Get cited in AI search.
ChatGPT, Claude and AI Overviews answer the question now, and the click never happens. I'm building the playbook for getting a brand cited in those answers, in public, on real accounts.
- Members
- 186 on Skool, opens in a new tabSeptember 2026
- Price
- Free, publicNo credit card, no countdown
- Results
- Two real accountsCharts further down the page
- Run by
- Richard MagallanesEnterprise SEO by day, Sydney
Nobody has a GEO playbook yet
AI search is rewriting how people find brands. The model answers the question, cites three sources, and the click you used to get never happens.
Meanwhile someone above you wants to know why organic is down, and “the models are answering it instead of us” does not survive that meeting on its own.
So we're building one, in public.
answers happening without you
- ChatGPT
- AI Overviews
- Claude
- Perplexity
The system that runs the plays
Every play in the Lab runs on MarketingOS: a set of plain files that hands Claude Code your business context, your instructions and your working history. You don't write code. You tell it what you want and review what comes back.
- Your Brain
- Voice, audience, offer and the accounts you work on. Written once, read every run.
- Your Skills
- One reusable capability each, for the jobs you repeat every week.
- Your Blueprints
- Skills ordered into a discipline, so a whole workflow runs start to finish.
- Your Work
- Research, briefs, drafts and decisions, kept where the next run can find them.
- BRAIN.mdthe operating contract
- CONTEXT.mdcurrent focus and constraints
- business/the only source of business truth
- brand/
- audience/
- offers/
- strategy/
- proof/
- decisions/
- knowledge/sources kept, wiki maintained
- sources/
- wiki/
- .claude/skills/one skill per repeatable job
- content/dated, one folder per channel
- campaigns/
- reporting/
- outputs/every run, kept
The folder layout every play runs on. Plain files you own, read by Claude Code or Codex.
The engine can change. Claude Code built most of it and I now use Codex as well, so the system outlasts whichever model is best this month.
Your business context and your working process stay in files you own.
Inside the Lab
MarketingOS (Your AI Brain)Build the AI Brain (MarketingOS) that I use to ship SEO/GEO strategies at scale., opens in a new tab
Tools, Templates & ResourcesSteal my entire tech stack., opens in a new tab
Vibe Coding MasterclassLearn to build and automate anything with vibe coding. Powered by MarketingOS™., opens in a new tab
PremiumThe Brand AuditGEO Pt. 1: Audit your brand and see how LLMs are perceiving your brand., opens in a new tab
PremiumThe Technical AuditGEO Pt. 2: Audit your website architecture to ensure that AI can read your content., opens in a new tab
PremiumThe Content PlanGEO Pt. 3: Learn how to build a content plan to get your brand cited in LLMs., opens in a new tab
PremiumThe Prompt LibrarySteal my most popular prompts., opens in a new tab
Build In Public RecordingsRaw and unfiltered recordings of me building with AI., opens in a new tab
Mirrored from the classroom on Skool, September 2026. Premium courses sit behind the paid tier; everything else is free.
Some lessons, Blueprints and Masterclasses are still being built. The structure above is what we're working toward, not a claim that every item is live today.
Built from real accounts, not a course
I do enterprise SEO by day and freelance for small business clients on the side. That's the whole reason this works: I get to test the same idea against enterprise constraints and one-person constraints in the same week.
I started building these systems because I wanted AI to handle the repeated marketing work properly. Not as a demo. As something I could run again next week.
That work now includes
- A content system that researches, outlines, drafts and routes each deliverable.
- A skills library of 80 working skills, and counting.
- A private research wiki compiled from 51 official Skool videos.
- Traffic turnarounds on an e-commerce store and a local gym.
- The MarketingOS engine, running on Claude Code and Codex.


The Lab is where I document the working bits, the broken bits and the decisions behind them, so you can run the same thing on your own accounts.
One honest gap: I don't have a wall of citation screenshots yet. When the first one lands, it goes up, and you'll have watched it happen.
Bring one page you need cited.
It will suit you if you
- Own organic traffic for an employer, a client or your own site.
- Can see AI answers taking clicks you used to get.
- Have to explain the drop to someone above you.
- Would rather run the playbook than read about it.
- Do not have a developer background.
It won't suit you if you
- Want a tool you can buy and forget about.
- Collect prompt packs but never run them.
- Prefer polished demos to visible process.
Come and build the playbook with me.
Start with the onboarding course, sit in on the weekly live build, and run the first play on a page that actually matters to you.
Join the Lab free(opens in a new tab)Free community on Skool. No credit card, no fake countdown.