DPAM: AI reshapes asset managers’ operating models

4 maart 2026

AI in asset management is often seen as a technology issue, but its real impact lies in reshaping how organisations learn, adapt and scale decisions, not merely in automating tasks.

By Frederiek Van Holle, Head of Technology & Innovation, Degroof Petercam Asset Management

AI confronts asset managers primarily with an operating model challenge. As the cost of applying intelligence continues to fall, competitive advantage increasingly depends on an organisation’s ability to embed learning loops into its core processes. Firms that treat AI as a tooling exercise may achieve efficiency gains. Those that redesign their operating models for AI-driven learning gain adaptability.

Understanding AI through this lens shifts the conversation from experimentation to strategy and from short-term optimisation to long-term organisational resilience.

Artificial intelligence is often discussed in asset management as a tooling challenge. Which models to use, which vendors to select, which processes to automate. This framing is understandable, but ultimately misleading. It reduces AI to a technological choice, while its real impact lies elsewhere.

AI is not primarily a technology problem. It is a question of how organisations are designed to create value, learn from their actions and scale decisions. In other words, AI confronts asset managers with an operating model problem, whether they actively deploy AI today or not.

This distinction matters. Because once AI is understood as an operating model issue rather than a toolset, the conversation shifts from experimentation to strategy, from efficiency to adaptability.

Why this AI wave is fundamentally different

Scepticism towards AI is not irrational. Asset management has experienced multiple waves of technological innovation: quantitative models, electronic trading, portfolio management systems, advanced analytics. Many failed to live up to their transformative aspirations and instead merely delivered incremental change.

What differentiates the current AI wave is not intelligence itself, but the economics of intelligence. Three forces have converged: exponential growth in data generation, massive increases in computing power, and advances in algorithmic architectures. Together, they have drastically reduced the cost of applying learning to complex tasks. In other words, the cost of cognition and problem solving has come down.

Generative AI made this shift visible. Activities traditionally associated with high-skilled knowledge work (for instance summarisation, pattern recognition, drafting, analysis) can now be supported or augmented at marginal cost. This does not necessarily eliminate expertise, but it does change roles.

For asset management, this implies a subtle but profound shift: competitive advantage increasingly depends on how effectively an organisation learns from its own activity at scale, rather than on the quality of individual decisions alone.

AI enables asset managers to tighten the feedback loop between research, decisionmaking and outcomes.

From technology adoption to operating logic

The most consequential mistake organisations make when approaching AI is treating it as an IT initiative. AI does not sit alongside existing processes. It reshapes how decisions are made, evaluated and improved within those processes.

In AI-enabled organisations, data is not a byproduct of operations but a core operational asset. Algorithms improve through use. Decisions feed learning loops. Over time, the organisation itself evolves continuously rather than episodically.

This requires an operating logic that differs fundamentally from traditional asset management structures. Siloed teams, fragmented data ownership and sequential decisionmaking limit the ability to learn at scale. Optimisation for control and risk containment, historically strengths of the industry, can inadvertently slow adaptation.

The result is rarely immediate failure. Instead, it is gradual divergence. Firms may look comparable externally, while their underlying learning capabilities drift further apart.

What AI actually enables in asset management

Discussions around AI in asset management often focus on prediction, automation or performance enhancement. These applications exist, but they are secondary to a more fundamental capability: accelerated organisational learning.

AI enables asset managers to tighten the feedback loop between research, decisionmaking and outcomes. Portfolio decisions generate data. Data improves models. Models inform subsequent decisions. This cycle can operate continuously rather than periodically.

The strategic value lies less in superior forecasts than in improved adaptability. Research becomes cumulative rather than fragmented. Operational processes generate insight rather than friction. Client interactions inform service design in near real time.

Importantly, this does not remove the need for human judgement. It changes how human attention is allocated. It signals a move away from repetitive analysis towards interpretation, oversight and strategic choice.

Why ‘doing nothing’ is not a neutral option

A persistent misconception is that firms can wait until AI becomes clearer or more mature. In reality, AI reshapes competitive dynamics asymmetrically.

Organisations that adopt AI-enabled operating models learn faster, experiment more cheaply and scale insight without proportional increases in cost.

Organisations that adopt  AI-enabled operating models learn faster, experiment more cheaply and scale insight without proportional increases in cost. Over time, this creates divergence in responsiveness and adaptability.

In asset management, regulation, long investment horizons and trust-based relationships can mask these dynamics for years. But when adaptation eventually becomes necessary, the gap may be difficult to close.

Choosing not to act is therefore not neutral. It reflects an implicit bet that existing operating models will remain sufficient in an environment where learning dynamics are fundamentally changing.

Scale, learning and responsibility

AI scales impact. Decisions embedded in systems propagate faster and further than human-centric processes allow. Errors, biases and blind spots embedded in systems therefore propagate more widely than in humancentric processes.

Bias rarely stems from malicious intent. It emerges from historical data, incomplete representation and optimisation choices. When encoded in learning systems, such biases become systemic rather than incidental.

For asset managers whose decisions influence capital allocation and long-term outcomes, this raises fundamental ethical considerations. Fairness, transparency and inclusion cannot be addressed retrospectively. They must be integrated into system design and governance from the outset.

Here, incumbents possess an often-underestimated advantage. Strong governance traditions and regulatory discipline provide a foundation for responsible scale. In a digital context, trust determines whether innovation is sustainable.

Conclusion

Artificial intelligence challenges the industry because it changes how organisations learn, adapt and scale decisions.

The core issue is not whether asset managers deploy AI tools, but  whether their operating models can absorb AIdriven learning without fragmenting or losing control. Firms that treat AI as a tooling problem may gain efficiency. Firms that recognise it as an operating model problem gain adaptability.

The strategic question, therefore, is not if AI will matter for asset management, but how organisations redesign themselves to engage with it responsibly and competitively.

Read the original article in Financial Investigator magazine