GTM/REVENUE AGENTS · وكلاء

Your OS doesn’t just advise — it ships the work.

GTM/Revenue Agents are generated from RIVET INFINITE OS. Each one inherits the same eight models, the same CR1→CR8 pipeline and the same shared context — so it doesn’t guess your motion, it runs it.

  • 33 agents. One context.
  • A level, an owner, a review.
  • The agent prepares. You decide.

Count the nodes: 25 on top, 8 underneath. The eight are why it’s a workforce, not a crowd. Now you know.

The roster

Meet the 33.

Pick a track. Each chip is the agent's own ceiling: the most it may run on its own. Your motion and your AI provider can lower it, never raise it.

Prospecting track download cover
Prospecting · CR1–CR2

10 agents, one zip.

Open the conversation: target, map, brief, first touch.

  • Cover of the TAM Agent agent
    TAM Agent
    Ceiling · Autonomous

    Sizes your market as TAM, SAM and SOM, split into Priority A, B and C for the team you have.

  • Cover of the ICP Scout agent
    ICP Scout
    Ceiling · Autonomous

    Scores every account to your rubric with checkable facts and finds more of your best customers.

  • Cover of the Target List Builder agent
    Target List Builder
    Ceiling · Autonomous

    Builds the named account list, gates first, so you only ever talk to a reviewed list.

  • Cover of the Prospect Prep agent
    Prospect Prep
    Ceiling · Autonomous

    A one-page brief before every first touch: about 40 minutes of prep in about 5.

  • Cover of the Signal Watch agent
    Signal Watch
    Ceiling · Autonomous

    The night shift: fresh, dated signals on your accounts, each mapped to the play it triggers.

  • Cover of the Account Mapper agent
    Account Mapper
    Ceiling · Agentic

    Maps who really decides inside an account, and names the one gap to close first.

  • Cover of the First-Touch Writer agent
    First-Touch Writer
    Ceiling · Orchestrated

    Drafts the first touch in your voice and their language, after the gates. You send.

  • Cover of the Outbound Sequence Writer agent
    Outbound Sequence Writer
    Ceiling · Orchestrated

    Turns a play into a dated sequence for one person, with the ask built in. You send.

  • Cover of the Cadence and Reply Handler agent
    Cadence and Reply Handler
    Ceiling · Orchestrated

    Answers replies first, honours every no, and drafts the right next move. You send.

  • Cover of the Inbound Qualifier agent
    Inbound Qualifier
    Ceiling · Orchestrated

    Every inbound answered the same day, scored, routed, and the reply drafted. You confirm.

Next in the library · taught in the course

Taught now. In the library next.

The course already teaches these three. Their files join the library next; until then their cards are dashed.

Generated, not assembled · من نظامك

An agent is only as good as the OS it’s cut from.

Prompts live in one rep’s chat history. These agents live in your operating system — so they know your models, work a named stretch of the pipeline and write to the record everyone reads.

Knows your models

It reads the same eight you run.

Revenue, Growth, Motion, Data, Communication, Moat, AI, Operation — so an agent never re-asks what kind of company you are.

RevenueGrowthMotionDataCommunicationMoatAIOperationRevenueGrowthMotionData
Owns a stage

One stretch of CR1→CR8.

Each agent owns a stage and hands the next a finished packet — not a to-do.

One record

The buyer is never asked twice.

Every agent reads and writes the same shared context.

Governed like a team

Level. Owner. Review. Log.

Four things every agent carries — the difference between a pilot that works in the demo and one that is still running in week three.

LevelOwnerReviewLogLevelOwnerReviewLogLevelOwnerReviewLog
How far an agent may run · الاستقلالية

Autonomy is a ceiling, not a switch.

Five levels, from Manual to Autonomous. An agent runs at the lowest of three ceilings — its own, your motion’s, and your AI provider’s. Drag them and watch which one holds it down.

  • Nothing starts above Assisted.
  • Climb the ladder and your job changes — it never disappears.
  • The weakest ceiling wins, every time.
This agent runs atOrchestrated

The weakest of three sets the pace. Now you know.

  • Human-onlyMove money
  • Human-onlyUse credentials or identity
  • Human-onlyRelease external approvals
  • Human-onlyMake commercial commitments

At any level, the agent prepares. You decide.

A relay, not a crowd · التسليم

Each agent’s output is the next one’s whole input.

Take the Prospecting track. Seven agents, one chain — market to first touch — and a dashed handoff into Selling where the next track picks it up. Hover a node: the one after it lights, because it starts exactly where this one stopped.

  1. CR1TAM Agent

    Turns your ICP into a sized, segmented, tiered addressable market.

    Hands over → tiered segments
  2. CR1Target List Builder

    Builds and tiers account lists from your ICP.

    Hands over → tiered account lists
  3. CR1ICP Scout

    Researches accounts, scores them to your rubric and finds lookalikes.

    Hands over → scored, researched accounts
  4. CR2Account Mapper

    Maps the buying committee — the seven roles — for every Tier-1 account.

    Hands over → stakeholder map
  5. CR2Prospect Prep

    Writes the overnight brief on every account before you touch it.

    Hands over → account brief
  6. CR2First-Touch WriterA human sends

    Drafts a personalised opener in your voice.

    Hands over → draft opener
  7. CR2Cadence & Reply Handler

    Runs the follow-ups and sorts the replies.

    Hands over → a qualified conversation
  8. CR3Selling track

    The conversation crosses tracks — Discovery starts with the packet, not a blank page.

    Handoff between tracks

The core chain of Prospecting’s ten. Voice Profile, from the shared eight, grounds every node that writes.

Inside RIVET Labs

Every agent has its own lesson.

Open an agent in the space and its lesson walks you through the setup and the handoff. These are real pages from Cohort 1.

  • Agent Operating Protocol lesson page in the RIVET Labs space
    Agent Operating ProtocolThe rulebook every agent runs on.
  • ICP Scout lesson page in the RIVET Labs space
    ICP ScoutA Prospecting agent.
  • Health Radar lesson page in the RIVET Labs space
    Health RadarA Customer Success agent.
  • Resource Hub lesson page in the RIVET Labs space
    Resource HubThe setup, in the order we teach it.
Now you know · الآن تعرف

Don’t just read about agents. Run your track’s.

AI Workspace GTM/Revenue Engineer: eight sessions over two weeks where you run these agents on your own accounts.

Needs a paid AI assistant — Claude (preferred), ChatGPT, Gemini or Copilot. Free tiers are not sufficient.