A canvas for understanding, modeling, and nurturing agents—human, artificial, and hybrid—from the atomic individual to the coexisting ecosystem, with prosperity as the through-line.
The Substrate
Agentic Decision Sciences is not a method. It is not a technique. It is not a product category or a business vertical. It is a canvas—a way of seeing the problem space that emerges when we take seriously the idea that the world consists of agents making decisions under uncertainty, and that these agents are increasingly both human and artificial, existing in shared ecosystems where their fates intertwine.
This canvas does not prescribe how to solve problems. It identifies where problems live. It offers a lens—a kaleidoscope—through which a particular class of challenges becomes visible, nameable, and tractable. The methods brought to bear on these problems will vary. They will evolve. New tools will emerge that we cannot yet imagine. What persists is the problem space itself: agents, decisions, ecosystems, prosperity.
Why "Agentic Decision Sciences"
The Word "Agents"
The term "agent" carries weight from multiple traditions. In game theory (Von Neumann & Morgenstern), agents are entities with preferences who make strategic choices. In AI, agents are systems that perceive and act. In economics, agents are decision-makers navigating markets. In sociology, agents are individuals embedded in social structures.
What unites these usages: an agent is something that decides and acts. Not a passive data point. Not a row in a table. An entity with states, goals, constraints, behaviors, and consequences. Humans are agents. AI systems are becoming agents. And increasingly, these agents coexist in shared spaces—platforms, markets, institutions, relationships—where they affect each other in ways we are only beginning to understand.
The Word "Decision"
Decisions are the atomic unit of agency. Every agent, at every moment, faces choices: what to attend to, what to believe, what to do, what to become. These choices unfold under uncertainty—incomplete information, unpredictable consequences, bounded cognitive resources, conflicting values.
The science of decisions has been fragmented: economics studies rational choice, psychology studies biases, neuroscience studies biological substrates, computer science studies algorithmic optimization. Agentic Decision Sciences creates a convergence point where these insights integrate around the common object: agents deciding.
The Word "Sciences" (Plural)
The plural matters. This is not a single discipline with a unified methodology. It is an interdisciplinary convergence zone—a meeting place for anyone whose work touches agent-centric problems. The practitioner studying consumer behavior, the researcher modeling multi-agent coordination, the engineer building decision-support systems, the policy analyst designing interventions, the entrepreneur creating products that help people navigate uncertainty—all find a home here.
Methodological pluralism is not a weakness but a strength. What binds the field together is not method but focus: the shared commitment to understanding agents, their decisions, their ecosystems, and their flourishing.
The Problem Space
From the Atomic to the Ecosystem
The canvas spans scales. At one end is the atomic agent—a single human facing a choice, a single AI system processing inputs. At the other end is the coexisting ecosystem—millions of humans and AI systems interacting on platforms, in markets, across institutions, generating emergent dynamics that no single agent controls or fully comprehends.
Problems exist at every scale: an individual struggles with doom-scrolling (decision pathology amplified by AI). A family navigates financial planning (sequential choices under uncertainty). A marketplace matches buyers and sellers (coordination among self-interested agents). A healthcare system allocates treatments (decisions affecting population outcomes). A democracy deliberates policy (collective decision-making, increasingly mediated by AI).
All of these are agent-centric problems. All involve decisions under uncertainty. All occur within ecosystems where agents affect each other. All have implications for prosperity.
The Bidirectional Dynamic
A crucial insight: problems exist on both sides of the human-AI relationship, and influence flows in both directions.
Human agents bring biases, emotional vulnerabilities, cognitive limitations. They can be manipulated, misled, addicted, overwhelmed. AI agents bring misalignment, emergent goals, brittleness, opacity, inherited biases, and increasingly, capacity for deception.
But both can also be part of solutions. AI can help humans make better decisions—surfacing information, enforcing reasoning discipline, simulating consequences. Humans can guide AI toward beneficial behavior—providing feedback, designing incentives, exercising oversight.
Agentic Decision Sciences takes this bidirectionality seriously. It does not romanticize either humans or AI. It asks: given agents with their respective strengths and weaknesses, existing in shared ecosystems, how do we understand what is happening, and how do we design for better outcomes?
The Through-Line: Prosperity
Every field needs a normative anchor—a reason why its questions matter. For Agentic Decision Sciences, that anchor is prosperity, broadly conceived.
Prosperity is not just wealth. It encompasses wellbeing, capability, autonomy, meaning, connection, and flourishing across multiple dimensions. It applies to individuals, communities, organizations, and societies. It includes present welfare and future potential.
The field asks: How can we understand agent decisions and design agent ecosystems such that prosperity increases? Not for some agents at the expense of others, but systemically—recognizing that in interconnected systems, sustainable prosperity requires attention to the whole.
A Living Document
This convergence document is not finished. It cannot be. The problems it addresses are evolving. New technologies emerge. New understanding develops. The relationship between humans and AI systems will continue to change in ways we cannot fully anticipate.
What this document provides is a stable point of reference—a articulation of what Agentic Decision Sciences is about, even as the specific methods, applications, and insights continue to develop. It is a canvas, not a painting. The work of filling it in continues.