The financial industry is currently experiencing a profound transformation, driven in large part by the rapid evolution and integration of artificial intelligence. While the public consciousness often equates "AI" with the groundbreaking generative capabilities of Large Language Models (LLMs) that surged into prominence in 2022, seasoned professionals have been leveraging AI’s power for much longer. Adam Cilio, Senior Portfolio Manager & Engineer at Guardian Capital, is a prime example of this extended engagement. For nearly a decade, Cilio has been at the forefront of applying AI within investment processes, specifically through Guardian Capital’s innovative i³ investment portfolios. This approach, rooted in a sophisticated understanding of AI’s diverse forms, strategically deploys different artificial intelligence models to tackle the nuanced challenges of portfolio management, recognizing that no single AI solution is universally optimal.
Guardian Capital’s i³ investment philosophy is built upon a foundational principle: the acknowledgment that artificial intelligence is not a monolithic entity. Instead, it encompasses a spectrum of distinct models, each possessing unique strengths and weaknesses. This core understanding allows the i³ team to meticulously select and apply the most appropriate AI tool for specific tasks within their investment framework. Cilio elaborated on this distinction, explaining that the i³ team has actively utilized machine learning (ML) AI tools since late 2017. This "traditional" form of AI, he noted, excels at numerical tasks where LLMs often falter. Simultaneously, the firm has embraced LLMs for their prowess in digesting vast amounts of qualitative data and extracting critical insights, a capability that complements the quantitative strengths of ML. This synergistic integration of diverse AI capabilities, Cilio posits, forms the bedrock of their alpha generation strategy, enabling controlled costs and fostering genuine differentiation in a competitive market.
"What truly sets us apart is not merely the AI technology itself," Cilio emphasized in a recent interview. "Algorithms can be implemented anywhere. The true differentiator lies in the fusion of applying AI to an investment universe analyzed through a multi-model, multi-feature approach, and then synergizing that with the invaluable experience of our portfolio managers. These seasoned professionals are responsible for the crucial aspects of portfolio construction and risk management at every single stage." He further elucidated the specific applications of AI within their process: "We leverage AI to identify companies exhibiting qualities such as robust earnings growth, coupled with a sustainable cash flow, and a consistent history of dividend payouts and growth over time. Our focus extends beyond the mere quantum of the payout; we scrutinize the sustainability and the growth trajectory of those dividends."
AI’s Role in Informing Sophisticated Dividend Strategies
This meticulously crafted AI-driven process is now actively informing two new fund offerings from Guardian Capital: the Guardian i³ Global Dividend Growth Fund and the Guardian i³ Canadian Dividend Growth Fund. Cilio detailed how the AI tools employed in managing these funds are specifically engineered to benchmark stocks against their global peers. For the Canadian fund, this comparative analysis is particularly vital, serving to counteract the inherent tendency of Canadian equity strategies to be dominated by a limited number of large-cap companies.
In both the Canadian and global equity strategies recently launched by Guardian Capital, Cilio explained that their AI tools empower the team to forecast a company’s forward dividend growth trajectory while simultaneously measuring the sustainability of those dividends. This enables them to strategically position portfolios towards companies that may offer a slightly lower initial payout but demonstrate a consistent capacity for dividend growth, thereby avoiding those companies at risk of failing to sustain high dividend growth over the long term. This data-driven approach, Cilio stated, has consistently correlated with superior long-term equity performance and, critically, with the preservation of capital.
The broader integration of AI into Guardian Capital’s investment framework is encapsulated in what Cilio terms their "GPS" investment philosophy. GPS stands for Growth, Payout, and Sustainability of cash flows. Cilio explained that the AI tools are instrumental in enabling his management team to maintain a sharp focus on these three core metrics throughout their analytical process. The outcome of this focused approach, he noted, is the creation of high-conviction portfolios characterized by lower turnover rates and a more concentrated selection of underlying holdings. The efficacy of this strategy, he stressed, is intrinsically linked to the judicious application of the right AI tools for the specific analytical demands.
The Strategic Deployment of AI Tools for Specific Tasks
The foundational AI tool that Cilio and his team have been utilizing since 2017 is not an LLM. Rather, it is a sophisticated, numerate machine learning model algorithm that transcends the limitations of traditional linear statistical models. In the context of these investment products, this specific AI model is employed to meticulously identify and analyze patterns related to dividend growth, dividend sustainability, and earnings growth. Cilio highlighted the advantages of this more "traditional" AI approach, noting its enhanced robustness in handling outliers and spurious data, and its significantly lower susceptibility to the risk of "hallucinations" – a phenomenon that can sometimes derail newer LLMs. Furthermore, this ML model offers superior control over potential biases inherent in input data and possesses the ability to discern complex interrelationships between data points that are often imperceptible to linear models.
However, Cilio was quick to acknowledge the significant role that generative AI tools, particularly LLMs, play within their broader analytical framework. He pointed to their invaluable application in the digestion and analysis of unstructured text data. Cilio illustrated the impact of LLMs by noting that they can effectively replace the work of three human analysts who would typically require days to meticulously review hundreds of documents. LLMs, therefore, dramatically accelerate qualitative analysis, while machine learning AI powers the quantitative aspects of their research. Crucially, Cilio reiterated that the entire process remains anchored by human oversight and expertise.
"AI provides the essential framework, but it is imperative that a human remains in oversight, monitoring every facet of the process," Cilio stated. "The scale of operation must still be such that a human can effectively manage and interpret the outputs. For our industry, this is particularly relevant. One cannot simply rely on an LLM’s output at face value. It necessitates specific software development, robust control mechanisms, and human-driven analysis that provides not only the summary or output of an LLM but also a clear indication of its confidence level. While all of this can be engineered, the ultimate validation and integration must be performed by a person."
Quantifiable Benefits of AI Integration for Investors
Cilio firmly believes that the strategic deployment of these diverse AI tools can yield significant benefits for investment strategies, particularly in terms of cost control and alpha generation. He pointed to the lean and efficient structure of his own team as evidence of this operational advantage, comprising four investment managers, one client portfolio manager, two data scientists, one data engineer, and one software developer – a team capable of high output and efficiency.
Moreover, AI can foster a more conviction-driven investment strategy, Cilio explained, by enabling a rigorous focus on the precise metrics by which a fund’s performance will ultimately be judged, such as the sustainability of dividend growth. This disciplined approach can lead to reduced portfolio turnover. It also serves to prevent portfolio managers from engaging in "closet indexing," a practice where a small number of holdings are used as a proxy for a specific index, thereby diluting genuine active management. The linchpin to achieving positive outcomes, Cilio underscored, is a profound understanding that certain AI tools are demonstrably better suited to particular tasks, and that an experienced human at the helm can harness the power of AI to construct an efficient and effective investment strategy.
"Our clients understand, at the end of the day, that the product is still managed by a seasoned team of portfolio managers," Cilio concluded. "These managers are now empowered to leverage their extensive domain knowledge in conjunction with AI technology to make demonstrably better decisions. I would even go so far as to characterize this as a renaissance in active management. It allows us to implement a superior process, which in turn fosters greater trust and confidence in active managers." This perspective suggests a future where the synergistic combination of human expertise and advanced AI capabilities will redefine investment management, leading to more robust, transparent, and ultimately more rewarding outcomes for investors.
