Every reply, read and classified

Every reply, read and classified

Sending a third touch to a prospect who already answered the first signals that nobody is paying attention, and it costs the conversation you had just opened.

Reply detection is built to prevent that. Every message that arrives in a connected mailbox is checked against the threads your sequences have sent. A match pulls the lead out of the sequence at once, classifies the reply, and places it in front of you with the noise removed.

How detection works

Connected mailboxes are polled over IMAP about every minute. When an inbound message is matched to a thread you sent, the lead is removed from its sequence before the next step can fire, so no further touches go out.

Quoted history is stripped from the reply shown in the app. You read what the prospect wrote, not several screens of your own email quoted back.

Detection runs on every mailbox you connect: Gmail or Google Workspace, Outlook or Microsoft 365, Zoho, or another IMAP provider.

The eight classifications

Not everything that arrives is a person answering you. Each reply is classified into one of 8 types, and the type determines what happens next.

• Interested: the lead is stopped and surfaced for you to handle, or answered by autopilot where you have enabled it.
• Not interested: the lead is stopped, and you can add the address to the permanent do-not-contact list in one step.
• Out of office: treated as a delay rather than an answer; the follow-up is rescheduled around the return date.
• Bounce: the address is marked as undeliverable and sending to it stops.
• Auto-reply: acknowledged and excluded from reply statistics.
• Unsubscribe: honored on the spot; the address joins the suppression list that spans every sequence and mailbox.
• Referral or hand-off: the prospect points you to a colleague; the thread is flagged for you to follow up with the right person.
• Already booked: the prospect confirms a meeting is set; the sequence stops and the thread is marked so no further times are proposed.

What happens next

An out-of-office reply is parsed for the date the person says they are back. The next touch is held and resumes on the next business day after that date, so it lands when the prospect is at their desk rather than at the top of the backlog they return to.

A hard bounce marks the address and stops sending to it. The lead is removed from the sequence.

An unsubscribe request is honored immediately. The address joins the account-wide suppression list described on the Sequences page.

Bounces, out-of-office messages and auto-replies are never counted as replies in any statistic in the app. Reply figures on the outcomes dashboard and on each sequence reflect only messages written by a person.

Unified inbox and conversation view

Replies from every connected mailbox appear in a single inbox. Each item shows its classification, the sequence and step it came from, and the mailbox that sent the original message.

You answer from the inbox, in the original thread, from the mailbox that sent the first email. The conversation stays in one place for the prospect.

A full conversation view is available for every contact. It shows each touch sent, each reply received and each classification applied, alongside the research notes and lead record.

Auto-reply and booking

Auto-reply and auto-book are per-sequence settings and are off by default. Nothing is answered on your behalf until you turn them on for a specific sequence.

When enabled, autopilot answers interested replies in your voice, using the writing controls set on your workspace. It offers meeting times drawn from your real calendar through a scheduling integration, books the meeting when the prospect agrees, and hands the thread to you whenever the reply calls for a judgment the system should not make alone.

Cold first touches never propose meeting times and never carry marketing links. Times are offered only once a prospect has shown interest.

Scope

• Detection covers the mailboxes you connect, and each mailbox needs IMAP enabled on the provider side.
• Replies are picked up by polling on a short cycle, not by instant push. Expect a reply to appear within about a minute, not the same second.
• Classification is a model judgment and can be wrong. Every reply is still shown in the inbox, so a misclassification is visible and can be corrected rather than lost.
• A reply sent from a different address than the one you wrote to, or a thread forwarded on by the prospect, may not be matched to the original send.
• Reply detection is not a helpdesk. There are no ticket statuses, assignment rules or response timers.

Credits

Reading, matching and classifying replies costs nothing. Each reply generated by autopilot costs 1 credit. Replies you type yourself are free and sent verbatim.

Credits are included with every plan, granted monthly or as a full year on annual billing, and top-up packs are available. Allowances are listed on the Pricing page; the full credit schedule is in the Docs.