Agentic Commerce: A Theory of Markets When the Shopper Is No Longer Human

The most comprehensive academic statement on AI-mediated commerce in existence.

The Paper

"AGENTIC COMMERCE: A Theory of Markets When the Shopper Is No Longer Human" is a 70-page integrated theoretical architecture published on SSRN in January 2026. It is the definitive academic statement on what happens to markets, firms, and consumers when autonomous AI agents assume the purchasing function.

This is not a whitepaper. It is not a trend report. It is not a consulting framework dressed in academic language. It is a formal theoretical architecture with defined constructs, falsifiable propositions, empirical simulation data, and multi-level analysis across micro (consumer-agent), meso (firm-platform-agent), and macro (market-societal) dimensions.

READ THE FULL PAPER ON SSRN →

Why This Paper Exists

By late 2025, the term "agentic commerce" had entered mainstream business vocabulary. IBM defined it. Salesforce published on it. McKinsey projected its market size. The World Economic Forum discussed it.

But no one had explained it.

Industry commentary described the phenomenon. What was missing was an integrated theoretical architecture that explained the structural mechanics: why this transition is happening, how it works, what it means for firms, consumers, and markets, and what testable predictions follow from the theory.

This paper fills that gap. It draws on principal-agent theory, information economics, platform economics, and marketing theory to construct a framework that is rigorous enough for academic scrutiny and actionable enough for executive strategy.

The Five-Part Architecture

Part I: The Paradigm Rupture (pp. 1–20)

The paper opens with a formal definition of Agentic Commerce — the first integrated academic definition of the discipline — and establishes the three boundary conditions that distinguish agentic commerce from adjacent phenomena: Delegation, Intermediation, and Optimisation.

Part I introduces the Spectrum of Delegation (five levels from Human-Only to AI-Autonomous), maps agentic commerce against adjacent concepts (e-commerce, conversational commerce, programmatic advertising, algorithmic trading, GEO), and presents the core constructs summary table and the Integrative Architecture diagram.

Part II: Theoretical Foundations (pp. 21–30)

Part II extends four bodies of established theory to the agentic commerce context:

Principal-agent theory is extended to algorithmic agents, identifying three levels of misalignment: consumer-agent, platform-consumer (the Shadow Principal Problem), and temporal instability.

Information economics is inverted: Akerlof's lemons problem, Stigler's search costs, and Spence's signalling theory are reconstructed for computational buyers who can eliminate information asymmetry.

Platform economics is extended to account for the platform-as-principal problem and learning-based lock-in.

Marketing theory's implicit assumptions are exposed: the AIDA model, the customer journey, brand equity, persuasion theory, and loyalty programmes are shown to be structurally insufficient when the buyer is a machine.

The central insight of Part II is the "influence to eligibility" transition: in human commerce, marketing influences purchasing decisions through psychological persuasion. In agentic commerce, marketing makes products eligible for algorithmic selection through data quality and operational verification. The discipline shifts from psychology to infrastructure.

Part III: The Integrative Framework (pp. 31–49)

The theoretical heart of the paper. Part III introduces the interdependent frameworks that constitute the Agentic Commerce architecture:

The Shopper Schism®: The structural separation of consumer and shopper, analysed across three dimensions — functional, temporal, and informational.

Delegated Consumption and Agency Costs: Three pathways of misalignment in AI-mediated purchasing.

The Agent Decision Preference Stack: The three-layer model of agent decision-making (Hard Constraints → Learned Heuristics → Real-Time Optimisation), including the 4.71× Operational Dominance Finding — the empirical result from 600 simulated purchase decisions showing that balanced agents overwhelmingly favour verifiable operational performance over accumulated reputation.

The Four Ds Framework™: Data Quality, Discoverability, Decisional Clarity, and Delivery Reliability — the four dimensions of agent accessibility.

The GEO/AIO Layer Mapping: GEO mapped to Layer 2 (Learned Heuristics), AIO® mapped to Layer 3 (Real-Time Optimisation).

The Trust Paradox™: The three-phase lifecycle of trust in agent intermediaries (Utility-Based Adoption → Monetisation-Driven Misalignment → Discovery-Triggered Disillusionment), with the Google Search historical precedent.

Algorithmic Readiness™: Five dimensions of organisational preparedness (Data Infrastructure, Process Adaptation, Capability Development, Governance Frameworks, Strategic Alignment).

The Automaton Economy™: The systemic consequences of algorithm-to-algorithm commerce — velocity transformation, opacity intensification, competitive regime change, and governance challenges.

Part IV: The Research Agenda (pp. 50–63)

Part IV presents 13 falsifiable propositions with quantitative thresholds and explicit falsification criteria — the hallmark of a testable theory, not industry commentary.

The propositions are organised across three levels:

Micro-level (Consumer-Agent): Propositions P1–P5 address delegation patterns, agency costs, trust dynamics, and preference divergence.

Meso-level (Firm-Platform-Agent): Propositions P6–P8 address competitive dynamics, platform power, and the operational versus reputational value shift.

Macro-level (Market-Societal): Propositions P9–P13 address market structure changes, price convergence, innovation effects, labour displacement, and regulatory responses.

Each proposition specifies what would need to be observed to confirm it, and — critically — what would need to be observed to falsify it. This is what separates theoretical architecture from opinion.

Part V: Implications and Conclusion (pp. 64–70+)

The paper closes with structured implications for four audiences:

For marketing theory: Brand equity must be reconceptualised as "Algorithmic Standing" — a firm's position in agent evaluation hierarchies, built through structured data quality and operational execution, not emotional association.

For management practice: Invest in Algorithmic Readiness. Rebalance from GEO to AIO. Treat data as product. Prepare for the Great Value Sort. Develop governance proactively. Think platform strategy.

For public policy: Consumer protection law needs updating for delegated purchasing. Antitrust frameworks must address algorithmic collusion. Platform power requires new scrutiny. Algorithmic fairness, liability frameworks, and transparency requirements are urgent.

For business education: 96 per cent of AI-related business school curricula address how leaders can use AI as a tool. Less than 4 per cent address what happens when AI becomes the customer. This gap is a curriculum crisis.

Key Findings at a Glance

Finding	Detail
4.71× Operational Dominance	Balanced agents place 4.71× more decision weight on real-time operational data than on accumulated reputation
Three Boundary Conditions	Delegation + Intermediation + Optimisation = Agentic Commerce
Five Delegation Levels	Human-Only → AI-Informed → AI-Guided → AI-Managed → AI-Autonomous
Three-Layer Decision Stack	Hard Constraints → Learned Heuristics → Real-Time Optimisation
13 Falsifiable Propositions	Testable predictions across micro, meso, and macro levels
Three Trust Phases	Utility-Based Adoption → Monetisation-Driven Misalignment → Discovery-Triggered Disillusionment
Five Readiness Dimensions	Data Infrastructure, Process Adaptation, Capability Development, Governance Frameworks, Strategic Alignment

About the Author

Paul F. Accornero is the Founder of The AI Praxis™ and the author of The Algorithmic Shopper (St. Martin's Press / Macmillan, forthcoming 2027). He is a Guest Lecturer at Harvard University, a Fellow of the Chartered Institute of Marketing (FCIM) and the Chartered Management Institute (FCMI), and holds a PhD candidacy at the University of Queensland. His prior career includes 25+ years of C-suite commercial leadership, including oversight of a €3.5 billion global P&L across 120+ markets.

SSRN Author Rank: Top 3% globally | 22 papers | ~8,000 views | 1,400+ downloads

READ THE FULL BIO | VIEW ALL RESEARCH

Read the Paper

"AGENTIC COMMERCE: A Theory of Markets When the Shopper Is No Longer Human" — SSRN #6111766

Cite This Paper

Accornero, P.F. (2026). "AGENTIC COMMERCE: A Theory of Markets When the Shopper Is No Longer Human." SSRN Electronic Journal. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6111766