--- chapter: 02 part: Map title: The Co-Intelligent Vision subtitle: Not Humans OR AI. Humans WITH AI. page: page-0024 readingMinutes: 15 published: 2026-05-13 status: published version: racecraft-v2 --- ## Not Humans OR AI. Humans WITH AI. In early days of automobiles, we called them "horseless carriages." The name revealed our thinking. We understood new technology only in terms of what it replaced. A carriage—without a horse. The frame of reference was old world. It took decades before we stopped thinking about what was missing and started thinking about what was possible. Roads, highways, suburbs, supply chains, global commerce—none of it was imaginable when we were still mentally tethered to horse. ![Beyond horseless-carriage replacement thinking](/images/book/ch01/ch01-beyond-horseless-carriage-replacement-thinking.png) **We're making the same mistake with AI.** Most organisations think about AI in terms of replacement. Which jobs will it eliminate? Which tasks can it do instead of humans? How many people can we remove? This is horseless carriage thinking. ![FIGURE:ch01-two-modes](figure) The Co-Intelligent Organisation doesn't ask: "What can AI do instead of us?" It asks: **"What can we become together?"** ## The Definition A Co-Intelligent Organisation is one where human and artificial intelligence are designed to amplify each other. In this book, we'll keep returning to Formula 1 as our living laboratory for that question — the sport where human and machine already think together in real time, under race conditions. Not humans using AI tools. Not AI replacing human workers. A genuine synthesis—where neither could achieve alone what they accomplish together. **Four characteristics define a Co-Intelligent Organisation:** ![The Four Characteristics of a Co-Intelligent Organisation](/images/book/ch01/ch01-four-characteristics.png) **1. Leaders orchestrate human and AI intelligence** The leader's job isn't to do the work. It's to direct the cognitive resources—human and artificial—toward outcomes. Like Hannah Schmitz on the pit wall, receiving AI simulations and human intuition, then making the call. The old identity: "I'm the leader of humans and expert who knows the answer." The new identity: "I'm the conductor who synthesises the system." As **Ruth Buscombe**, former Head of Race Strategy at Alfa Romeo F1, puts it, gut instinct is "a summation of all your experiences" and "a summation of numbers that are going on in your head that your conscious brain isn't aware of." The Co-Intelligent Leader uses this "new intuition"—a synthesis of data and experience. **2. Teams include AI as collaborative members** In Lloyds Banking Group's Project Turing experiment, three teams competed: human-only, AI-only, and human+AI. The hybrid team's work was judged "most human" by a blind panel. The explanation is counterintuitive. AI didn't replace human creativity—it gave humans more time for it. Volume enabled depth. When AI handled the collecting, humans could focus on the connecting. The Co-Intelligent Organisation designs teams where AI is a member, not just a tool. **3. Structures are designed for human-AI workflows** BNY Mellon has 100 "digital employees" in production. They have login credentials. Email addresses. Human managers. Onboarding processes. Performance reviews. This sounds like a stunt. It's not. It's the inevitable conclusion of treating AI as organisational infrastructure, not just technology. If AI is doing work, it needs to be managed, governed, and integrated—just like human workers. The Co-Intelligent Organisation doesn't bolt AI onto existing structures. It redesigns structures for human-AI flow. **4. Workflows continuously learn and adapt** The most important shift isn't about adding AI. It's about reimagining and building learning loops. A traditional workflow is static: input → process → output. A co-intelligent workflow evolves: input → process → output → feedback → adaptation → improved process. The AI learns from the work. The humans learn from the AI's patterns. The workflow gets better over time. This is what separates automation from intelligence.
Pit Stop · 02.1

Design the relationship, not the tool.

Co-intelligence is not a feature list. It is the designed relationship between leadership, teams, structures, and workflows so the system can think better than either humans or AI could alone.

--- ## This Isn't Just Theory The concept of co-intelligence might sound visionary, but organisations are already living it. Let me show you two contrasting examples—one gradual, one rapid—that reveal the same pattern. ### Commonwealth Bank: The Shield's Patient Build Commonwealth Bank, Australia's largest bank with 16 million customers, took the Shield archetype approach to co-intelligence. They started with static risk models and advisory systems but spent five years on something most organisations skip: **the data foundation**. Five years. Fixing their data lake before scaling AI. By 2024, this patience paid off. They migrated 61,000 data pipelines to AWS cloud with zero downtime. They deployed ChatGPT Enterprise—largest in global banking. They launched "Benefits Finder," which found $1B+ for customers. Today, Commonwealth Bank processes **55 million AI decisions daily** across 2,000+ AI models. They have moved from basic operator and advisory use cases into genuine partnership and optimisation territory—using AI not just to assist staff, but to shape and increasingly run dynamic decision systems under strong governance. The result? App assessment time dropped from 6 weeks to 1 hour (99% reduction). Modernisation cycles went from 16 weeks to 1 week (94% reduction). **The lesson:** Commonwealth Bank treated AI as organisational transformation, not a technology project. They built trust capital through "AI for Good" initiatives—helping customers find money rather than just selling products—before deploying at scale. ### Klarna: The Shotgun's Rapid Push Klarna, the fintech Buy Now Pay Later company with 150M+ consumers, took the opposite approach—The Shotgun archetype. In 2024, CEO Sebastian Siemiatkowski made a directive: **"Replace human work with AI"** to reach profitability. They pushed aggressively into high-agency automation. AI handles 2/3 of customer chats autonomously. Resolution time dropped from 11 minutes to <2 minutes (82% faster). They reduced staff from 5,000 to 3,000 (40% reduction)—replacing 700 full-time agents with AI. The financial impact? $40M profit improvement in 2024, with $60M projected by late 2025. **But something went wrong.** Customers got stuck in "bot loops"—endless automated conversations that couldn't resolve complex issues. The over-automation damaged trust. Klarna had to re-hire humans and implement an "escape hatch" for complex queries. They corrected course toward a more sustainable hybrid: AI handles simple queries, humans handle complex ones, and the system is governed around the work's actual complexity rather than a blanket ideology of replacement. **The lesson:** Speed without service quality = trust withdrawal. Klarna proved that the most effective customer service is co-intelligent—AI for speed/efficiency, humans for empathy/judgment.
Pit Stop · 02.2

Speed spends trust.

Commonwealth Bank and Klarna moved at different speeds, but both ran into the same physics: higher-agency AI only works when the organisation has enough trust capital to let the machine act.

### The Common Thread Both Commonwealth Bank and Klarna are successful co-intelligent organisations, despite radically different speeds. **Commonwealth Bank:** Gradual, patient, 8-12 year journey. Built from low-agency tools into partnership and selective optimisation. Trust was earned through governance and reliability first. **Klarna:** Rapid, aggressive, 2-3 year journey. Pushed from assistance into high-agency service automation too fast, then corrected when speed compromised quality. **Both learned:** You cannot move into higher-agency, higher-complexity territory without building trust capital. Commonwealth Bank built it through governance excellence (99.999% reliability, gold-standard AI ethics). Klarna built it through measured value ($40M profit improvement) and then learned the cost of overreach. ![FIGURE:ch01-two-paths](figure) This isn't unique. It's the pattern of co-intelligent transformation. And it is not confined to incumbents. A new generation of AI-native firms is turning this pattern into raw business physics. Lovable reached $100M ARR in 8 months with 45 employees. Cursor reached $100M ARR on roughly $11M in funding, generating close to $9 of ARR for every $1 raised. Gamma reached $100M ARR with about 50 employees while already profitable. The top lean AI-native firms are averaging around **$3.48M in revenue per employee** — roughly **8-10x** the traditional private SaaS median. Those are not just better startup tactics. They are evidence that the co-intelligent organisation is already a real business model. --- ## The Question That Changes Everything Here's the question that separates a Co-Intelligent Organisation from everyone else: **"Who does the thinking?"** Not "how do we make decisions faster?" or "how do we reduce headcount?" or "how do we become more efficient?" Who. Does. The. Thinking. For every workflow in your organisation, someone (or something) is doing cognitive work. Analysing. Recommending. Deciding. Executing. In the old world, the answer was always "humans"—with varying degrees of tool support. In the new world, the answer can be: - Humans, with AI assisting - Humans and AI together, sharing the cognitive load - AI, with humans overseeing Each answer has implications. Different trust requirements. Different governance needs. Different organisational designs. **The 9Q Grid maps these answers.** **Your organisation doesn't have an "AI strategy." It has a position on the 9Q Grid. Or more accurately, a shape—a constellation of positions across your initiative portfolio.** That shape reveals your strategic intent. Or its absence. **The 9Q Grid is the strategic canvas for the co-intelligence era.** It doesn't ask how much AI you have. It asks where your AI portfolio is positioned, where it should be next, what kind of work it is actually doing, and whether you have the trust capital to operate there. The next chapter unpacks the 9Q Grid in full. For now, hold onto just one question: **Who does the thinking in your organisation's key workflows?** That question is the map. --- ## Why Now: The Window In October 2025, Sam Altman announced OpenAI's vision: "We're going to build 1 GW AI factories every week." One gigawatt is South Australia's (where I live) entire annual energy production. Every week. This is not hyperbole. It's infrastructure planning. The factories are being built. The compute is being deployed. The intelligence is becoming abundant. SpaceX and Google are talking about data centres in space to achieve these goals. For us, the question is this: *What does abundant intelligence mean for organisations?* **It means the constraint flips.** When intelligence was scarce, the challenge was acquisition. How do we hire enough smart people? How do we train them? How do we retain them? When intelligence is abundant, the challenge is deployment. How do we use this safely? How do we govern it? How do we build the trust required to let it act? This is the window. Right now, organisations that understand this shift have an advantage. They're building trust capital while others are still arguing about whether to adopt AI. But windows close. Within two years—maybe less—the laggards will catch up on the technology. Every organisation will have access to the same AI capabilities. The differentiator won't be the AI itself. It will be the organisational capacity to deploy it. **The trust capital you build now compounds. The trust capital your competitors aren't building doesn't.** --- ## How Long Does This Take? The journey to co-intelligence doesn't happen overnight — and the timeline depends heavily on your industry, your risk environment, and how much trust capital you have already built. Part 2 maps the readiness gates in detail. For now: the organisations that try to skip the journey almost always fail. --- ## The Flip Here's what most organisations get backwards: They think governance is the enemy of speed. In the old world, they were right. When intelligence was scarce, governance was drag. Every approval chain, every committee review, every compliance check—these slowed down the already-limited cognitive capacity of the organisation. Speed meant cutting governance. Agility meant fewer controls. **In the new world, this logic inverts.** When intelligence is abundant, governance is what lets you deploy it. Governance manufactures trust. Trust is the permission to let the machine act. Think about it: - Without governance, you can't let AI speak to customers - Without governance, you can't let AI approve transactions - Without governance, you can't let AI make hiring recommendations The organisations with the strongest governance won't be the slowest. They'll be the fastest—because they'll have the trust capital to deploy AI where others can't. This is the flip. **Old World:** Governance = Drag (slows you down) **New World:** Governance = Downforce (lets you go fast safely) In F1 terms: drag makes you slow. Downforce gives you grip. The car with maximum downforce and minimum drag wins. The Co-Intelligent Organisation designs governance for downforce, not drag.
Pit Stop · 02.3

Governance is grip.

Bad governance slows the car. Good governance lets it hold the track at speed. In an intelligence-abundant world, governance is the operating condition that turns capability into permission.

--- ## The Leader's New Job The most important shift in a co-intelligent organisation isn't in the workflows. It's in how leaders think about their role. The leader stops being the expert with the answers. They become the conductor of a system that thinks with them. Part 3 unpacks what that shift looks like in practice — and how to build the teams, structures, and governance that make it possible. --- ## The 87% Opportunity RAND Corporation reports that 80%+ of AI projects fail. McKinsey finds that 87% fail to scale beyond pilot. BCG confirms that only 7% achieve transformational value. These aren't technology failures. They're organisational failures. The organisations in the 7% don't have better AI. They have better organisational design — workflows, teams, governance, and leadership redesigned from first principles for the new physics. **The 87% are trying to win the 2026 championship with a 2021 car.** Part 2 gives you the readiness framework. Part 3 gives you the design playbook. The gap between 87% and 7% is closable — but only if you treat this as organisational transformation, not a technology project. --- ## The Road Ahead **Part 1: The Map** — Where your AI portfolio is. Where it should be. What shape you're running. **Part 2: The Licence** — What earns you the right to advance. The five readiness gates that separate the 7% from the 87%. **Part 3: The Human and The Machine** — How to design the Engine, the Crew, the Conductor, and the Container for a world where intelligence is abundant. **Part 4: The Race** — How to put it in motion. The 90-day season that turns strategy into execution. The next chapter zooms into the map itself. --- ## The Real Question Let me leave you with the question that matters most. Not "Should we adopt AI?" That question is settled. Not "How much AI should we have?" That question is misleading. The real question is: **What kind of organisation do you want to become?** An organisation where humans and AI compete for relevance? Or one where they compose together, creating value neither could achieve alone? An organisation where governance slows everything down? Or one where governance enables speed, building trust capital that lets you deploy AI where competitors can't? An organisation stuck in the 87%, running pilots that never scale? Or one in the 7%, redesigning from first principles for the new physics? **The Co-Intelligent Organisation is a choice.** Red Bull chose to redesign from first principles when the regulations changed. Mercedes chose to adapt their existing approach. The results spoke for themselves. Hamilton chose Ferrari, betting they'd crack the ground-effect puzzle. They hadn't. The greatest driver of his generation spent 2025 fighting a machine that wasn't designed for the new physics. You have a choice too. The rules have changed. The scarcity has flipped. The window is open. **What will you choose?** *Next: Chapter 3 zooms into the map itself — how the 9Q Grid works, how to locate your current position, and what it will cost to move.*

Chequered Flag03 takeaways

  1. Co-intelligence is a design choice. It is not humans using tools and it is not AI replacing humans.

  2. The new strategic question is who does the thinking. That question reshapes leadership, teams, and workflows.

  3. The map comes next. Vision without positioning is still just ambition.