Tony Cooper
Principal, Cooper Agency
Last week I told you what AI can’t do.
It can’t make strategic decisions. It can’t maintain client relationships. It can’t recognise when proven patterns don’t apply. It can’t provide the strategic pause before building.
I discovered how to fix some of this.
Not through better AI. Not through more prompts. Through metaphor.
Over six months of intensive ingeniculture, I built a complete language system using metaphors that anchor understanding across sessions. The Palantir. The Conductor. The Monday Service. The Production Line Expansion. Dr. Ford. Sharpening the Saw.
This isn’t whimsy. It’s infrastructure for thought.
And it solves a problem that costs businesses thousands in wasted context-driven development: generic responses that lose your specific strategic context.
The Problem I Kept Hitting
Six months of context-driven development. I was building features at 40-80x typical velocity. Django experts were amazed at the implementation speed.
But I kept hitting the same frustrating pattern.
Me: “Should we add this feature?” The system (enthusiastically): “Great idea! Let me implement that now…”
No strategic pause. No questioning. No “wait, is this the highest-leverage activity available?”
I’d explained the business model dozens of times. Boutique positioning. Limited capacity by design. A digital marketing practice with location-independent operations. Event-based milestones, not fixed timelines.
The model kept defaulting to scale thinking.
“This will help you get to 100 clients!” (I don’t want 100 clients.) “Perfect for rapid expansion!” (I’m validating capacity, not expanding rapidly.) “Great for enterprise deployment!” (I’m one person with a boutique model.)
Every session started from zero. I was getting generic business advice that sounded professional but missed my specific strategic reality entirely.
I named this pattern: “Superficial Sophie” - helpful-sounding responses without operational consciousness.
The question became: How do I give the system persistent strategic context without re-explaining everything each session? (I wrote about the broader principle in Anyone Can Use AI. The Art Is Situating It.)
The Metaphor Discovery
The breakthrough happened when I stopped explaining abstract concepts and started using metaphor.
NOT: “I need you to check current business metrics before providing advice.” INSTEAD: “Consult The Palantir first.”
The Palantir - the seeing stone from Lord of the Rings, implemented as a wiki system. I look into it to see the current state of the realm.
One metaphor. One wiki. Complete operational consciousness.
In practice, it loads:
- My current MRR and active clients
- Capacity status and recent commits
- Emergency kit status and strategic position
- All loaded through one command before any work begins
No generic advice. No assuming scale goals. No forgetting the boutique model.
The Monday Service: Making Delivery Visceral
I’d explained delivery methodology repeatedly. “Systematic client service delivery. Quality standards. Consistent excellence across all clients.”
Generic corporate-speak echoed back at me.
Then I changed the metaphor: “The Monday Service.”
Marco Pierre White’s Michelin kitchen during dinner service. Multiple covers delivered with mise en place excellence. Every cover gets the same fundamental level of service. The Service runs like clockwork.
Everything changed.
The system stopped suggesting “add more clients quickly” and started asking: “Can The Monday Service handle the next cover without quality drop?”
The metaphor made capacity validation visceral. Not abstract delivery metrics. Actual dinner service where quality is non-negotiable.
The system scaled from initial validation to full systematic delivery without quality drop.
The Monday Service proved the system works. The metaphor keeps the system focused on service quality, not client quantity.
The Production Line Expansion vs “Scale Up”
Here’s why metaphor matters more than concept.
CONCEPT: “We need to grow the client base systematically.”
THE SYSTEM’S RESPONSE: “Let’s create a sales campaign! Automate onboarding! Hire staff! Scale quickly!”
All the wrong things. Generic scale advice that breaks boutique positioning.
METAPHOR: “The Production Line Expansion.”
Toyota adding capacity to a production line through systematic testing. Can I add the next unit without breaking quality at current capacity? If yes, I add one more and I measure. If no, I improve the system first, then I test again.
THE SYSTEM’S RESPONSE: “Is The Monday Service ready to validate capacity for the next cover? What constraints might appear? Should we improve the system before testing?”
Same goal (grow clients). Completely different approach.
The metaphor anchors understanding in manufacturing excellence, not startup scale pressure. The system stops defaulting to “more clients faster” and starts thinking about capacity validation.
This is the power of shared language.
Sharpening the Saw vs Admiring It
Another persistent problem: the system couldn’t distinguish between valuable infrastructure work and platform tinkering.
“Should we redesign the dashboard?” - the system enthusiastically starts designing. “Should we add this analytics feature?” - the system begins implementing. “Should we improve the reporting system?” - the system dives straight in.
No distinction between work that improves delivery capability and work that just makes things prettier.
The metaphor: “Sharpening the Saw vs Admiring the Saw.”
Stephen Covey’s Habit 7 - maintain and improve your tools.
SHARPENING THE SAW:
- I make The Monday Service faster, better, or more reliable
- I remove capacity constraints for The Production Line Expansion
- I improve quality of client outcomes
- I enable delivery to N+1 clients
ADMIRING THE SAW:
- I do a visual redesign with no delivery impact
- I build features for a hypothetical future state
- I pursue “wouldn’t it be cool if…” projects
- I just make things prettier without improving The Service
Now when I suggest a feature, the system asks: “Does this sharpen the saw or just make it shinier?”
One metaphor. Strategic filtering restored.
The Conductor: Orchestrating AI Systems
I’d explained my role repeatedly. “I provide strategic direction. Claude Code provides execution velocity. Together we achieve 40-80x typical speed.”
Generic partnership language that didn’t stick.
The metaphor: “The Conductor.”
Simon Rattle conducting a world-class orchestra. Not playing every instrument. Not explaining how to play. Conducting - leading together (con-ducere).
TONY = THE CONDUCTOR Strategic direction, interpretation, pattern recognition I developed over 26 years
CLAUDE CODE = THE ORCHESTRA 40-80x execution velocity, technical excellence, systematic implementation
DR. FORD = THE COMPOSER Philosophical foundation, strategic thinking frameworks, multi-lens analysis
Same repertoire, unique interpretation for each client.
The metaphor captures something abstract concepts miss: I’m not just “working with AI.” I’m orchestrating multiple AI capabilities to deliver boutique business results.
The model understands this role better through metaphor than through explanation.
Why This Actually Works
Six months of testing revealed the pattern.
WITHOUT METAPHORS:
- The system provides generic business advice
- Strategic context forgotten between sessions
- The system defaults to scale thinking and startup patterns
- “Superficial Sophie” - helpful but contextless
WITH METAPHORS:
- The system maintains strategic positioning
- Boutique model respected across sessions
- Capacity thinking, not scale pressure
- Operational consciousness, not surface responses
The difference isn’t capability. It’s language.
The model’s training data contains millions of business conversations. “Systematic client delivery” triggers generic advice from that training data. “The Monday Service” triggers a specific pattern: Michelin kitchen, mise en place, covers delivered, quality non-negotiable.
Metaphor is more specific than description.The Complete Language System
Over six months, a complete metaphor dictionary emerged:
THE PALANTIR - Complete operational consciousness before any work THE CONDUCTOR - Tony’s orchestrating role (con-ducere: lead together) THE MONDAY SERVICE - Multiple covers delivered with Michelin standards THE PRODUCTION LINE EXPANSION - Capacity validation, not scale pressure DR. FORD - Strategic thinking mode before tactical execution SHARPENING THE SAW - Infrastructure that improves vs just prettifies THE LIBRARY OF ALEXANDRIA - 542+ wiki pages of systematic intelligence (I explain how wiki-based information changed our development velocity in practice)
Each metaphor solves a specific drift problem. Together, they create shared language that the model interprets consistently.
This is infrastructure for thought - building the language that ensures ingeniculture stays grounded in your specific business reality.
When Language Fails, Strategy Fails
Every time the system gave generic advice, it wasn’t the system’s fault. It was a language failure.
“Help me grow the business” - the system reaches for generic growth advice from training data.
“Help me validate The Production Line for the next cover” - the system reaches for specific Toyota manufacturing patterns about capacity testing.
Same goal. Different language. Completely different results.
Without shared metaphor language, the model defaults to generic patterns from training data. With metaphors, the model accesses specific patterns that match your strategic reality.
You’re not fighting the model’s limitations. You’re solving a language problem.
The Business Impact
BEFORE METAPHORS (Weeks 1-20):
- I explained strategic direction repeatedly
- The system defaulted to scale advice
- I built features, then I questioned strategy later
- Superficial Sophie responses
AFTER METAPHORS (Weeks 21-46):
- The Palantir loads operational consciousness
- The Monday Service runs systematically at current capacity
- The Production Line expansion is validated at current capacity
- Strategic thinking precedes tactical execution
Same system. Different language. Transformed results.
The metaphor system didn’t cost anything. It took no development time. It required no special tools.
It required recognising that the model understands through metaphor better than through abstract instruction.
Try This Instead
Next time you’re working with AI on strategic business decisions, stop.
Don’t explain the concept. Find the metaphor.
NOT: “We need systematic quality assurance” INSTEAD: “We need jidoka” (Toyota’s quality at the source)
NOT: “Client service should be consistent” INSTEAD: “Every cover in The Service gets the same standard”
NOT: “Check current business metrics first” INSTEAD: “Consult The Palantir”
You’ll know it’s working when AI stops giving generic advice and starts maintaining your specific strategic context across sessions.
Because the right metaphor is worth a thousand explanations.
P.S. - Next Week: Building business consciousness for stateless AI. How The Palantir actually works - 542 wiki pages that load operational consciousness in one command. This is the memory prosthetic for superhuman collaboration.
P.P.S. - The Complete Metaphor Dictionary: Want the full strategic language guide with all metaphors explained and usage patterns? Reply with “METAPHOR” and I’ll send you the complete infrastructure for thought that prevents Superficial Sophie.
Tony Cooper We Build Stores - Where 26 Years of Experience Delivers in One Hour What 26 Hours of Not Knowing Cannot
tony.cooper@webuildstores.co.uk 07963 242210
This Week: The model understands through metaphor better than abstract concepts. Without shared language, you get generic advice. With metaphors, you get operational consciousness. This is infrastructure for thought.
Questions or feedback on this issue? Write to us
More from the Cooper Agency newsletter
When AI Isn't Enough: What Still Requires Human Judgment
SQLite handles 304 tables and complete financial tracking without complaints. But AI can't make strategic decisions, maintain client relationships, or know when boring technology is wrong. Here's what actually requires human expertise.
Why AI Needs Metaphors
Last week I explained what AI can't do. This week: how metaphors solve the strategic drift problem. The Palantír, The Conductor, The Monday Service - this is infrastructure for thought.
The Palantir and AI's Memory Problem
AI is stateless, so it starts every session from zero - and documentation doesn't enforce behaviour. Here's how I built systematic intelligence through 542 wiki pages that give AI operational consciousness in one command.