The landscape of investment management is undergoing a significant transformation with the integration of artificial intelligence (AI). As firms increasingly incorporate AI into their operations, a critical challenge arises: defining the type of intelligence being utilized. This article explores the concept of agentic AI and the need for a comprehensive AI taxonomy to enhance governance and scalability in investment strategies.
AI agents, distinct from traditional models like ChatGPT, can observe, analyze, make decisions, and sometimes act autonomously within defined parameters.
As investment firms adopt these technologies, they must clarify whether their AI functions merely as a decision-support tool, an independent research analyst, or a delegated trader. This classification is essential for effective management and oversight of the technology.
Table of Contents:
The importance of AI classification
Implementing AI in investment management presents a prime opportunity to delineate boundaries and establish governance frameworks. Without a shared language or classification system, firms risk mismanaging technology, leading to both over-reliance and under-utilization. Our research team, in collaboration with DePaul University and Panthera Solutions, has developed a multi-dimensional classification system for AI agents in the investment sector to provide a common framework for evaluating these systems.
This classification system assists practitioners, board members, and regulators in understanding the operational capabilities of AI agents based on their autonomy, function, learning ability, and governance structures. By utilizing this framework, investment leaders can adopt a structured approach to AI integration, transitioning from reactive to proactive strategies.
Strategic vs. tactical AI adoption
Investment managers typically adopt AI in one of two ways: as a functional toolkit or as an integral part of the investment decision-making process. The functional approach employs AI for specific tasks, such as risk assessment, sentiment analysis, and summarizing portfolio positions. While this enhances operational efficiency, it does not fundamentally alter the human-centric decision-making architecture of the organization.
Conversely, a smaller group of firms embraces the systemic approach, where AI agents are integrated into the investment design process as adaptive participants rather than merely supportive tools. This approach requires clear definitions of autonomy, learning capacity, and governance, resulting in a decision ecosystem where human insight and machine intelligence collaboratively evolve.
Designing an effective AI framework
The distinction between functional and systemic AI adoption is crucial. While the former leads to faster tools, the latter cultivates smarter organizations capable of sustaining a competitive edge. Neuroscientist Antonio Damasio noted that all forms of intelligence seek to achieve homeostasis, a balance with their environment. Similarly, financial markets are complex adaptive systems that rely on equilibrium among data interpretation, automation, accountability, and sustainability.
A robust AI framework should map AI agents across three vital dimensions: the investment process, comparative advantage, and complexity range. Understanding where in the value chain an AI agent operates is essential for accountability and governance. The investment process typically includes five key stages: idea generation, evaluation, decision-making, execution, and monitoring. AI agents can enhance any of these stages, but it is critical that decision-making rights align with the interpretability of the AI’s functions.
Enhancing decision quality
Next, consider the comparative advantage that AI brings to investment firms. It is vital to identify whether AI enhances informational, analytical, or behavioral edges. Importantly, while AI does not inherently create alpha, it can amplify existing advantages. For instance, a quantitative firm might deploy reinforcement learning models for deeper analytical insights, while a discretionary firm might utilize AI co-pilots to maintain reasoning quality and behavioral discipline.
Finally, assess the complexity range within which the AI operates. Investment environments fluctuate between measurable risks and radical uncertainties. Understanding this spectrum enables firms to navigate market dynamics effectively and ensure that AI contributes meaningfully to decision-making processes.
AI agents, distinct from traditional models like ChatGPT, can observe, analyze, make decisions, and sometimes act autonomously within defined parameters. As investment firms adopt these technologies, they must clarify whether their AI functions merely as a decision-support tool, an independent research analyst, or a delegated trader. This classification is essential for effective management and oversight of the technology.0
AI agents, distinct from traditional models like ChatGPT, can observe, analyze, make decisions, and sometimes act autonomously within defined parameters. As investment firms adopt these technologies, they must clarify whether their AI functions merely as a decision-support tool, an independent research analyst, or a delegated trader. This classification is essential for effective management and oversight of the technology.1
