Intent classification
Eight categories: positive / negative / call_request / closing_request / unsubscribe / question / objection / out_of_office. Drives every downstream decision.
Every inbound is classified, then the AI generates a goal-driven reply on the same channel as the inbound. When the agent's goal is reached (call requested, demo booked, deal close-ready), auto-reply pauses itself and the rep is notified, never before.
Eight categories: positive / negative / call_request / closing_request / unsubscribe / question / objection / out_of_office. Drives every downstream decision.
Each agent has one goal (BOOK_CALL / BOOK_DEMO / QUALIFY / CLOSE / NURTURE). The reply nudges toward that goal with proof, urgency, or a calendar link.
When the AI hits the goal, it pauses further auto-replies on the lead, writes a ProspectActivity row, and notifies the rep on Slack + in-app + email.
You choose how the AI handles negative replies · either it stops on the first firm no (default), or it tries one more time with a negative-streak cap you set (3 by default).
Cap auto-reply at N turns or leave it Unlimited (continues until the goal is hit). One toggle in the wizard.
Your choice per agent. 'Unlimited' means the AI keeps replying until the goal is reached, then auto-reply pauses and the rep is notified. Bounded mode hands off after N replies, OR earlier if the goal is reached.
By default the first firm negative reply (or an unsubscribe) stops the AI immediately and notifies your rep. You can also let the AI try one more time, with a negative-streak cap you set, to give a last shot before handing off.
No. The meeting link comes from your agent config (e.g. https://cal.com/you/15min). The AI inserts it verbatim into the reply when the call_request or closing_request intent fires.
Trial on request · contact our sales team.