AI native marketing | One-week sprint, day 1: Marketing Brain, context engineering
Generic AI output is a context engineering problem.
Every draft starts on brand, because Claude reads your buyers, positioning and voice before you type.
- Build an ICP dossier from your customer list, reviews, tickets and call notes
- Fill a messaging hub that names the buyer’s job, the category and the alternatives
- Measure your voice from published copy, with “write this, not that” examples
- Check every output against a quality gate built from nine expert reviews
- Replaces: Pasting context into every prompt, and the generic drafts that come back when you don’t.
Sprint day 1: Tuesday, 10 November 2026 | 4 PM CET / 9 AM CST / 3 PM UTC | 90 minutes
Included in a CXL annual subscription, with 3 months community access.
32 marketers posted a working marketing brain in the same week as the last workshop.
Trusted by marketers at
Your live instructor and messaging expert
4x’d efficiency & leads CXL’s growth, experimentation & product. Trained 25k+ marketers with 80+ courses, 4000+ blogs & playbooks.
The cause is generic context. The model guesses what you never wrote down.
Here is what the last workshop’s demos surfaced when marketers ran it on their own business:
- An ICP built from CRM deals that didn’t match the company’s own brand guidelines
- A brand book that sounded sure, where every line about the buyer turned out to be a guess
- A live website breaking its own style rules, caught by the voice check
- A clear line between who buys today and who the business wants to win next
Every gap gets tagged, so you fix it once. You leave with an ICP dossier, a messaging hub and a voice guide built on your own business, checked and saved where your AI tools read them.
Included in a CXL annual subscription, with 3 months community access.
📅 Day 1 of 3, 90 minutes
Tuesday 10 November 2026
4 PM CET / 9 AM CST / 3 PM UTC
Includes 3 months community access so you can keep building and get support after the session.
Came to previous workshops? This is a teardown of the updated system, not a rerun.
Before the session: install it and run it on your business
Install the Marketing Brain plugin, then run its three commands on your customer data, or on the sample company if yours isn’t ready.
- A few minutes to install, and 30 minutes at most to run on the sample data
- On your own customer data the results are richer, but it can take longer. More time and more connected MCPs, richer results
- Outcome: Your first ICP dossier, messaging hub and voice guide, ready to take apart live
Part 01: Your ICP, from your own customer data
Start from what your customers say, not from your pitch deck. Live, we open the ICP step: how it reads your customer list, reviews, tickets and call notes, and how to point it at your own sources.
- Segments, buying triggers and the exact words your buyers use
- Every line traced to a source, and anything AI guessed tagged so you can check it
- How we built it: why every line needs a source, and why a guess gets tagged instead of deleted
- Outcome: An ICP step you can point at your own data sources
Part 02: Positioning and messaging
You filled the messaging hub in prep. Live, we show how the framework behind it became a template Claude can fill, and how to change it for your market.
- Who it’s for, the job they hire you for, the category they shop in, and what you replace
- Messaging pillars, each one mapped to a problem your buyer has
- Outcome: A messaging hub, and the template behind it, both yours to change
Part 03: Your voice, measured from your best copy
Voice extracted from work you’ve published beats voice described in adjectives. Live, we show how the voice check measures your copy, and how to change what it measures.
- Three or four traits you can hear, each with “write this, not that” examples
- The words you own, and the ones you never use
- Outcome: A voice guide AI applies to every draft
Part 04: Check it before you use it
Every output runs a quality gate built from an expert’s video reviews of nine cohort demos. Live, we open the gate: how those reviews became checks Claude runs, and how to add your own.
- Fix notes on the page that say what’s wrong, what to do, and the test to run
- Post your demo in the community for feedback
- Outcome: A checked marketing brain, saved where your AI tools read it
After the session: post your demo
Post a finished or attempted customization or rebuild of the system for your needs in the community.
- A piece or the full system: the more you show, the more we can review
- Demos close Friday 20 November, end of day
- Outcome: Feedback on your build and your certificate
What you need before you start
- You should be comfortable with: running campaigns and reading your own performance data. No coding needed.
- You do not need: prior Claude Code or n8n experience. Every build starts from a template.
- A paid Claude account, Claude Code preferred, plus a free Miro account. The reporting agent and AI visibility builds also use n8n.
- About 2 hours of prep in the app on the Monday before, matched to your level by the quiz. If you get stuck, we help you through it.
- Any other tool is explained live, with alternatives. You are not locked into one stack.
What marketers said after building their marketing brain
AI-Forward Marketing Executive
“I was in your workshop 12 hours ago. I already have a workflow live… I feel like a freaking WIZARD!”
Co-Founder, Feral MKT
“We’ve never used VS Code or Claude Code before. We’ve been running in AI-assisted mode only. This has changed the game entirely… Before the cohort, I was having issues with our chosen ICP, but the dossier exercise helped me nail it down fast.”
Creative Director, Feral MKT
“A brand brain built on what you can prove is smaller, but much sharper, and the voice guide finally reads like us because it’s measured from our real copy (which will change anyway), not from adjectives.”
CRO Manager | Insurtech
“It will make the review process with the agency lead very focused, since you have already pointed out the issues – I basically have a list of fixes for them already!”
Founder, Mesh Creative Studios
“Seeing that distinction mapped out has already made the business feel much more structured. It’s impressive how much clarity came out of only the second workshop.”
Who has built one
A pension fund. A B2B payments provider. ESG software. A market research firm. Clinic management software. A fractional CMO practice. A mental wellbeing nonprofit. A social club that hasn’t launched yet.
Social club for people 60+, pre-launch
No customers yet, so the ICP is built from reviews, Reddit threads and US research, with every unproven line tagged.
B2B eCommerce platform, rebuilt
Separated what past customers prove from who the rebuilt product needs to win next.
Fractional CMO practice, launching
Tested a new fractional CMO offer on her own website, using public data only.
Marketing agency for health clinics
Ran the brand book through it: every claim about the buyer came back tagged hypothesis or blank.
Build it in the one-week sprint. The best demos are selected for expert feedback like this.
The next AI native cohort
One week. Three working workflows. Starts 10 November 2026.
Marketing Brain
One document every AI tool reads, so drafts start on brand.
Campaign engine
One brief in, a full campaign draft out, with sources you can check.
Reporting agent
A weekly summary of what changed and what to do next.
Build three AI native workflows you use at work, in a one-week sprint.
You build a marketing brain your AI draws from, a campaign engine that checks its own claims, and a weekly reporting agent.
Keep up with AI native marketing tactics and workflows.
Subscribe to our weekly educational newsletter
We find and report on AI native marketing case studies, success stories, and playbooks that are working out there right now.