How I Built My Marketing Intelligence Layer (+ The Skill and Template)
The skill + template for turning years of marketing experience into an asset every AI tool you use can learn from.
I’ve spent 15+ years accumulating marketing experience. It lives in a sprawling mess, across the parts that become the unconscious way I do marketing, the sprawl of learnings across Google Docs. Presentations. Meeting notes. Slack threads. Random files. And most of it is lost.
That’s just how things worked.
Today, that intelligence is the most valuable asset you have.
Your accumulated knowledge is the number one thing you need to begin monetizing with AI. The frameworks you’ve developed. The marketing campaigns you’ve run. The experiments that succeeded. The experiments that failed. The counterintuitive things you’ve learned. The benchmarks you trust. The mistakes you don’t want to make twice.
For all of us, that experience is something you lived through, at an incredible fast pace, but rarely get all the benefit from.
Some of it we can apply subconsciously, some of it we can prompt to an AI assistant, but so much of it is lost.
AI drastically changes the value of that experience.
Turn it into structured context, and suddenly every AI system you use can reason from what you’ve learned.
So I built something to do exactly that.
I call it my Marketing Intelligence Layer.
And I’m giving you the skill and template to build your own.
This is solving a forever problem. It dates back to Darwin.
Charles Darwin accumulated a huge amount of information over his lifetime. Observations. Quotes. Ideas. Questions. Learnings. Things he’d notice when traveling. Things he’d pick up from conversations.
He kept commonplace books and notebooks to organise what he was learning. It meant ideas captured years earlier could eventually contribute to much bigger pieces of work.
His system mattered because human memory is terrible infrastructure for accumulated knowledge.
Everyone has a version of this problem, including marketers.
You’ve run experiments, campaigns, paid advertising, SEO, brand, product marketing. You’ve built teams. Learned hard lessons across multiple companies.
You know what works, what hasn’t worked. You understand the flaws of attribution. You have strong opinions on the impact of brand and the importance of product positioning because you’ve seen what works, what hasn’t worked.
But where does all of that intelligence live?
For most of us, everywhere.
And nowhere.
That’s becoming a much bigger problem because your AI can only reason from the intelligence it can access.
Every time you open Claude or ChatGPT without that context, it starts cold.
You re-explain your experience. Re-describe your frameworks. Provide context you’ve probably provided 100 times before.
Or worse, you don’t provide it at all.
Then we complain that the output is generic.
So I wanted to make my intelligence portable.
The Marketing Intelligence Layer
The idea is pretty simple.
I wanted one structured artifact containing the transferable marketing intelligence I’ve accumulated across my career.
Frameworks.
Wins.
Losses.
Metrics.
Formulas.
Counter-intuitive learnings.
Operating principles.
All organised by topic and available to whatever AI assistant I’m using.
Here’s the system I built:
The workflow is deliberately simple.
I drop raw material into a folder.
I run /build.
The system extracts the useful intelligence, classifies it, strips confidential company information and adds the transferable learning to my Intelligence Layer.
Then I can give that layer to Claude, ChatGPT, Gemini or an agent I’m building. I can do this by either pointing it at the artifact we’re going to build on my local laptop, or a Google drive folder.
The AI starts with my accumulated context instead of starting from zero.
But that raises the obvious question.
Does it actually make the AI better?
I tested the same problem with and without my Intelligence Layer. A real marketing problem and ran the exact same prompt twice.
One with a clean new AI session. One with my Marketing Intelligence Layer loaded.
First, here’s the prompt:
I'm taking on a new CMO role.
We're trying to grow the pipeline by 100% over the next 12 months for a vertical SaaS AI agent company. It’s a Series B company that is at $50M ARR and wants to double to $100M ARR within 12 months.
Give me the most impactful recommendation you have, one that most CMO’s wouldn’t know and would fail to suggest.
Please start with the TL;DR bullet points.
It’s hard to show in a simple graphic how different the outputs are between these two things, but even the TL;DR does highlight how much better the ‘with intelligence’ version is.
Here’s the summary without my intelligence layer loaded.
Good, but here is it building on top of my intelligence gathered from previous roles, advisory roles, and conversations with smart people.
There is a huge improvement.
Here’s how I built it
There are 4 parts to the system.
1. The Inbox
What it replaces: knowledge scattered across dozens of places.
I don’t want another knowledge-management ritual, another note-taking process, another productivity tool; I already know those things end up in the same scattered mess.
Instead, I drop everything into a /raw folder.
Meeting notes.
Docs.
Transcripts.
Voice memo transcripts.
Bullet lists.
Old presentations.
I don’t need to clean the files, tag them, or decide where anything belongs. The system will take care of that.
2. The Extractor
Running /build tells the skill to read everything in /raw and identify the distinct transferable learnings inside each file.
This is an important distinction.
I don’t want summaries of my documents. I’ve seen people make this mistake time and time again.
I want the intelligence buried inside them.
A single meeting transcript might contain:
a useful benchmark
a channel learning
something we tried that failed
a framework I used to make a decision
a counter-intuitive insight
an operating principle worth remembering
The Extractor pulls those apart into individual pieces of intelligence.
3. The Router
Once a learning has been extracted, the system needs to decide where it belongs.
My default Intelligence Layer covers 8 areas:
Demand Generation
Paid
SEO
Content
Attribution
Product Marketing
Activation
Leadership
But you can add whatever matters to you.
AI in Marketing.
Partnerships.
Community.
Brand.
Developer Marketing.
Whatever you’ve spent your career getting unusually good at.
The important part is that you don’t have to organise it manually.
If the system finds:
“How we structured the paid team...”
That might belong under Leadership, not Paid.
If it finds:
“Changing our activation flow increased...”
That’s Activation.
The routing lives in the build.md skill. It also populates specific learnings attributed to the different companies you’ve worked at. You’ll see this structure in the HTML template as part of the giveaway below.
4. The Confidentiality Layer
It’s going to be fascinating to see how ‘intelligence’ evolves over time. Who owns what intelligence. When working in a company, you have access to proprietary data that you can’t take with you.
Confidential company information belongs to the company. But, the learnings you acquire at that company, the intelligence you gather, is yours to own and capture.
That’s why I built a confidentiality filter into /build.
It strips things like:
exact company revenue
internal targets
employee names
customer names
unreleased products
company-specific sensitive information
But there’s an important distinction between company data and the transferable lesson you learned from working with that data.
For example:
The goal isn’t to create a personal archive of company information.
It’s to preserve the judgment you developed while doing the work.
What you actually get
After /build finishes, I have a single HTML file containing my Marketing Intelligence Layer.
I can open it in a browser.
I can give it to Claude.
I can give it to ChatGPT.
I can use it as context for an agent.
And because it’s mine, I’m not rebuilding my marketing brain every time the hot AI tool changes. You’ll see the template as part of the package I’m giving away below.
You can build mine
I’m giving away the actual system I use.
→ [DOWNLOAD: THE INTELLIGENCE LAYER HTML TEMPLATE]
The template gives you the self-contained artifact that stores and organises your intelligence.
And you can take the skill that builds it:
→ [DOWNLOAD: THE /build CLAUDE CODE SKILL]
Drop the skill into your Claude Code commands folder:
/.claude/commands/
Then run:
/build setup
It asks for your name, your role and whether you want to add custom topics beyond the defaults.
Drop your source material into:
/raw
Then run:
/build
That’s it.
I also built it to work with Google Drive
I didn’t want the system tied to one laptop.
So there’s another version:
/build drive
That finds your Drive folder, reads the source material from /raw and writes the updated Intelligence Layer back to Drive.
Same system.
Now your intelligence is accessible wherever you’re working.
The first version of your Intelligence Layer probably won’t be particularly impressive.
It takes time to build that intelligence. The more you use the system, the better it gets.
The value compounds.
Every meeting.
Every campaign.
Every experiment.
Every failure.
Every framework you develop.
Every time you learn something you don’t want to learn twice.
Drop the source material into /raw and run /build.
Your Intelligence Layer gets richer; it becomes a huge asset.
This is the most impactful way to monetize your experience.
Until now, most of that intelligence has been trapped inside your head.
It’s now feeding your AI assistant.
Build the intelligence layer first.
→ [DOWNLOAD: COMPLETE MARKETING INTELLIGENCE LAYER BUNDLE]
Skill + HTML template + setup instructions. Free, no email required. Available on GitHub
If you build one, I’d love to see what you create. Drop me a note.
Until Next Time,
Happy AI’fying,
Kieran