Quant strategies have always been driven by data and technology – and AI is the latest chapter in that evolution. It’s important to note that AI in investing isn’t completely new. ML and NLP techniques have been researched and applied in our strategies for decades.

What’s different now is the magnitude of what these AI techniques can do, and the variety of data they can handle. Rather than replacing the tried-and-true quant methods, AI is enhancing them. Think of it as moving from a toolbox of hand tools to power tools – the job is the same, but we aim to do it faster and more effectively.

Machine learning

For example, machine learning (ML) can uncover complex patterns in data – including non-linear relationships and subtle interactions that traditional models might miss. This allows us to refine existing signals or discover new sources of alpha. For instance, by combining a classic reversal signal with news sentiment data, our quant team improved its ability to identify stocks likely to rebound after an overreaction. ML can also group similar securities based on shared traits like supply chain or technology use, adding depth to risk management. These kinds of insights help make quant models more adaptive to today’s fast-moving markets.

Natural language processing (NLP)

AI’s ability to interpret language has advanced rapidly - moving from simple word counts to large language models that can understand nuance, tone, and intent. By analyzing textual data such as news, financial reports, or employee reviews, and even listening to the tone of voice on earnings calls, natural language processing (NLP) techniques help reveal deeper insights into sentiment. Whether it’s gauging a CEO’s confidence, detecting subtle shifts in customer mood, or identifying key players in a company’s supply chain, NLP enriches our understanding of the broader corporate landscape. These richer signals feed into our quantitative models, potentially giving an edge in gauging market sentiment and corporate health.

Alternative data

The realm of alternative data – unconventional or non-traditional types of data that haven’t been used in the past for investment decisions – makes these AI techniques possible and increasingly essential. Unparalleled in volume, variety and speed of generation, alternative data offers real-time insights into economic activity that traditional data often misses. For example, credit card data can signal shifts in consumer spending, while shipping manifests can track global trade flows.

Separating noise from signal and converting the torrent of alternative data into tangible investment insights means combining innovative AI techniques with sensible economic hypotheses. Incorporating AI into investment processes offers both quant and fundamental investors the advantages of faster speed, broader sources of information, and better information processing. But it also necessitates being purposeful, clear-headed and transparent when using AI in our processes.

With so much new data and computing power at our fingertips, quant investors today can tap into data sources and techniques that were unimaginable a decade ago. Robeco’s quant team, for example, has expanded its toolkit to include everything from sector-specific signals to novel alternative datasets.

To ensure quant strategies evolve in a healthy, evidence-based way, preserving the robustness clients expect, we’ve developed dedicated AI-driven strategies in our incubator. This has led, for example, to the launch of our Dynamic Theme Machine (DTM) ETF.

Dynamic Theme Machine

Thematic investing holds great appeal, but it also comes with challenges. Investors often gain exposure to a theme only once it’s already well underway, missing out on early upside. Emerging themes may be too small to justify a standalone product. And once invested, it can be difficult to know when a theme is fading, leading to missed opportunities elsewhere.

Using AI, Robeco’s Dynamic Theme Machine (DTM) ETF addresses these challenges through a multi-thematic quantitative approach. Our AI engine, using techniques such as NLP, scans vast and diverse data sources to detect patterns, identify thematic momentum, and signal when a theme is gaining or losing traction. This allows us to dynamically allocate across emerging and established themes, using a systematic process helps us rotate efficiently. Our established stock selection and portfolio construction algorithms ensure rigorous risk management throughout.

AI has also emerged as a dominant theme for fundamental investors. With its seemingly infinite capacity to take on diverse tasks from computer coding to drug development, AI has been heralded as the next era in technology, akin to the personal computer, the internet, and the smartphone. While the impacts of such technological shifts are broadly felt, there are typically fewer beneficiaries.

Therefore, we focus on identifying providers of foundational technologies and those adopting the technology within the core of their business to create value.

AI foundations

Understanding competitive differentiation and the real value-add of companies supplying technology in the development of AI is critical to building long-term returns. In previous technology eras, a handful of firms providing foundational intellectual capital captured the majority of the value creation across the cycle. For instance, this was the case for IBM in the mainframe era, Microsoft and Intel in the PC era, Apple and Qualcomm in the mobile era, and Amazon and Google in the internet era. While many other technologies played a role in those cycles, often these enablers and component suppliers proved interchangeable. We suspect a similar dynamic will play out in the AI era.

AI adopters

In addition, AI adopters are companies in multiples sectors that integrate AI to improve their products or operations. Virtually every company will use AI in some form – much like every company uses computers or the internet today – but the impact on their financial performance will differ. For instance, the internet enabled a new class of competition to emerge from e-commerce to streaming media, a development that transformed their respective industries. Similarly, cloud computing enabled a new business model for software as Software as a Service (SaaS) providers not only distributed incumbent providers, but also broadened the market for enterprise applications. AI has already created opportunities for existing SaaS providers to enhance their offerings as well as for new entrants to rethink the model yet again.

Digital Innovations strategy

These opportunities require careful analysis – looking at the substance behind the AI claims. Does the company have proprietary data? AI talent? A defensible strategy to monetize their AI capabilities? By asking these questions, we aim to separate the hype from the reality and potential. Our Digital Innovations strategy aims to include exposure to established players and promising emerging names, when appropriate, always weighed against the risks.