The monitor counts
Minimum sends, minimum reads, all recipients contacted, or a keyword in a report.
Run an objective on a CRM list over WhatsApp, email, or phone. You pick a library template, preview the Organization wallet, then approve the send. Success is sends, reads, or contacted — not a booked appointment.
A Project-scoped list task: objective, audience, channel, library template, then approve. The planner does not create templates. Adaptive nurture is not shipped.
Minimum sends, minimum reads, all recipients contacted, or a keyword in a report.
Appointments, pipeline Bookings, or “Interested.” A human or a workflow still moves the card.
Three runtimes. They do not share a stop-when-booked contract.
Live threads in the omnichannel inbox: skills, knowledge bases, router, human handoff.
Clock-based broadcasts. Conversion is reply rate unless you track a pipeline stage.
Agents → Autonomous. Activity shows the Autonomous tab. Replies still land in Inbox.
The product surface is the task, not a vertical clone of a playbook.
The brief the planner compiles. Playbooks in product are inferred from objective text in places; they are not a hidden closer.
CRM query, appointment query, or deal query. CSV import is v1 if the list is still a spreadsheet. No marketplace-to-task handoff.
WhatsApp, email, and phone only for autonomous outbound. Instagram, web chat, RCS, and SMS stay on campaigns, workflows, or inbox.
You pick an approved artefact. Closed-window WhatsApp is template-only; the planner does not auto-pair session plus template follow-ups.
Minimum sends, minimum reads, all contacted, or report keyword. Not appointment booked. Not site visit booked. Not token paid.
Agents → Activity unifies AI replies, workflow runs, and autonomous tasks so ops can see which runtime fired.
The job write-up is cold-buyer win-back. This page is the mechanism.
Use this when someone asks the autonomous agent to “just book the visit.”
| Runtime | Trigger | What success means today | Does not mean |
|---|---|---|---|
| Autonomous agent | Approved plan on a list | Sends, reads, contacted, or report keyword | Appointment or site visit booked |
| Conversation agent | Buyer messages inbox / voice | Replies, fields, handoff | A 2,000-row blast |
| Growth campaign / drip | Clock + matched audience | Reply rate labeled conversion | First-class Interested or Bookings |
| Workflow | Event (e.g. stage changed) | The actions you configured | A scheduled list objective |
| Appointments / calendars | Slot confirm on a thread | A real booking | Autonomous monitor output |
What the product does not do today.
Pick an existing library item. The planner does not create or submit WhatsApp templates.
Closed-window follow-ups are not auto-paired as session plus template.
No campaign-reply or marketplace list into the task. CRM, appointment, or deal query.
WhatsApp (Gupshup), email, phone. SMS is campaigns and workflows when MSG91 is connected.
Phone “replied” elsewhere is speech-to-text, not intent. Campaign conversion is reply rate.
Preview the Organization wallet before you approve a large list. No seat tax.
Operators wiring Agents → Autonomous for the first time.
Not today. Success specs are minimum messages, minimum reads, all recipients contacted, or a report keyword. Move Bookings on the pipeline with a human or a workflow. Appointments and calendars remain the slot truth on live threads.
No. Clarify asks you to choose an existing library template. Submit utility or marketing copy in the library via Gupshup first. AI draft there is text-only; you upload IMAGE headers.
Not as an automatic source. Audiences are CRM query, appointment query, or deal query. Campaign engagement and marketplace leads do not flow into autonomous tasks yet.
No. Drips are clock sequences with reply-rate conversion. Autonomous tasks take an objective, clarify a plan, and require approve. Neither adapts from last week’s outcomes. Staff the omnichannel inbox for replies either way.
WhatsApp, email, and phone. Not Instagram, not web chat, not live SMS. Connect those in Setup → Channels (Gupshup, SES, Plivo). Campaigns support a wider channel set than tasks.
Agents → Activity, Autonomous tab, alongside AI responses and workflow runs. Human handoff still happens in the inbox, not in the task monitor.
No. Wallet-only pricing. Usage (messages, minutes, AI) meters from the Organization wallet. Details on pricing.
Mechanisms here; jobs on playbooks; how-tos on guides.
Studio, skills, knowledge, router — the inbox runtime.
Agents →Event-driven stage change, CRM properties, start agent.
Workflows →Clock-based sends. Conversion is reply rate today.
Campaigns →Where replies and human handoff actually happen.
Inbox →Omnichannel playbook — not a vertical clone of this page.
Playbook →Organization wallet. Preview before approve.
Pricing →Connect Gupshup, SES, or Plivo, pick a library template, preview the Organization wallet, and keep Bookings on the pipeline.
Get started free