further steps run. **Does an out-of-office reply stop a sequence?** No, Swan filters out automated out-of-office and similar replies so they don't count as a real reply. **What happens if a contact never accepts the connection request?** Nothing pauses while Swan waits for it — the sequence keeps moving. Only once it reaches a LinkedIn message step does Swan check: if the invite is still unaccepted after 2 weeks (from when it was sent), that LinkedIn message and any later LinkedIn steps are skipped, while email steps in the same sequence continue. **Why did a contact disappear from a sequence I didn't touch?** Check its status reason at [Outreach → Sequences](https://agent.getswan.com/outreach/sequences) — open the contact's row to see it — common causes are a reply, the sender being disconnected, the sequence being deleted, or the contact being excluded. See status meanings in [reference/statuses.md](../reference/statuses). **If I disconnect and reconnect my LinkedIn or email account, do my sequences pick back up automatically?** Yes — contacts that were paused because the sender was disconnected resume automatically once you reconnect the same account. **What should I do with a contact marked as errored?** Fix the underlying issue (for example, correct their LinkedIn URL) and manually restart them — errored contacts don't retry on their own.