How to Build a 4-Person AI Marketing Team for Under $24
Four agents. One pipeline. Zero employees.
All powered by markdown files.
What if you could hire four marketing employees who research your competitors, plan full campaigns, create ad assets, build landing pages, and publish everything to Meta?
And what if the whole thing cost you less than $24?
That is not a hypothetical. One founder built this entire AI marketing team using Claudebot. Every single "employee" runs on nothing but markdown files. No code. No complicated automations. Just plain written instructions that tell the AI exactly what to do.
What one run produces
What this AI marketing team actually produces
Before getting into the how, look at what this system actually spits out.
You start by giving it a single competitor URL from the Meta Ads Library. About 10 minutes later, you get a complete analysis of every ad that competitor is running. Every landing page they are sending traffic to. Every funnel they have built.
Then, without you lifting a finger, the system takes those insights and generates a full campaign proposal. We are talking 4 landing pages, 11 ad creatives, 7 video scripts, and 3 funnel stages. All mapped out with targeting, budget allocation, and creative specs.
But it does not stop at planning. The system actually builds the assets. Real landing pages. Real image ads. Real carousel ads. Real video scripts ready for you to record.
And then it pushes everything to your Meta Ads account. Campaigns, ad sets, ads — all created in draft mode, sitting there for you to review before anything goes live.
The entire output is ready for human review
Nothing gets published without your approval. But all the heavy lifting? Done.
The 4-role structure that makes it work
The system is built around four distinct roles, each one acting like a specialized employee on your marketing team. Each role automatically hands off to the next one. It is a chain, and once you kick it off, it runs itself.
Ads Analyst
Handles all competitor research. Feed it a Meta Ads Library link and it extracts and analyzes every ad the competitor is running — images, videos, landing pages, funnels, and strategy breakdowns.
Head of Marketing
Takes the competitor research and plans your entire campaign strategy. Studies your website, extracts your brand guidelines, and builds a tailored campaign proposal with full creative specs and budget allocation.
Creative Director
Produces the actual assets — ad images via Imagen, video scripts, landing pages, and carousel ads. Reviews its own work for hallucinated logos, wrong text, and quality issues before passing anything on.
Performance Marketer
Takes everything the creative director built and pushes it to Meta. Campaigns, ad sets, and individual ads — all properly configured with targeting and ready for your review.
Think of it this way:
Each role is a separate markdown file. You share them with Claudebot and say "import these skills." That is the entire setup. No code. No integrations. No automation workflows.
Free & open source
Build Your Own AI Marketing Team
Download the free skill files and let your AI agent import them. Four roles. One pipeline. No code required.
Get the Skills on GitHub →No code required: how skills actually work
Here is what might surprise you most about this whole setup.
Every single piece of this AI marketing team is just a markdown file. No code. No automation workflows. No complicated integrations. Just plain written instructions that tell the AI agent exactly what to do.
Skills: Markdown files with step-by-step instructions for each task — competitor analysis, campaign planning, asset creation, Meta publishing
Sub-skills: Specialized instructions for specific tasks like ad design, script writing, and landing page creation
Orchestrator: A top-level skill that coordinates the handoffs between all four roles automatically
The skills also teach the agent how to work with other skills. The campaign planner skill explicitly says it works best when combined with the website brand analysis first, then the Meta ads analysis, then the asset creation. It is a self-aware system.
Setup is one sentence:
"Hey, I am sharing with you a set of skills for a marketing team. Can you use your skill creator tool to import all of these skills for yourself?"
And the agent handles the rest.
What it actually costs to run
The cost depends on which AI model you use. For marketing work, Claude from Anthropic is the recommended choice. Opus 4.6 is the strongest model right now and was used for everything in the demo.
That might sound like a lot at first glance. But think about what you would pay even one marketing freelancer for a day of work. It is not even close.
A note on Claude subscriptions
A Claude subscription at $20 a month is cheaper than API pricing. But there have been reports of Anthropic shutting down accounts that use the subscription to run Claudebot or OpenClaw. The API is the officially supported route. If you go the subscription route, that risk is on you.
Frequently asked questions
Common questions about the AI marketing team setup.
Continue reading
OpenClaw for Paid Ads — Fire Your PPC Agency and Keep the Results
Use OpenClaw to audit, optimize, and report on Google Ads and Meta Ads campaigns through natural language.
OpenClaw for Social Media Marketing
Automate scheduling, competitor monitoring, and engagement across every platform with OpenClaw and Mixpost.
Conclusion
This is what marketing is starting to look like. Not one person doing everything. Not an agency charging you thousands a month. A team of AI agents, each with a specific job, each handing off to the next, all running on simple written instructions.
You still make the final call. You still review the targeting, approve the creatives, and hit publish on the campaigns. But the grunt work — the research, the planning, the asset creation, the campaign setup — that is handled.
For under $24.
The question is not whether AI can do marketing work anymore. It clearly can. The question is whether you are going to keep doing it all manually.
Nikhil Kumar (@nikhonit)
Growth Engineer & Full-stack Creator
I bridge the gap between engineering logic and marketing psychology. Currently leading Product Growth at Operabase. Builder of LandKit (AI Co-founder). Previously at Seedstars & GrowthSchool.