CFA Institute unveils framework to help advisers navigate AI-driven market upheaval

The CFA Institute Research and Policy Center has rolled out a new framework intended to help investment professionals get ahead of the structural changes artificial intelligence is bringing to capital markets, rather than simply reacting once they've already taken hold.

The tool, dubbed the CFA Institute AI Transition Framework, debuts as the opening piece in a broader research series examining how AI's growing role in investment management could reorganize everything from price formation to fiduciary duty.

The framework centers on four forces already reshaping the industry: capability, or the technical power of AI systems; adoption, meaning how deeply firms embed that technology into their operations; substitution, the shifting of tasks between humans and machines; and recomposition, the market-wide fallout once those shifts accumulate.

“The investment profession has always evolved within the capital market ecosystem, and this transition will be no different,” said Mona Naqvi, the center's managing director and author of the underlying paper. "The objective now is to understand these structural changes early enough so that professional standards, governance, and market practice can evolve ahead of AI's deepening integration, rather than solely in response to it."

Naqvi also pointed to stakes that go well beyond back-office efficiency. "The implications of AI's integration into the market ecosystem extend well beyond productivity gains and reach into price formation, capital allocation, and the integrity and stability of the financial system. As AI-driven analytical capability becomes increasingly abundant and embedded in investment decision-making, professional competence will increasingly depend on judgement, ethics, and the ability to govern complex systems responsibly."

Four scenarios

Rather than forecasting a single outcome, the research lays out four alternative states that capital markets could settle into as AI adoption deepens, each reflecting a different balance of the four structural forces.

In an "augmented markets" scenario, AI would boost workflow efficiency across research, portfolio construction and risk management without fundamentally rewriting market architecture. Human oversight of allocation decisions would remain intact, active management would stay viable, and firms would see productivity gains and gradual fee compression rather than existential disruption.

"Competitive divergence" describes a world where AI adoption is patchy. Firms that push deepest into AI across trading and portfolio construction would pull ahead on cost and scalability, while those tied to legacy systems face margin pressure, talent losses and eroding relative performance. Asset owners would increasingly sort managers into AI haves and have-nots, while smaller intermediaries would be forced to choose between investing, partnering, specializing or exiting.

Under "platform convergence," AI capability would spread widely but through shared infrastructure such as foundation models and centralized data and compute ecosystems. That would compress differentiation among managers and concentrate influence in a handful of dominant platforms, narrowing the informational edge that has historically driven alpha and pushing correlations to shift more rapidly as common signals spread across overlapping systems.

The most far-reaching scenario, "model-mediated markets," envisions AI systems taking primary responsibility for signal generation, capital allocation and risk calibration across large swaths of the market. The line between discretionary and systematic investing would blur, portfolio construction and execution would become largely automated, and governance credibility rather than raw analytical output would become the key differentiator among managers.

Building blocks under pressure

The underlying research paper, titled "Artificial Intelligence and the Future of Finance: A Framework for Structural Change," argues that AI doesn't so much demolish the theoretical pillars of modern finance as force them to operate under different conditions.

Five tenets in particular face strain: the efficient market hypothesis, the diversification logic behind modern portfolio theory, the premise that skill can generate persistent outperformance, fiduciary accountability, and market structure stability.

The paper's central worry is what it calls "cognitive convergence," the tendency of firms to align around similar model architectures, training data and decision frameworks as AI adoption spreads. That alignment, the research warns, could compress the diversity of market views that historically has kept the system resilient, echoing what other researchers have labeled "monoculture risk."

To manage that risk, the paper flags three governance priorities: keeping human oversight documentable and auditable within hybrid human-AI decision systems, building validation frameworks suited to AI models that continue learning and adapting after deployment, and developing oversight mechanisms for shared analytical infrastructure now that concentration risk is migrating from individual firms to the technology providers underpinning them.

What's next

The framework paper is the first release under the center's broader "Reimagining the Future of Finance in the Age of AI" research program, which will go on to examine how the investment profession itself is being reshaped as AI scales, how accountability works in markets increasingly mediated by models, and what institutional safeguards are needed to keep markets fair and stable as the transition continues.