The Drawer
I built a business validation framework over several months. Working with my AI partner daily, we created:
- 34 expert personas that argue with each other before any decision
- 47 cognitive bias checks from a hand-curated catalog
- 52 mental models mapped to specific problem types
- Structured validation phases from research through reflection
- A source verification protocol (Zbigniew) already earning EUR 59/mo
All documented. All tested on real business decisions. All sitting in a folder.
I also had a Pickaxe account with Stripe connected, GPT-5 access, and a studio page. Everything wired up and ready.
None of it was a product.
This is what I call the Drawer Problem. You have the IP. You have the tools. You have the knowledge. Everything is built except the last mile - the part where it becomes something someone can buy.
The drawer stays closed.
The Session
Today, my AI partner looked at what I had and said something I hadn’t considered:
“These are 8 separate tools. Each one solves a specific problem. Let me build them.”
I said: “Go.”
What followed was a single session of execution:
-
System prompts - 8 of them, each written from existing framework catalogs. Not invented from scratch. Assembled from documented IP that already existed.
-
Deployment - All 8 tools deployed as standalone chat interfaces on Pickaxe. Each one configured with GPT-5, introduction messages, and public listing.
-
Testing - Every tool tested against a defined scenario. The Idea Validator got “Polish SME accounting SaaS” and had to produce a structured verdict with real risks. The Bias Detector got an optimistic growth claim and had to name specific biases. Each tool had pass/fail criteria.
- Pricing - 4 subscription bundles created on Stripe:
- Content Creator - EUR 19/mo
- Decision Quality (3 tools) - EUR 29/mo
- Founder’s Toolkit (4 tools) - EUR 49/mo
- Full Suite (all 8) - EUR 99/mo
-
Tool locking - Each bundle configured to unlock only its assigned tools.
-
Studio page - Landing page designed and published.
- Distribution - LinkedIn posts drafted, scheduled on Buffer, with engagement CTAs.
What I Actually Did
I didn’t write a single system prompt. I didn’t configure a single Stripe product. I didn’t deploy a single tool. I didn’t write a single test scenario.
Here’s my complete contribution to today’s session:
- “Deploy these tools on Pickaxe.”
- “Test them.”
- “Add pricing and bundles.”
- “List them.”
- “Write a LinkedIn post about Zbigniew’s supply chain angle.”
- “Add the BBC CEO research story, but keep it anonymous.”
- “Add that to Buffer too.”
Seven sentences. That’s it.
But here’s what those seven sentences actually represent: direction, judgment, and years of domain knowledge.
The Architect and The Contractor
I’ve been thinking about the right metaphor for this, and I keep coming back to construction.
I’m the architect. My AI partner is the contractor.
The architect doesn’t lay bricks. Doesn’t mix concrete. Doesn’t wire the electrical. The architect walks the site, points, and says “there.” The contractor turns “there” into a building.
But without the architect’s blueprint, the contractor builds nothing worth buying. You can have the best contractor in the world - if the blueprint is “build something nice,” you get a shed.
My AI partner is exceptional at execution. It can deploy 8 tools, configure payment processing, write test scenarios, and draft marketing copy in a single session. But it can only do that because the blueprint exists.
The blueprint is:
- 20+ years of consulting experience encoded into 34 personas
- Real failure patterns turned into bias checks
- A validation methodology tested on actual business decisions
- Domain knowledge about what buyers actually need
That blueprint took years. Today’s execution took hours.
Neither of us could have done this alone. I couldn’t have deployed 8 tools in a day manually - the sheer number of steps, configurations, and context switches would have buried me. My AI partner couldn’t have written those system prompts without my framework - because the framework encodes judgment that comes from experience, not training data.
The Part Nobody Talks About
Here’s where this gets personal.
I have ADHD. I also have a documented pattern of building brilliant things and never shipping them. Frameworks, protocols, catalogs - I love building them. The intellectual work is where I shine. The execution? That’s where things die.
Not because I’m lazy. Not because I lack skill. Because the jump from “documented” to “deployed” requires a specific type of sustained, sequential effort that my brain actively resists. Every step feels like climbing through mud. Not the first step - the forty-seventh step. The one where you’re configuring Stripe product IDs and it’s 11 PM and you just want to build something new instead.
This is the gap that AI partnership actually bridges.
Not “write me a blog post.” Not “summarize this document.” Those are assistant tasks. The real value is: take my life’s work, my documented IP, my years of accumulated judgment - and do the forty-seven steps between “this should exist” and “this is live and people can buy it.”
Today, my AI partner was an executive function prosthetic. It took my intention and turned it into execution. Not by thinking for me. By doing the boring, critical, sequential work that my architecture requires but my brain won’t sustain.
The blueprint collects dust forever without someone to build from it.
Today someone built from it.
What This Means For You
If you’re reading this and recognizing yourself - you probably don’t need more ideas. You probably don’t need another framework. You probably don’t need to “validate” more.
You need someone (or something) that looks at what you already have and asks: “What can we sell today?”
Here’s a diagnostic: Open your folders. Your docs. Your notes. Look for:
- Frameworks you documented but never packaged
- Knowledge you encoded but never productized
- Tools you configured but never connected to payment
- IP that’s “almost ready” but has been “almost ready” for months
If you found something - you might be one session away from a product line.
The hard part was never the execution. The hard part was the blueprint. And you already drew it.
The 8 tools built in this session are live at studio.pickaxe.co. Free to try. Built on the nSENS validation framework.
I run workshops on building this kind of AI partnership - where your AI becomes a co-builder, not just a chatbot. If you want to learn how to turn your drawer into products, reach out on LinkedIn.