The Winning Intelligence Series: A New Theory of the Firm in the Age of AI
Existing theories of the firm ask three questions: why do firms exist (Coase), what makes them competitively different (Barney), and how do they adapt to change (Teece). The Winning Intelligence series proposes a fourth question that these frameworks were not built to answer: how do firms compound intelligence over time, and how do they govern the compounding so that it moves in the right direction? Across eight papers, the series develops a theory in which competitive advantage is not a stock a firm possesses but a flow it governs — a continuous process of accumulating, directing, and compounding both machine intelligence and human creative capability through deliberate organizational design.
Paper 1 — The AI Intelligence Stack argues that firms compete across five layers of AI capability — data, models, decisions, workflows, and interfaces — and that as AI commoditizes each layer, sustained advantage migrates upward to a Purpose and Values Layer that specifies what the intelligence stack should be doing and for whom.
Paper 2 — Programmable Strategy shows how strategy is no longer formulated and then implemented but compiled: translated layer by layer from human intent through AI decision logic into system behavior, making the quality of that translation the primary determinant of strategic execution.
Paper 3 — The Strategic Envelope develops the governance architecture that bounds what AI systems may do: not a constraint on AI action but the primary mechanism through which human purpose and values are operationalized, requiring deliberate design before AI systems are capable of circumventing it.
Paper 4 — Learning Power introduces the firm’s machine-side compounding formula — Learning Power = Density × Velocity × Scale × Directionality — showing that machine intelligence compounds through structured engagement with data, and that Directionality, the governing dimension, is what organizations most consistently fail to manage.
Paper 5 — From Alliances to Swarms reframes competition itself: as AI agents increasingly act as autonomous economic actors, the unit of competition shifts from the firm to the swarm, and advantage accrues to the organizations that design swarm coordination protocols and values specifications that reliably align collective AI behavior with organizational intent.
Paper 6 — Programmable Incentives shows how AI enables behavioral orchestration at organizational scale — aligning human and AI behavior around strategic intent through real-time adaptive incentives — while identifying the conditions under which orchestration fails if the values layer is not sovereign over the measurement systems it employs.
Paper 7 — Generative Power introduces the human-side compounding formula — Generative Power = Depth × Velocity × Variety × Trajectory — as the parallel to Learning Power, specifying how human creative capability compounds through AI-mediated interaction when that interaction is designed prosthetically rather than substitutively, and developing the Creativity Stack as the operational architecture for building it.
Paper 8 — Agentic AI and Superintelligence addresses what happens when AI agents move beyond assisting and begin producing frontier knowledge autonomously, identifying the four human roles (Orchestrator, Interpreter, Verifier, Applier) that agentic knowledge systems require, and introducing the knowledge sovereignty trap — the recursive dynamic in which more capable AI increases governance requirements while simultaneously, through substitutive deployment, eroding the governance expertise available.
Theoretical contribution and distinctiveness
The series makes four contributions that distinguish it from the existing literature on AI and organizations.
The first is the compounding theory of competitive advantage. The existing frameworks — the resource-based view, dynamic capabilities, absorptive capacity — describe how firms acquire and reconfigure resources. The Winning Intelligence series describes how firms compound intelligence: the machine-side compounding of Learning Power and the human-side compounding of Generative Power are the twin engines of advantage in the Age of AI, each governed by a directing dimension (Directionality; Trajectory) that organizations consistently fail to measure and therefore consistently allow to drift toward the wrong default.
The second is the Design Not Destiny principle applied throughout. The series consistently argues that AI’s impact on organizations — on strategy, on human capability, on knowledge production, on competitive dynamics — is not technologically determined. It is the product of design choices that most organizations are not yet making deliberately. The series specifies those choices at each level: the Intelligence Stack, the strategic envelope, the learning power architecture, the Creativity Stack, the four human roles in agentic knowledge production.
The third is the parallel treatment of machine intelligence and human creativity as two sides of the same compounding flywheel. No existing framework addresses both simultaneously at the level of theoretical specification. The Learning Power and Generative Power formulas are parallel in structure, share the Velocity dimension, and both require governance of a directing dimension that organizations systematically neglect. This duality gives the series an integrative architecture that neither the AI strategy literature nor the human capital literature provides alone.
The fourth is the Knowledge Sovereignty Trap as a conceptual contribution at the superintelligence horizon: the recursive dynamic in which organizations that fail to invest in human governance capacity during the current window will arrive at the threshold of broad superintelligence least equipped to govern it. This is a genuinely new framing of an AI governance challenge that the existing literature has not named.
The post The Winning Intelligence Series: A New Theory of the Firm in the Age of AI appeared first on Portulans Institute.