gomarketagent
New — Promoter agent for paid GTM

Promotion agents
that respect your voice.

Drafts, prospects, replies, and paid-ad campaigns — generated in your project’s voice, queued for your review. Nothing ships until you approve it.

Sign inHow it works

Provisioned by invite — get in touch about access.

draft essaysubstack~1500 words
just now

Anchor essay

Deploying AI with intention

We are not a productivity service. We are not a course. We are a small group of practitioners working with a small group of clients, deliberately. The work begins with one question: what would a quiet week with AI look like for you?

Voice check

0.92

passes register

Why GTMAgents

Built for teams who write deliberately.

Three principles you can’t paste into a system prompt. Architectural, not aspirational.

01 — Voice

Your register, embedded

Load 5+ on-register samples per project. Every draft, prospect note, and ad headline is matched against the corpus and scored before it reaches the queue. No SaaS theatre unless your brand asks for it.

02 — Approval

Human in the loop, always

Nothing publishes. Nothing sends. Every output lands as an artefact in the approval queue with the full agent plan, the model used, and the voice score. Approve, edit, or reject — the agent learns from rejections.

03 — Channels

Every GTM lever, one queue

Anchor essays, prospect outreach, paid-ad strategy, inbound replies — the same approval flow, the same project-level guardrails, the same voice. One place to oversee everything.

The agents

One queue. Four specialists.

Each agent owns a single GTM concern and writes one artefact kind. You stay in command of every external action.

drafts on register

Scribe

live

Anchor essays, posts, threads, case studies. Reads voice samples, guardrails, and the project objective at the top of every run. Voice-checks the output before queuing.

producesdraft essay

finds matching prospects

Scout

live

Plans search queries from your ICP, scores candidates against it, dedupes against existing prospects, masks PII. Returns ranked introductions with the “why match” reasoning attached.

producesprospect

drafts and stages paid ads

Promoter

live

Platform-aware ad strategy — targeting, creative variants, budget, success metrics. Stage approved X campaigns as paused; finalise media and enable spend in the platform. Google, LinkedIn, Meta to follow.

producesad strategy

drafts inbound replies

Concierge

soon

Classifies new mail, drafts a considered reply in your voice, attaches an SLA clock. Coming in M3 — Gmail integration in flight.

producesdraft email reply

How it works

Four steps. No surprises.

The same flow whether you’re drafting an essay, sourcing prospects, or staging an ad campaign.

  1. 1

    Configure once

    Voice samples, ICPs, channels, guardrails, credentials. Embeddings attached on save so retrieval finds them at run-time.

  2. 2

    Trigger a run

    From the project page, name a brief and a campaign. The agent reads every piece of config and produces an artefact — typically in 30-60s.

  3. 3

    Review in the queue

    Side-by-side draft, voice score, run plan, model used, every step the agent took. Approve, edit, or reject with a reason that feeds the next run.

  4. 4

    Act with confidence

    Copy approved drafts into your channels. Stage approved ad campaigns in X (paused) and finalise in the platform. The agent never publishes directly.

Paid GTM

Strategy AI can draft.
Spend, you always approve.

Promoter produces a complete platform-aware strategy — objective, targeting in named terms (locations, interests, lookalike handles, keywords), 3-5 creative variants on your register, budget, schedule, success metrics, rationale.

Approve it in the queue and stage as a paused campaign in X. Finalise media and enable spend in the platform — the moment money moves is yours, not the agent’s.

  • no_paid_ads guardrail blocks the agent at the project level — paid is opt-in, not default.
  • X today. Google, LinkedIn, Meta strategies draft now; staging follows the same shape.
  • Budget cap enforced in the brief, not invented by the model.
stagedin X — paused

campaign · 8f72e4d1

Practitioner cohort — founding round

Budget
$500

$30/day · 14d

Variants
4

on register

Voice
0.91

check passed

LocationsGB · US
Lookalikes@ethanmollick · @benedictevans
Objectivewebsite_visits

Quiet, considered, in your voice.

Multi-tenant from the schema up. Per-project credentials encrypted with AES-GCM. Postgres row-level security on every tenant-scoped table. Built on Claude via OpenRouter so you can pick the right model per task.

Sign inPricing