Artificial Intelligence continues to gain momentum in investment management, but true transformation remains uneven.
Artificial Intelligence continues to gain momentum in investment management, but true transformation remains uneven. In our previous article (link here), we explored how AI is becoming a strategic enabler for asset and wealth managers — transforming research, portfolio construction, advisory productivity, and client engagement. Yet the industry is still far from capturing AI’s full potential.
This second chapter of the MDOTM × EY research focuses on one of the key reasons why: the Pilot Project Paradox — the recurring pattern where firms start many AI experiments, yet only a few progress to meaningful, enterprise-wide adoption.
Why the Industry Is Stuck in Pilots
Over the last eighteen months, investment managers have launched an impressive number of AI initiatives. But most remain small, isolated, and limited to narrow use cases — typically chatbots, Q&A tools, or simple automation tasks.
This initial experimentation helps organizations build familiarity with AI. But it also highlights a structural problem: pilots do not scale by themselves.
They often lack strategic alignment, governance, the right data foundation, and the change-management support needed to move from “testing” to “transforming.”
This disconnect is what the research identifies as the Pilot Project Paradox:
It is relatively easy to build an AI proof of concept — and surprisingly difficult to turn it into a production system that creates real business value.
The Main Barriers to Scaling AI
The stagnation after initial pilots is not primarily technological.
The research shows that the most common barriers fall across three areas — business, technology, and risk.
1. Business and Organizational Barriers
Many firms still approach AI as a set of disconnected experiments rather than as part of the broader strategic plan. This creates several issues:
- Lack of strategic integration: AI is often tied to efficiency metrics, not to growth or top-line ambition.
- Fragmented sponsorship: pilots are typically championed by CIOs or COOs, while commercial, investment, and distribution teams remain less involved.
- Isolated use cases: firms focus on “small” productivity pilots rather than on value-driving use cases in portfolio management, advisory, or client experience.
- Application on outdated processes: applying AI to inefficient workflows yields only incremental value unless processes are redesigned for automation.
- Insufficient skills: very few employees receive AI training, limiting adoption and long-term scalability.
These challenges reflect a common theme: firms are experimenting with AI, but without a holistic plan for where it can truly create differentiation.
2. Technological Barriers
Scaling AI requires more than model development. It demands the right data, infrastructure, and architecture.
The research highlights three critical issues:
- Data quality and readiness: raw data is often fragmented, inconsistent, or hard to access — especially unstructured data.
- Legacy systems: many infrastructures are outdated, making integration and scalability difficult.
- Proliferation of uncoordinated stacks: each pilot often introduces its own tools, models, and workflows, slowing down productionization and complicating risk management.
To unlock real value, firms need to invest in reusable model archetypes, scalable data platforms, and API-driven architectures — allowing new use cases to be deployed faster and more consistently.
3. Risk and Regulatory Barriers
The evolving regulatory landscape adds complexity to scaling AI.
Firms face challenges such as:
- Defining AI-specific risk metrics and integrating them into the enterprise risk framework.
- Establishing a consistent risk appetite across different use cases.
- Managing compliance, data privacy, intellectual property, and model transparency.
- Navigating evolving regulatory frameworks, including the upcoming AI Act.
Without clear governance and controls, firms tend to limit AI applications to non-critical activities — reinforcing the Pilot Project Paradox.
Why Firms Are Now Beginning to Break the Paradox
Despite these barriers, the research highlights a growing shift: more investment managers are now ready to move from pilots to scalable programs.
Several factors are driving this transition:
- Increased familiarity with AI after initial experimentation.
- Stronger pressure to innovate as peers accelerate adoption.
- More mature technological foundations (cloud, data platforms, orchestration layers).
- Improved understanding of where AI creates real business value — not only in efficiency, but in portfolio management, advisory scalability, product differentiation, and client experience.
Firms that succeed in breaking the paradox share a common feature:
They treat AI not as a series of experiments, but as a structural capability embedded into their strategic roadmap.
A Turning Point for Investment Management
The industry is moving from curiosity to conviction. AI is no longer something to “explore,” but something to operationalize. This requires a clear AI strategy linked to business outcomes, an operating model that supports cross-functional implementation, scalable data and technology foundations, robust governance and risk controls and an organization trained and confident in using AI.
As the next chapter of the MDOTM × EY paper will explore, the real opportunity lies in creating an AI-enabled operating model — one that removes the barriers to scale and helps investment managers unlock the full value of the technology.
