Run it · The roster
Five specialists and a coordinator, in a fixed order. The order is the motion.
The workspace includes GTM CEO as coordinator, plus Leads, Intent, Outbound, Copy and QA specialists. Each role has a defined toolset and responsibility. Campaigns reuse these roles.
Idle
Resolves the ICP, pulls the matches in enriched
Watching
Reads what fired, decides who just got warm
Sending
Writes, sends, handles the reply, books the call
Ready
Writes in a human voice, rewrites what QA flags
Checking
Checks drafts against their evidence
What is actually going on
Every claim below is followed by the mechanism under it, because a claim without a mechanism is a slide.
Only one of them can send
The outbound agent holds the send tool. Nothing else in any belt reaches a person, which is what makes the pre-send gate the single path rather than one of several.
Identity is a column, not a prompt fingerprint
A preset id plus a partial unique index on organisation, workspace and role. That is what lets every prompt be editable without the self-heal losing track of which agent is which under concurrent page loads.
They are also tools on the agent screen
The assistant you talk to delegates to them the way you delegate to it. A slash command is the same delegation with the choice already made.
A sub-agent's belt is a subset of yours
Filtered by your real membership first, then by the role's allowlist. Delegation can never reach a permission you do not have, and it is one level deep and acyclic.
Step by step
How the agent roster runs
- 1
Leads
Resolves the ICP into a query, pulls the matches in enriched. Daily. Does not send.
- 2
Intent
Reads what fired, watches contacts' own reactions, decides who just got warm. Every minute. No send tool at all.
- 3
Outbound
Writes, sends, handles the reply, books the call. Every minute. The only one that sends.
- 4
Copy & Quality
Copy writes in a human voice and rewrites what quality says is not landing. Quality checks drafts against evidence and cannot release one.
The numbers
Facts, not adjectives
Every figure here is a real value from the running product. A cadence from the scheduler, a price from the credit table, a limit from the ramp. A marketing page that invents a limit is a support ticket with a delay on it.
- Roster size
- Five, fixed
- Plus
- GTM CEO coordinates the five specialist roles
- Cadences
- Daily, 15-minute, hourly, 6-hourly
- Editable
- Every prompt and every label, with a reset
- Delegation
- One level deep, acyclic, subset belts
The agent roster, honestly
How is the coordinator different from the specialists?
GTM CEO coordinates the workspace, while Leads, Intent, Outbound, Copy and QA have separate execution roles. Campaigns reuse that roster.
What can I actually change?
Every prompt, every piece of display copy, the schedule, the model tier and the send policy. What you cannot change is the hidden platform layer above your prompt. No invented claims, no implied prior contact, no arguing with the gate.
Run it
The team that does all of the above
Keep reading
All featuresThe agent roster is one part of it. The rest runs too.
Sourcing, enrichment, intent, copy, sending, replies and reporting are one product with one bill and one gate. Connect a channel and the whole motion starts.
Start with sales. $2,000 or $5,000 onboarding, then at least $250 monthly credit funding.