GTM Brigade in your AI assistant
Everything below is a sentence you can type.
Connect GTM Brigade to Claude, ChatGPT, Gemini or Perplexity and your LinkedIn engagement stops being a set of screens you click through. It becomes something you ask for. These are real examples. Each one shows the tools that actually run, and marks the ones that change your account or act on LinkedIn.
https://api.gtmbrigade.com/mcp- reads your data and changes nothing
- writes something inside GTM Brigade
- acts outside on LinkedIn, email or your CRM
Know where you stand
The numbers you would otherwise open four screens to assemble, answered in a sentence.
How did my LinkedIn posts do last week, and am I hitting my targets?
linkedin_get_my_analyticsreadslinkedin_get_my_progressreads
Impressions, reach, followers gained, comments and reactions you gave and received, your activity score, and how each of those sits against the weekly targets set for you.
Which of my posts got the most comments this quarter, and what did they have in common?
linkedin_get_my_postsreads
Every post you published in the range with its own reactions, comments, impressions and engagement rate. Enough for the assistant to find the pattern rather than guess at it.
Show me the team leaderboard for this month and who has slipped since last month.
linkedin_get_team_leaderboardreadslinkedin_get_team_overviewreads
Ranked team activity and the org-wide roll-up, so the comparison is against real movement rather than a screenshot someone pasted into Slack.
Owner access. A member asking this gets a clean "this needs owner access", not somebody else's numbers.
Is our company page pulling its weight compared to my personal profile?
linkedin_get_company_page_analyticsreadslinkedin_get_my_analyticsreads
Company Page impressions, reach, follower growth and per-post performance, side by side with your own.
Owner access, and the page has to be connected.
Find the right people, and prove they are the right people
Your lead database is already filterable in the app. Here you describe who you want instead.
Find every VP of Marketing in Germany in our leads with a relevance score above 70, sorted by followers.
gtm_get_leadsreads
The same filters the Leads page runs (role, country, industry, company, lead source, LinkedIn badges, score range, enrichment status, assignment), with sorting and paging.
Add these five LinkedIn profiles to the watchlist and enrich them.
gtm_add_leadswritesgtm_enrich_leadswrites
The profiles land on the watchlist and enrichment runs in the background; ask again later for its status.
Enrichment costs credits, so it is quoted first: you see the exact cost and your balance afterwards, and nothing is spent until you approve that quote.
Build an ICP for mid-market ecommerce ops leaders and show me who already matches.
gtm_create_icpwritesgtm_recompute_icpwritesgtm_get_icpreads
A saved ICP with its signals, scored against your existing leads. The same scoring the app uses, not a second opinion invented in chat.
Put the twenty best matches in a list called Q4 outbound and mark them ready for outreach.
gtm_create_lead_listwritesgtm_add_leads_to_listwritesgtm_set_ready_for_outreachwrites
A real lead list your team sees in the app, with the outreach flag set.
Push this lead to HubSpot.
gtm_push_lead_to_crmacts outside
The lead is created or updated in your connected HubSpot, Salesforce or Pipedrive.
Owner access, and the lead needs an email, the only key either CRM will dedupe on.
Turn what your company knows into posts
The knowledge base stops being a folder nobody opens and starts being the source your drafts come from.
What do our docs actually say about onboarding time?
gtm_search_kbreads
The passages that answer it, from your organization documents plus your own private ones. Never a teammate's private ones.
Draft a post about the pricing change using our positioning docs, in my voice.
gtm_generate_postwritesgtm_create_draftwrites
A draft grounded in your knowledge base and written to your own writing style, saved to your content pipeline.
Generation costs credits and is quoted first.
Change the opening line and schedule it for Tuesday at 9.
gtm_update_draftwritesgtm_schedule_postwrites
The edit is saved and the post is queued, in the same queue the app publishes from.
Actually, publish it now.
gtm_publish_postacts outside
The post goes live on LinkedIn under your own name, immediately.
This one leaves GTM Brigade. Your assistant asks before running it.
Work the engagement loop without opening LinkedIn
The daily grind: see what the people you track posted, say something worth saying, move on.
What did the people I watch post today?
gtm_get_feedreads
Today's posts from your watchlist, with the engagement they are picking up.
Write me a comment on the second one, in my voice.
gtm_generate_commentwrites
A comment written in your own style. The connector learns that style from comments you have actually written, including their typical length.
Costs credits, quoted first.
Post it, and react to the other two.
gtm_post_commentacts outsidegtm_react_to_postacts outside
The comment and the reactions appear on LinkedIn under your name.
Both leave GTM Brigade, and both are capped at twenty public LinkedIn actions an hour, so a looping agent cannot spam on your behalf.
What is sitting in my comment automation queue, and can you approve the good ones?
gtm_get_automation_queuereadsgtm_approve_automation_draftacts outsidegtm_reject_automation_draftwrites
The suggested comments waiting on you, reviewed one at a time. Approving publishes, rejecting just clears it.
Run the team
Owner work: who is on, what they are aiming at, and what it is costing.
Invite these three teammates and set everyone a target of three posts a week.
gtm_invite_memberacts outsidegtm_set_targetswrites
Invitations go out by email and the weekly targets are set, the same targets the progress numbers are measured against.
Owner access. The invite sends a real email, so it is flagged.
Assign the Q4 outbound list to Alex.
gtm_assign_leadswrites
The leads show up as Alex's in the app.
How many credits have we burned this month, and on what?
gtm_get_consumptionreadsgtm_get_plansreadsgtm_get_invoicesreads
Consumption broken down by what spent it, against your plan allowance, with the invoices behind it.
Use it as a research source
One search across everything you can see, and a fetch that opens any result in full. These are the two tools ChatGPT's research mode looks for by name.
Search my GTM Brigade for everything we have on Acme.
searchreadsfetchreads
Matching leads, your posts and drafts, and knowledge-base assets. Each one is citable, and each links back to the LinkedIn profile or the record in the app.
ChatGPT's deep research is read-only by design. It will read all of this and cite it; it will not run anything that writes.
What it will not do to you
The connector can publish, comment and delete. That is only safe because of the rules underneath it, so here they are in full.
Nothing spends money without a quote you approved
Enrichment, post generation and comment generation cost credits. Each one prices the run first, shows you the cost and what your balance will be afterwards, and only executes against a quote you approved. There is no path where an assistant quietly burns your allowance.
Six tools reach outside GTM Brigade, and they are marked
Publishing a post, posting a comment, reacting, approving an automation draft, inviting a member and pushing a lead to your CRM are the only actions that leave the product. They are flagged in the protocol, so your assistant asks before it runs them.
Six more can delete things, and they say so
Deleting a draft, leads, a lead list, an ICP or a knowledge-base asset, and removing a profile from the watchlist, are flagged destructive, which is what makes your assistant warn you first.
You only ever reach your own account
Who you are and which organizations you belong to are worked out on our side from the account you signed in with, never from anything the assistant sends. Members see their own work; owners see the team. Ask for another organization and the request simply does not resolve.
Your private documents stay private
Knowledge-base assets you uploaded for yourself are isolated from your teammates'. Searching, reading and generating from the knowledge base all respect that boundary, in the connector exactly as in the app.
You can pull the plug from your side
Remove the connector in your assistant and the access token stops working. Access tokens last an hour; refresh tokens rotate on every use, so a stolen one is single-use at best.
Questions
- Which assistants does this work with?
- Claude, ChatGPT, Gemini and Perplexity, from the same URL and the same account. Claude and ChatGPT register themselves when you paste the URL. Gemini Enterprise and, sometimes, Perplexity ask for a client ID and secret instead. Ask us and we issue a pair for your organization.
- Do I need a developer to set this up?
- No. In Claude or ChatGPT you paste one URL, sign in on our consent page and click Allow. The setup steps for every assistant are on the connector page.
- Can it post to LinkedIn without asking me?
- Publishing, commenting and reacting are flagged as actions that reach outside GTM Brigade, so your assistant asks before running them. They are also rate-capped at twenty public LinkedIn actions an hour, so an agent stuck in a loop cannot spam your network.
- Does it cost extra?
- The connector itself is included. Anything that already costs credits in the app (enrichment, AI generation) costs the same through the connector, and is quoted before it runs.
- What can it not do?
- It cannot upload files to the knowledge base (there is no file-transfer path over the protocol), change anyone's role, remove a member, or manage billing. Those stay in the app on purpose.