The next chapter for asset managers: generating alpha with AI and ML

Key Takeaways

• The benefits of generative AI are seen as outweighing the costs.
• The technology is expected to transform the industry with enhanced decision-making and returns.
• Operational hurdles are high, and infrastructure will need to be modernized.
• Skill shortages will also need to be addressed for the technology to be successful.

Although a hot topic for some time, it may be years before artificial intelligence (AI) in asset management reaches its full potential. While Natural language processing (NLP) and machine learning (ML) are still evolving, generative AI (GenAI) is creating most of the buzz today.

GenAI and the future of alpha

Unlike traditional AI models that rely solely on historical data, GenAI in finance can create fresh perspectives by synthesizing diverse data sources. New opportunities are uncovered, risks are better managed, and client engagement is personalized at scale. This sharpens the decision-making process and the competitive edge.

It is no surprise that, despite the high implementation costs, investment in AI and ML is top of the investment priority list for around 75 percent of CEOs in KPMG’s 2024 Asset Management CEO Outlook report.

A separate study from Oliver Wyman and Morgan Stanley – The Generative AI Tipping Point – confirms that the benefits outweigh the costs.

How ML and AI enhance alpha generation

On the performance front, the report highlights the ability of AI algorithms to process data in real time, which can help managers navigate different market conditions. This advantage has become increasingly important in the current environment, where ongoing geopolitical tensions and unpredictable tariff policies are causing ructions across the asset class spectrum.

The technology also adds value by incorporating alternative data sources such as social media, satellite data, and digital consumption behavior into financial analysis. Patterns and trends are identified along with risk tolerance, return expectations, investment objectives, and dynamic market conditions.

Moreover, the tools have accelerated the pace of automation – so, tasks like generating financial reports, risk management, and document handling have become faster and more precise. This reduces human error, giving asset managers more time to focus on AI in alpha-generating ideas and high-impact strategic investment strategies.

The technology also offers advanced reasoning tools such as the ‘Chain of Thought’, which analyses complex problems using logical thought processes. This enhances the understanding of asset interrelationships and produces more accurate projections.

Looking farther down the line, as with all technologies, it is a journey. Overall, AI will continue to evolve, becoming increasingly integral to investment strategies. Market participants believe that one of the most significant developments will be the refinement of predictive analytics, using ML, which allows asset managers to anticipate market movements with greater accuracy.

AI and ML will also play a greater role in fundamental analysis. Instead of merely optimizing existing strategies, these tools have the potential to develop entirely new investment paradigms and identify asset correlations that human analysts would never detect. This could lead to the emergence of novel asset classes and trading strategies with opportunities for ML in alpha generation. Indeed, ML is already actively contributing to alpha generation in modern investment strategies, opening the door to further prospects.

Strategic implications for asset managers

There will be challenges, and the operational barriers remain high. Adam Graham, Global Head of Product at FE fundinfo, contends that many fund managers do not have the right infrastructure to streamline data, integrate systems, and remove or shift outdated tools that cannot scale to meet modern expectations.

Many are still operating on a silo basis, with disconnected systems managing everything from fund documentation to performance analytics. For example, it’s not unusual for product information to be maintained in one format, compliance data in another, and reporting tasks handled manually across Excel files.

To get the best results, Graham advises adopting a two-prong approach. This means applying AI to specific, high-impact use cases today, while continuing to modernize the underlying infrastructure. Technology needs to connect fragmented systems, embed data governance, and simplify workflows. Clean, structured data and integrated processes remain essential, but they are no longer prerequisites; they are the enablers.

It is not just about the nuts and bolts. The requisite expertise is also needed. Currently, there is a shortage of financial and technical talent. Many finance professionals lack AI skills, while data scientists may not grasp the nuances of financial markets.

Analyst researchers and investment managers are seen as being the most exposed. The former must be able to evaluate AI-driven approaches critically. Those that cannot risk either overlooking superior strategies or, worse, endorsing flawed ones. Meanwhile, the latter face growing pressure to assure clients they are harnessing AI and not just paying lip service or misapplying it.

Conclusion

Asset managers are also advised to invest in specialized training to ensure the ethical use of data. Several advanced models function as complex opaque boxes, making it difficult for risk managers to interpret how decisions are made. Such opacity raises concerns around model validation, transparency, and fairness, especially as regulatory bodies worldwide are scrutinising the use of AI in trading.

While firms will adopt different implementation strategies, industry participants all agree that if they fail to address these skill gaps, they will struggle to use AI effectively and fall behind their peers.