Reshuffle

1. AI's True Impact Lies in Coordination, Not Just Automation

AI’s real power lies not in automating individual tasks but in coordinating entire systems.

Intelligence distraction. We often fixate on AI's human-like intelligence, asking "How smart is it?" or "What can it do?" This "intelligence distraction" blinds us to its true economic and systemic implications. AI's power isn't in replicating human thought but in its performance and ability to adapt to complexity through pattern prediction.

Systemic blind spots. The Singapore KTV lounge COVID-19 outbreak illustrates how even advanced technology (contact tracing, vaccines) fails when deployed in a system with hidden fault lines and interdependencies. The tools were sound, but the system behaved differently than assumed, leading to failure. This parable highlights the limits of technological solutionism.

Coordination, not just speed. Like the shipping container revolutionized global trade by enabling coordination across disparate logistics, AI's five core functions—sense, model, reason, act, learn—make it uniquely suited to solve coordination problems. It transforms how decisions are made and where attention is directed, changing the structure of the system it enters.

2. Beyond Jobs: AI Reshapes Entire Systems of Work

Your job exists because the current system of work needs someone to handle certain tasks and responsibilities in a specific way. But if the system of work itself changes, that set of tasks and responsibilities may no longer hold the same value in the new system.

The wrong frame. The common mantra "AI won't take your job, but someone using AI will" offers false reassurance. It assumes jobs are stable bundles of tasks and AI merely augments or automates them. This task-centric view misses how AI fundamentally reconfigures the entire system of work, much like the Maginot Line was built for an outdated system of warfare.

System-centric view. Jobs are temporary groupings of tasks that make sense within a specific system. When AI changes the system, the job bundle gets unbundled, and tasks are reassembled into new roles.

• Typists: Word processors didn't eliminate typing but collapsed the cost of editing, making the dedicated typist role redundant.
• Basketball: Analytics shifted focus from fixed positions to versatile players, rebundling skills around team-level coordination.
• Amazon Kiva: Robots aren't just augmenting workers; they're components in a fulfillment system that continually restructures workflows to minimize human variability and maximize predictability.

Value and agency. The real issue isn't whether a job still exists, but whether the work still holds economic value and if the worker retains control or agency. AI can diminish both, even if tasks persist, by making tacit knowledge explicit and centralizing algorithmic management.

3. Coordination is the New Power: AI Rewrites Value Distribution

A growing system does not mean shared prosperity. In fact, growth and inequality often rise together, because the very mechanisms that expand the pie also determine how it’s divided.

The growing pie, unevenly sliced. Debates about AI often polarize between job loss and universal prosperity. The reality is that AI expands the economic pie, but the mechanisms that enable this growth also determine how value is distributed, leading to new forms of inequality.

• Walmart vs. Kmart: Both adopted barcodes, but only Walmart built a "barcode-native" architecture for centralized coordination, gaining leverage over suppliers and capturing a disproportionate share of value.
• British Empire: Controlled India not through direct ownership but by redesigning its system through five levers of coordination: representation, decision-making, execution, composition, and governance.

Coordination without consensus. Traditionally, coordination required either a dominant player imposing standards (Walmart) or participants agreeing on shared rules (containerization). AI offers a new path: coordination without upfront consensus.

• Climate Corporation: Used AI to create a digital representation of farms, defining performance on its terms and shifting power to itself, even without farmers' explicit agreement on data standards.
• Trucking platforms: Centralize pricing and routing, reducing truckers' autonomy and pricing power by making their tasks measurable and substitutable, even without self-driving trucks.

New tensions. AI's coordination capabilities introduce new tensions: between workers and tools (loss of autonomy), within organizations (centralization vs. decentralization), and across competitive ecosystems (tool providers gaining leverage over solution providers).

4. Rebundle Your Role Around New Constraints, Not Just New Skills

Work is not simply a bundle of tasks but a response to constraints in a larger system, and when constraints shift, the value of work migrates with them.

Beyond reskilling. The constant treadmill of reskilling for new tasks is a losing game. Jobs are structured to resolve constraints in a larger system. When a constraint collapses, the logic for the role weakens, regardless of individual skills.

• Sommelier: Wine knowledge became commoditized, but sommeliers gained value by managing new constraints: overwhelming choice, emotional risk, and the need for curated experience.
• Fugu Chef: Commands a premium not for culinary skill, but for managing the lethal risk associated with the dish, highlighting the value of absorbing liability.

Types of constraints. Value shifts when constraints move. Understanding these types is crucial:

• Scarcity-based: Limited supply of resources or knowledge (e.g., radio operators before automated communication).
• Risk-based: Actions with significant consequences (e.g., anesthesiologists managing patient stability during surgery).
• Coordination-based: Many moving parts needing alignment (e.g., movie producers coordinating complex productions).

Rebundling around new constraints. When AI removes scarcity-based constraints (e.g., expert knowledge), new risk-based and coordination-based constraints emerge.

• Radiologists: As AI excels at image analysis, radiologists rebundle their role around clinical judgment and contextualizing findings within a patient's broader case.
• Nurse Navigators: In digitized hospitals, they manage coordination gaps and patient experience, gaining contextual value by resolving system-level friction.

5. AI Transforms Organizations from Silos to Coordinated Systems

Coordination is not just how players pass the ball or how teams communicate. It’s the shared mental model of the field, where everyone knows where they will be when the ball is passed.

The coordination tax. Organizations face a trade-off between team autonomy and cross-team coordination. Too much autonomy leads to silos (Real Madrid's Galácticos), while too much coordination stifles innovation. Coordination failures, like NASA's Mars Climate Orbiter disaster, impose hidden costs.

• Organizational knowledge: Historically managed through clerical roles and tools like filing cabinets, but still fragmented across emails, chats, and reports.
• Coordination tax: The silent cost of redundant meetings, chasing documents, and rebuilding shared context across teams.

Eliminating the coordination tax. AI offers the first scalable antidote by transforming knowledge management:

• Encoding knowledge: AI extracts structured information from unstructured sources (voice notes, contracts), codifying tacit knowledge.
• Organizing knowledge: AI synthesizes scattered data into actionable insights, reducing clerical burden.
• Deploying knowledge: AI proactively serves relevant insights into workflows, reducing the need for constant meetings (e.g., Ramp's customer podcasts).

Agentic execution. AI enables goal-oriented systems that execute chains of decisions and actions semi-autonomously.

• Ramp's sales reps: AI agents monitor funding news, extract emails, and refine message templates, scaling lead generation without manual oversight.
• Palantir Foundry: Unbundles organizational knowledge from silos, creating a shared asset and enabling agentic workflows for decision-making and execution.

6. The Building Blocks Economy: Rebundle Capabilities for New Advantage

When knowledge is unbundled from human labor and becomes accessible as capital, it gains three essential traits. It becomes rentable, as you can access it without long-term commitments. It becomes recombinable, since different forms of expertise can be recombined without the overhead of coordinating across siloed teams. And it becomes scalable: once a solution is built, it can be deployed repeatedly at near-zero marginal cost, unlike human labor, which scales linearly with cost.

Capabilities as building blocks. Just as cloud computing unbundled infrastructure, AI unbundles expertise from human labor, making it rentable, recombinable, and scalable. This enables new business models.

• MrBeast Burger: Launched an empire overnight by recombining rented building blocks: network capital (audience), ghost kitchens (Virtual Dining Concepts), and third-party delivery.
• Solopreneurs: Leverage AI for content creation, outreach, and analytics, running knowledge-intensive processes at scale without traditional workforces.

Leverage, not just labor. Power shifts from owning assets to knowing how to rebundle modular components.

• Michael Smith (Spotify fraud): Exploited Spotify's algorithm by manufacturing demand with bots and supply with AI-generated music, demonstrating "algorithmic awareness."
• SAG-AFTRA strikes: Studios seek to commoditize actors' likenesses into reusable digital assets, creating tension between capital (studios) and labor (actors). OnlyFans creators, with network capital, can design systems where AI avatars complement their persona.

Positive constraints. Success in the building blocks economy depends on managing constraints.

• Negative constraints: Hidden bottlenecks that weaken the system (e.g., MrBeast Burger's quality control issues).
• Positive constraints: Deliberately chosen design limitations that proactively shape system behavior (e.g., Muji's minimalist branding creating a unique supply chain).

7. AI as an Engine, Not Just a Tool, Redefines Competitive Advantage

A company that utilizes AI as a tool may improve efficiency, but it still competes on the same basis. A company that treats AI as an engine unlocks entirely new levels of performance and changes the basis of how it competes.

Tools vs. Engines. A tool performs an isolated function; an engine is a performance-driving component around which the entire system is structured.

• Formula 1: Renault won by building an integrated system around its engine, not just optimizing individual components like Ferrari.
• TikTok: Used AI as an engine to build a "behavior graph" (what you watch) rather than a "social graph" (who you know), making traditional network effects irrelevant and redefining social networking.

Tool integration trap. The more a solution provider integrates a third-party AI tool, the more dependent they become on the tool provider.

• Google Maps vs. Uber: Uber's performance was tied to Google Maps, giving Google leverage. Uber eventually invested in its own mapping.
• Tool provider advantages: Learning advantage (data from multiple customers), scope expansion (horizontal rebundling, vertical encroachment), and clockspeed (faster innovation).

Shifting pricing power. When AI tools become industry engines, they become chokepoints, shifting value from solution providers to tool providers.

• Charlie Munger's loom analogy: Productivity gains in commodity markets accrue to technology providers or end customers, not the producers in between.
• Netflix: Created its own content to avoid studios increasing licensing fees as its streaming platform grew.

8. Solutions Absorb Risk: The Moat Against Commoditization

Tools amplify performance, but solutions absorb risk. And it is that absorption of risk that assures a customer of the solution’s viability.

Tools vs. Solutions. Tools offer capability; solutions deliver reliability and performance in the real world by solving for cost, complexity, and change.

• Robotics in manufacturing: Adoption is low because robots are not accessible (high upfront cost), usable (disrupt workflows), or reliable (require specialized maintenance).
• Residential solar: Adoption was unlocked by financial innovation (leasing, PPAs) that absorbed upfront cost and maintenance risk from homeowners.

System builders. These orchestrate the entire architecture needed for a tool to function, solving constraints across financing, operations, software, and services.

• Formic: Offers "robots-as-a-service," leasing robots by the hour, designing custom setups, and absorbing downtime risk. It uses customer contracts as collateral to fund deployments, aligning its success with customer productivity.
• Erie Canal: Reengineered economic infrastructure by managing every critical constraint (geographic, financial, technical, operational) to guarantee predictable transport.

Business models of solution providers. Different levels of risk absorption lead to different models:

• Work-as-a-Service: Charges for keeping equipment running smoothly (e.g., Winterhalter's "Pay per Wash," Rolls-Royce's "power by the hour").
• Results-as-a-Service: Charges for guaranteed outcomes (e.g., Orica's "Rock-on-Ground" contracts for optimal rock fragmentation).
• Outcome-as-a-Service: Ties profitability to strategic customer outcomes (e.g., pharmaceutical Personal Reimbursement Models linked to therapeutic response).

9. Simplify Decisions: Turn Customer Confusion into Control Points

In a crowded ecosystem of competing providers, AI offers a way to reduce decision friction and guide customers through journeys that build trust at every step.

Decision support as advantage. In a world of abundant choice, simplifying customer decisions builds loyalty and competitive advantage.

• Best Buy: Bucked Amazon's dominance by turning stores into "decision-support hubs," training staff to guide customers through complex electronics purchases. It used price parity to retain customers after they made decisions in-store.
• Sephora: Used Color IQ scanners to simplify foundation matching, creating an entry point into the customer journey and rebundling related services (tutorials, virtual try-ons).

Control points. These are strategic positions in the customer journey where reducing complexity earns trust, locks in customers, and creates leverage over adjacent players.

• Relationship-based: Local car dealers, salon stylists.
• Workflow-based: Figma owning design collaboration.
• Intelligence-based: Google Maps/TripAdvisor shaping travel decisions.

Direct vs. derived demand. Companies owning direct demand (primary needs) dictate terms to those in derived demand (products satisfying those needs).

• Sephora: Owns direct demand for "beauty and confidence," dictating terms to beauty brands (derived demand).
• AI assistance: Can be deployed through intermediaries (salon professionals) to build trust and shift liability, creating intelligence-based control points for brands.

10. Control Without Consensus: AI Orchestrates Fragmented Ecosystems

Control isn’t established just by owning the interface. It’s established by resolving the coordination burden that stands between users and the outcomes they seek.

Alexa's failure. Amazon's Alexa, despite scale and ambition, failed to dominate the smart home because its AI couldn't effectively coordinate fragmented partners. It lacked contextual intelligence and couldn't chain intents across skills, leaving the coordination burden on users.

• Control by dependence: In ecosystems, control is earned by solving the hardest coordination problem for partners, making you indispensable.

Establishing control through coordination.

• CCC Intelligent Solutions: Coordinates auto insurance claims across insurers, repair shops, and suppliers. It established control by:
  • Unified representation: Creating a shared digital model of damaged vehicles.
  • Unified decision: Aligning insurers and shops on repair costs.
  • Unified execution: Integrating workflows for approvals and parts procurement.

Coordination without consensus. AI enables coordination where consensus is impractical.

• Tractable: A competitor to CCC, uses AI to analyze unstructured smartphone photos for repair estimates, achieving unified representation, decision, and execution without requiring standardized codes or upfront agreements from participants.
• Construction industry: AI extracts design requirements and project dependencies from disparate tools (CAD, Excel, PDFs), creating a unified project representation and coordinating actions across architects, engineers, and contractors without forcing tool adoption.

AI-driven rebundling. AI unifies fragmented information, decisions, and execution across tools, teams, and workflows, creating functionally integrated systems from fragmented ecosystems.

• Rox (enterprise sales): Consolidates customer data from CRM, emails, calls, and marketing platforms into a unified view, then uses agentic execution to automate follow-ups and sales workflows.
• Ramp (finance automation): Integrates corporate cards, bill payments, and expense tracking into a single, AI-coordinated system, enforcing policies and automating transactions across the enterprise.

11. Beyond an "AI Strategy": Reshape the Playing Field, Don't Just Play Faster

True strategic advantage in the age of AI doesn’t come from doing old things faster. It comes from seeing how the system itself is changing, and then deliberately reshaping it to your advantage.

Misunderstanding strategy. In rapidly changing environments, traditional strategy (positioning in stable markets) fails.

• Chegg's collapse: Exemplifies how a shifting playing field (COVID-19, then ChatGPT) can render market leaders obsolete overnight, despite continuous reinvention.
• Common traps: Chasing short-term wins, overemphasizing executional agility (Yahoo), or perpetual experimentation without clear direction.

Getting AI strategy right. Don't start by asking what AI can automate. Start by asking:

• Where to play: What new system is emerging, and what role can we play at its center? (e.g., containerization expanding trade, TikTok shifting social networking to behavior graphs).
• How to win: How do we establish control points by identifying and resolving the new constraints in this evolving system? (e.g., Singapore's port strategy combining intermodal coordination and risk management).

Strategic postures to flux. Companies leverage AI with different approaches:

• Reactive optimizers: Speed up existing tasks (e.g., agribusiness predicting farm yield better).
• Anticipators: Spot new opportunities but remain in old structures (e.g., precision agriculture improving input efficiency without changing farm decision-making).
• Logic shifters: Rewire who makes decisions and how (e.g., John Deere moving farm input decisions into machines).
• Field reshapers: Reorganize the entire system and coordinate multiple players (e.g., Climate Corp integrating seed, input, risk, and carbon markets).

Reshuffle or get reshuffled. The apparent chaos of technological change conceals a new structure. AI offers the power to redesign the playing field, tilting it in your favor by mastering coordination and control.

Last updated: March 10, 2026