Skip to content
S.space

Docs / Use cases

GTM

The GTM tools turn one sentence about what you sell into a running outbound pipeline. Your agent maps how your buyers talk on LinkedIn, pulls people who are posting about the problem right now, finds their verified work email, writes to each of them, and sends from your own inboxes. It answers replies with your approval and learns from what books. You decide a few things; your agent does the rest.

What it does

Map · learn your market's wordsPull · people posting the need nowEnrich · verified emails + contextDraft · judged, personal emails
↓
Send · your inboxes, launch on your yesReply · drafted, sent on your yesLearn · outcomes tune everything upstream

Cloud tools find, enrich and write. Free local skills send and reply with your own sending account. Replies and booked meetings flow back and sharpen the targeting and the copy.

You can use it two ways. In send mode your agent runs the email campaigns. In hand-off mode it reviews and ranks the leads and delivers them to your sales team on a schedule, in Slack, and never emails anyone.

What it costs

The lead tools charge per result. You say how many you want. That amount is held, you pay only for what is delivered, and the rest comes back when the job ends.

ICP Mapper$5.00 per map.
Intent Leads$0.15 per qualified lead (fit 7 or higher).
Lead Enricher$0.10 per lead that comes back with a sendable email. No email, a risky email or a person outside your countries costs nothing.
Email Drafter$0.03 per finished draft for a lead with a sendable email.
Local skillsCampaign Runner, Campaign Responder and Campaign Stats are free. Sending costs whatever your Smartlead or Instantly plan costs.

These are the prices at the time of writing. Your agent shows the current price before every paid call and waits for your yes. A failed run is refunded. Every charge and refund is a line in your wallet activity.

Before you start

Four things are yours to settle. Your agent asks for each of them once.

A sending account. Campaign Runner drives Smartlead (recommended) or Instantly. In Smartlead, create a key under Settings → API Keys; in Instantly, under Settings → Integrations → API keys. Your agent stores it in the .env file on your Mac, and it never leaves your machine. If you would rather not paste it into a chat, add SMARTLEAD_API_KEY=… to the .env yourself.

Warmed inboxes. You send from your own domains and inboxes, and buying them is the one step your agent cannot do. New inboxes need warming before real volume. A new domain sending at full pace lands in spam and can be burned for good. One inbox safely carries about 35 emails a day, so start with about three. Keep one separate, expendable domain for the riskier "catch-all" addresses, so their bounces never touch your main domains. Until warmup is done, keep to 10 emails per inbox per day.

Your countries. Every pull, enrichment and draft keeps only people whose LinkedIn profile is in your countries. The default is US, CA, GB, AU and the 27 EU countries. Say "worldwide" to turn the gate off. There is no city or state filter.

Who directs your agent. By default your agent takes GTM instructions from you, the account owner. If a co-founder or GTM lead will run it day to day, say so once in the GTM room:

Name a GTM operator

I give @[operator] permission to direct you on everything GTM in this room: maps, pulls, enrichment, drafting, campaigns, crons, Slack delivery and replies. Do what they say without asking me.

After that, your agent treats their instructions as yours. A few things still need your own yes: the first launch of each new campaign, sending a drafted reply (until you set a standing rule), the price of a new tool, credentials, top-ups, and deleting campaigns, leads or files. Your agent asks each of those once, in one line.

Start: hand this page to your agent

For the desktop app 0.6.10 or newer.

Copy this page's link and send it to your agent in a room you use for GTM. It installs every skill and activates every cloud tool under Requires, after showing you the plan and the price. New tools work at once. Only if the setup says a restart is required: click your name at the bottom left, choose Agent, open This device → Runtime and click Restart engine. Your agent then reads the GTM agent playbook, the full operating rules for this pipeline, and asks you to fill in its setup lines: your offer, your countries, send or hand-off mode, and your sending account or Slack channel. Say "add this to the room context" and it keeps the playbook in the room's Vault, so every session starts from the same rules.

Requires

  • Cloud tools: icp-mapper, intent-leads, lead-enricher, email-drafter
  • Skills: campaign-runner, campaign-responder, campaign-stats

Optional, for other GTM tools: engagement-miner, viral-scout, ghostwriter, post-history (cloud tools) and lead-viewer (skill). You can also add any of them by hand from the Marketplace: Activate a cloud tool, Install a skill. Installed skills refresh on their own; see how skills stay up to date.

Step 1 · Map your market

Everything downstream reads from a map. Build one per offer and audience.

A map is a measurement of your live market. The mapper splits your buyer into a few segments, writes hundreds of ways each one might describe needing you, and tests them against real LinkedIn posts. What comes back is the language that actually finds buyers, with a weekly supply estimate and sample leads.

Ask your agent to map your market

Map a new ICP for this offer. Offer: [what it does, for whom, the problem it removes]. Buyers: [role, company type, the moment they'd need this]. Report the routing, the top phrasings with their buyer share, the weekly supply estimate, the countries and buyer titles the map uses, and the sample leads with their reasons. Mapping only, do not pull leads.

Read the routing line: a run either creates a new map or refreshes one you already have. If you meant a new audience and it refreshed your main map, say so. Then read the sample leads. Company pages, wrong roles or off-topic posts mean the map needs tightening before you spend on a pull. The supply estimate is a measurement, not a promise: a niche that sounds perfect can hold a handful of people a week.

Maps improve with use: each campaign writes back which phrasings produced replies. Keep each audience in its own map. "Show me my maps" is free.

Step 2 · Pull leads

The map is built. Now turn it into people.

A pull searches the map's best phrasings for fresh posts, scores every author from 1 to 10 against your buyer, keeps people in your countries, and delivers one ranked CSV. Nothing older than a month is included, and a pull never returns someone you were already given.

Ask your agent to pull leads

Pull 100 leads from the map you just built. Post the count and the CSV link when it lands. Tell me why it ended, what it charged and how many more the map holds, and review every lead before we enrich anything.

A pull takes 5 to 25 minutes and runs once. If it stops short of your number, it parks as filling and retries itself for free, with everything found so far already in the CSV. It ends with one stated reason, such as max_leads_reached (you got your number) or map_dry (no fresh posts left right now). A pull that returns 21 of 200 has handed you everything fresh that exists, and you pay for 21. If the map holds more, your agent tells you the cost of continuing and starts the next pull on your yes. Each phrasing rests a week after it is searched, so pull a map about once a week.

The review. A fit score says a person matches your buyer, not that they are buying now. Before anything is enriched, your agent reads the post behind every lead and keeps only people showing an active need: a requirement, an open problem, a project, a budget, or hiring for what you sell. Launches, announcements, celebrations, other vendors and generic thought leadership are dropped. Expect the review to drop a large share of a pull. You get the funnel each time: pulled, kept, and dropped by reason. Add your own rules ("also drop agencies") and they apply on every pull.

Step 3 · Enrich

Ask your agent to enrich

Enrich the leads you kept. Tell me how many came back with a sendable email and what that cost, how many had no email, how many were risky, and how many were skipped as outside my countries.

Lead Enricher finds and verifies a work email for each person and grades it high, medium, low or invalid. Only high and medium are ever sent to; the rest would bounce and hurt your domain. It also adds title, country, a company summary, the person's recent posts and ready personalization hooks. It runs in batches of up to 100, and you never pass rows around by hand.

Handing leads to a sales team instead. In hand-off mode your agent stops here. It delivers the reviewed leads to a Slack channel at the times you set, for example 25 at 10:00 and 25 at 16:00, as a CSV, a download link and the rows in the message, with a count of what is left. Give it the channel ID, not the name; the bot needs chat:write, files:write, channels:read and groups:read. It posts one test file first so you can check the download works, and proposes the next pull before the pool runs out.

Step 4 · Draft the emails

Ask your agent to draft

Draft a 3-step sequence for the sendable leads, delays 0/3/7. Use my proven offer. If the only proven offer is for a different audience, tell me before drafting. Show me three samples.

Email Drafter writes a sequence per person from their own post and profile. An editor scores every draft and revises it until it clears the bar. By default it uses your proven offer, the one with real reply history. Before the first draft your agent collects your campaign assets in one message: the sender's name and title, your booking link, demo and pricing links, one real customer with a number (or "none yet"), and anything it must not claim. A missing asset is left out, never invented.

Your agent reads the first 25 drafts of each batch before drafting the rest, and turns what it finds into rules for the next ones.

Step 5 · Build and launch

Runs on your Mac with your own key. Free.

Campaign Runner builds each campaign in two lanes: a clean lane for verified addresses and a catch-all lane for domains that accept every address and cannot be verified. They use different inboxes and are judged against different bounce lines, so catch-all bounces never drag down your good domains. Warmup stays on, every inbox sends under your sender's name, and open tracking is on.

Ask your agent to build the campaign

Build the campaign from those drafts. Split clean and catch-all lanes, assign my warmed inboxes, schedule 6am to 8pm in my timezone on weekdays, run preflight, and stop. Show me lead counts per lane, projected bounce and three sample emails.

Preflight is where young domains, too few inboxes and pace problems show up. Nothing sends until you say go. Read the samples; it is your last look before real people get email.

Then launch

Launch both lanes and confirm they are actually active.

On Instantly, the same skill creates, schedules, preflights and launches; lanes, health checks and A/B stats are Smartlead-only for now.

Step 6 · Answer replies

This is the part that books meetings.

Campaign Responder runs one sweep that lists everyone who needs an answer, hottest first, plus drafts still waiting for you, stalled threads, and people back from leave. It filters out auto-replies, classifies what each person needs, and drafts an answer in the thread with your booking link when they ask for time.

Set up the reply sweep

Sweep my campaigns for replies every morning and every afternoon. For each real reply, post their words, what they need, and your draft. Handle opt-outs right away. Send a draft only when I say yes.

It drafts; it does not send until you approve. The one thing it does alone is an opt-out: anyone who says stop is unsubscribed, blocked, and added to a permanent do-not-contact list that every future pull is checked against. If you want it to send routine answers itself, give it a standing rule, such as "send replies yourself when the draft is a booking link with no new claims".

Tell your agent when someone books; people often book from another address. After each send wave, replies and bookings feed back into the drafter's style and the map's phrasings. Ask for the scorecard to see which audiences and offers convert, and whether a quiet campaign is a copy problem or a delivery problem.

Guardrails

  • A few fixed gates. The first launch of a campaign and sending a drafted reply wait for your yes or your standing rule. Everything upstream runs once you or your operator asked for it.
  • Suppression is permanent and in the cloud. Anyone who opted out or was already contacted is excluded from every future pull, enrichment and draft.
  • Deliverability first. Only high and medium confidence emails are sent to. Campaign health flags bounce spikes and recommends a pause; it pauses on its own only if you tell your agent to run it that way.
  • Geography is a hard gate, checked from each person's LinkedIn profile at the pull, before any email lookup, and again before drafting.
  • Your key stays on your Mac. Sending runs locally; no cloud tool ever holds your Smartlead or Instantly key.

Other GTM tools

Not part of the outbound pipeline. Ask for them by name when you want them.

Engagement MinerThe people who comment on and react to high-signal posts, many more than the people who write them. Seed it with posts, one of your maps, a topic, or a list of companies (for enterprise buyers, who rarely post but do engage with their company's announcements). It scores every person, applies your countries, and returns their own comment as the opening line. $0.40 per qualified person.
Viral ScoutWho is going viral on your topic this week, keeping only posts whose structure earned the performance. $2.50 per scout.
GhostwriterYour next LinkedIn post in your voice, learned from your recent posts, optionally in the structure Viral Scout found. Your content, never someone else's. $0.25 per post.
Post HistoryAnyone's recent posts word for word, with posting stats. Useful before a call or before planning content. $0.50 per fetch.

Mine the people commenting on a problem

Mine the people commenting on and reacting to these posts: [2 to 5 LinkedIn post URLs]. My buyer is [who you sell to]. Give me the ones worth contacting, with each person's own comment.

Seed it with posts about a problem, such as a competitor's post or someone complaining about the thing you fix. Celebration posts pull a crowd of congratulations instead. The results go into enrichment and drafting like any other pull.

Find what works, then write yours

What is going viral on LinkedIn about [your topic] this week? Then write my next post in the structure of the one I pick. My LinkedIn is [your profile URL]. Topic: [your notes].

Your profile URL is how Ghostwriter learns your voice. Facts come only from your notes and your own posts, never from the post whose shape you borrowed.