In a live room, five to ten percent of attendees ever type in the public chat. A setter who stares at that feed, waiting to catch a signal, misses everything else. The cockpit replaces that with three panels: public chat, notification center, private messaging. Every message tags itself and every hot lead surfaces at the right moment. The setter still writes every word.
A setter can't stop to analyze mid-live. So the AI does it for them, re-reading the room every fifteen to thirty seconds: public messages, private messages, time connected, clicks on the offer, poll answers, forms, quizzes. Every attendee carries a score that updates in real time. The setter never sets that score by hand. They act on what the cockpit surfaces, hottest lead first.
The role has a name: Moderator-Setter. Never closer, that word got dropped. While the host runs the pitch for the whole room, the setter works each hot lead in DM, one at a time, all the way to the booked call. Mass broadcast for the room, one-on-one for the lead.
The public chat runs down the left. Private messaging sits on the right, the open conversation in view. In the middle: the notification center, where the setter actually spends the live. No more bouncing between public and private, hoping nothing slipped past. Urgent actions sit up top, ranked, so the thing that has to go out right now is the first thing they see.
Every notification arrives with the information, the lead's context, and a suggested action. Here's what shows up today:
On a hot lead who just left, the notification carries a one-click WhatsApp link to catch them outside the app. Payment itself happens on the client's own tool, outside the live. But every step of the buying journey comes back into the cockpit: the second someone stalls on a payment step, the setter knows, and decides what to write.
Each notification carries four actions: reply privately, reply publicly, mark handled, ignore. Handled and ignored aren't decorative. That pair trains the models and sharpens what surfaces next. One thing never moves: the system detects and suggests, the setter writes the message. Nothing goes out on its own.
Every message in the public chat gets tagged on the fly: price objection, question, buying intent, hesitation, spam. In a room typing fast, the setter sorts at a glance instead of reading line by line. This one needs a public reply, that one a private DM. The rest needs nothing, and gets nothing.
Spam gets the shadowban: the disruptor keeps typing, convinced they're still visible, while nobody else sees a word of it.
Before opening a private conversation, the setter pulls up the attendee's profile. It carries a timestamped timeline of everything they've done since the live started: what they wrote, how they answered the polls, what they clicked, when they left if they did. The setter types the first message already knowing who they're talking to. They're not writing to a stranger.
The ratio we recommend: one setter per three hundred to five hundred registrants. On bigger rooms, several setters work in parallel, and notifications get routed between them so no lead gets worked twice.
We look at your format, your volume, and what the tool would change for you. No strings attached.