Your First GTM System at Seed (One Motion, Small Stack)

There is a loud story on X that "old school GTM is broken": hire a row of reps, buy a pile of legacy tools, and hope the motion appears. Fivos Aresti has made versions of that argument, and has also called Clay plus Claude Code "a great starting point." Others have posted starter stacks they would learn on if they were beginning now.

Pramod Kumar, co-founder of Kingdom of Kumar. On X

·

I agree that a seed team should not copy a late-stage org chart. I do not agree that a first system is a pair of fashionable tools. A first GTM system is one motion, running in accounts you own, with a fit rule, a way to reach people, approvals, and a Friday review. Tools come after that sentence is true.

This is the smallest system I would build. For the stack itself, see A GTM tech stack for non-technical founders. For the weekly habit, see A founder-led outbound system and Your first outbound motion in 30 days. The craft is also on GTM engineer and GTM engineering. If the product is not ready to sell, take the App Launch Score and do not build a system around a shrug.

"Old school GTM is BROKEN"

The part I would keep: you probably should not hire a team of SDRs before anyone has learned the conversation by hand, and you should not buy a full enterprise stack to look serious. Those moves assume a motion that does not exist yet, and they burn cash and calendar while you find that out.

The part I would not keep: that everything from the last decade is useless, or that "AI-native" is a substitute for talking to users. A CRM with five honest fields, a mailbox you set up properly, and a sheet of fit rules are not "old school" in a shameful way. They are how you see what is happening.

If someone is selling you a complete replacement of "legacy GTM" in a week, they are selling a demo.

"Clay + Claude Code is a great starting point"

It can be *a* starting point for a team that already knows who they sell to and will cap spend. It is a poor first purchase if you are still arguing about the offer, or if no one will approve the drafts the model produces. Clay is a workbench. Claude, or Grok, can help with research and drafts. Neither one is a system. I have also seen the stronger claim that a GTM team "only needs" those two; I do not believe that, and I unpack it in the stack article.

A starting point I would defend for most non-technical seed founders is blander: a written fit and never-contact rule, a small trustworthy list, a properly set-up sending domain, a model you use as a copilot behind approval, a CRM or a serious sheet, and a weekly review. Add Clay, or n8n, or a sequencer, when a specific step is painful enough to justify it.

The first system, in six parts

1. One motion. Outbound to a tight fit, or inbound follow-up done properly — not both on week one, and not five plays from a Twitter thread.

2. A one-page rule. Who deserves a message, who never gets one, and what evidence a first line is allowed to use.

3. A way to get and check contact data. One primary source and a verification step. Not a waterfall of five vendors on day one.

4. A sending path you can stop. Company-owned mailboxes, authentication in place, volume you can watch. See the infrastructure and deliverability pieces if this is new.

5. Approval, replies, and follow-up. A person signs off before mail goes out. A person answers. Follow-up is a rule, not a mood.

6. A Friday review. What ran, what should not have, one change or none.

If any of those six is missing, you do not have a system yet. You have tools and intention.

A 30-day build that stays small

Week 1: write the rule, pick the motion, set up sending and a place to record conversations. No volume.

Week 2: run a small approved batch. Fix what breaks. Do not add a second channel.

Week 3: keep the same motion. Improve the list and the first lines. Add a tool only if a step is clearly the bottleneck.

Week 4: write a handover note as if you were going to miss a week: where the rules live, how to send, how to stop, what to review on Friday. If you cannot write that note, the system is still in your head.

What I would not put in the first system

  • A second outbound channel "while we see what works"
  • An AI SDR that can send without approval
  • A scoring model with ten dimensions
  • A forecast dashboard for a pipeline you cannot yet name
  • A hire whose first job is to choose more tools

Those can be right later. They are how a first system becomes a construction site.

A note on "AI-native" as a first identity

I would rather you be known for answering the people who already raised their hand, or for writing ten truthful first lines, than for having an AI-native label on the stack. Models are useful inside the system I described: research, a draft, a summary of the Friday note. They are a poor substitute for the six parts. If a vendor's first slide is the identity and the sixth part — review — is missing, you are buying a story.

The same caution applies to hiring. A first "GTM engineer" who only wants to assemble agents will not give you a seed system. A founder who will run the six parts, with or without help, will.

When to graduate the system

Graduate a piece when it hurts in a specific way. Move off a sheet when records conflict. Add a sequencer when follow-up is failing because of bookkeeping, not because you dislike writing. Add enrichment automation when you already know which fields matter. Hire, or buy a build-and-run team, when the motion would continue if you were on the road for two weeks — or when it should, and currently will not.

Do not graduate because a thread made your stack look small.

How you know it is working

Not by the number of tools connected. You know it is working when a normal week produces approved outreach, follow-up that actually happens, replies handled the same day, and a Friday note you believe. Conversations may still be few. That is not automatically failure. Silence plus a review that changes the rule is progress. Silence plus a new stack every Monday is not.

If you want this built and kept running with you in the loop, Kingdom of Kumar does that work for $500 a month, three-month minimum, on Grok and Claude. You own the system and you approve what goes out. We will start with one motion, not a theatre of tools. It is not equivalent to a full-time GTM engineer, and it should not be sold as one.

FAQ

Is a sheet enough to be a "system"?

If it holds the rules, the queue, and the outcomes, and people actually use it, yes. Move to a CRM when the sheet is lying or when more than one person needs to work it at once.

Do I need a GTM engineer to build this?

No. You need a few focused weeks and the willingness to keep the Friday review. Hire or buy help when the motion is real and your calendar cannot be the only engine.

What if inbound is already trickling in?

Then inbound follow-up might be the first motion, not outbound. One motion still applies. Do not neglect a demo request because a thread told you outbound is the modern way.

Contact

hello@work.kingdomofkumar.com