When the Machine Becomes the Portfolio Manager

The article “When the Machine Becomes the Portfolio Manager” was originally posted on Alpha Architect blog.

For decades, discretionary portfolio managers (PMs) have been prized for judgment, pattern recognition, and intuition — skills honed through experience and resistant to automation. But the rapid rise of artificial intelligence (AI) and large language models (LLMs) is challenging that assumption. Today, machines are not only processing data but also interpreting narratives, forecasting returns, and constructing investment theses once reserved for humans. This paper examines how AI is reshaping the role of the discretionary PM, arguing that the edge isn’t disappearing — it’s migrating.

The Disappearing Edge: AI, Machine Learning, and the Future of the Discretionary Portfolio Manager

• Fabozzi, Chin, Yelnik, & Liew
• Financial Analyst Journal, 2025
• A version of this paper can be found here
• Want to read our summaries of academic finance papers? Check out our Academic Research Insight category

Key Academic Insights

AI is moving from analysis to decision-making. Large language models (LLMs) and deep learning systems now perform tasks once considered uniquely human such as interpreting earnings calls, constructing macro narratives, and identifying latent market signals.

The line between discretionary and systematic is dissolving. Every investment process now blends human intuition and machine structure. The key differentiator is not whether discretion exists, but how intelligently it is combined with scalable, model-driven insights.

The “hybrid PM” model is real, but fragile. Successful integration of AI requires shared accountability, redesign of workflows, and PMs fluent in model literacy, not just market intuition. Most firms today still treat AI tools as optional inputs, limiting their impact.

Discretionary edge is being redefined. Human skill now lies less in spotting signals and more in interpreting, governing, and constraining model outputs like ensuring alignment with investor mandates, risk budgets, and ethical standards.

Practical Applications for Investment Advisors

Redefine what “discretion” means.
The traditional edge of the discretionary manager—judgment, intuition, and experience—is not disappearing but moving higher in the decision hierarchy. The modern advisor’s edge comes from interpreting and governing model outputs, not competing with them. Knowing when to trust, challenge, or override AI signals becomes the new discretionary skill.

Focus on model literacy, not model worship.
Advisors do not need to become coders, but they must understand how models work, what data they rely on, and where their blind spots are. This literacy allows advisors to explain results, identify risks of overfitting, and maintain accountability when models misfire.

Reframe the advisor’s role as curator, not predictor.
As AI systems take over much of the pattern recognition once done by humans, the advisor’s comparative advantage lies in curating insights—selecting which model signals align with client objectives, regulatory constraints, and investment philosophy. The role shifts from forecaster to interpreter.

How to Explain This to Clients

“AI isn’t replacing human portfolio managers — it’s changing what they do. Instead of relying solely on intuition, today’s PMs act as interpreters and stewards of machine insights. Their value lies in knowing when to trust the model, when to challenge it, and how to turn complex analytics into clear, responsible investment decisions.”

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged and do not reflect management or trading fees, and one cannot invest directly in an index.

Abstract

The discretionary portfolio manager’s role is evolving as artificial intelligence and machine learning increasingly supplement or replace traditional investment insight. This article explores how advances in large language models and deep learning are narrowing the discretionary edge once defined by judgment and narrative skill. A new model is emerging in which the portfolio manager acts as an allocator and model steward, rather than a sole decision-maker. We examine the implications for governance, performance, and risk and argue that firms that retool talent, workflows, and oversight may be best positioned to harness the promise—and manage the limits—of AI-driven asset management.