Where are you seeing AI tools add the most value across the investment process today? 
AI is becoming an increasingly valuable tool across the investment process, particularly for smaller teams where professionals are involved across multiple stages of the investment value chain. Within fixed income investing, some of the most promising applications are in credit research and portfolio construction.
In my experience, large language models can help process large volumes of information more efficiently by summarising earnings results, comparing issuers with peers, identifying changes in credit fundamentals and exploring different scenarios. This has the potential to reduce the time spent gathering and organising information, allowing investment professionals to focus more on analysis, judgement and client advice.
AI also has potential applications in portfolio construction. By incorporating mandate guidelines and investment constraints, investment teams can evaluate a broader range of portfolio outcomes, test scenarios and identifyopportunities to improve diversification or risk-adjusted returns within defined parameters. However, these applications remain at an early stage and require further testing and refinement to ensure more consistent and reliable outcomes.
However, the quality of both inputs and outputs remains critical, and I see AI as a tool that can enhance productivity and support investment recommendations while risk assessment and accountability remain with investment professionals.
How significant of a challenge is data quality when it comes to effectively implementing AI on a trading desk?
Data quality is one of the biggest challenges when applying AI within investment processes. In fixed income markets, investors benefit from access to large amounts of data through providers such as Bloomberg, covering pricing, spreads and market activity. However, fixed income remains predominantly an OTC market, making data validation particularly important.
Ensuring data quality and validating information has always been a core responsibility for fixed income investors. AI can process large datasets quickly and identify patterns that might otherwise be missed, but it remainshighly dependent on the quality of the underlying information. Inaccurate, incomplete or outdated data can quickly lead to misleading conclusions.
In credit research, AI is particularly effective at analysing company disclosures, earnings reports and news flow. However, it can occasionally generate plausible conclusions that are not fully supported by the underlying information. As a result, validation and human oversight remain essential.
As AI adoption increases, investment firms will need to focus as much on data quality and governance as they do on the models themselves.
As the use of AI increases, how are firms approaching governance and oversight around these tools?
Governance and oversight have become central considerations in firms’ adoption of AI, particularly given the risks associated with data security, confidentiality and model reliability.
At Nippon Life Global Investors Europe, the focus has been on approved enterprise solutions such as Microsoft Copilot, which operate within established security, compliance and governance frameworks. We are also evaluating other applications, including the beta version of Bloomberg’s AI capabilities. In contrast, the use of public AI tools is subject to greater scrutiny due to concerns around confidential information, data protection and regulatory obligations. As a regulated financial institution, it is also important for us to closely monitor the evolving regulatory landscape surrounding AI.
As the technology develops, I expect firms to remain cautious about how AI is used within investment processes.
Looking ahead, how do you expect the role of AI within the investment process to evolve over the next few years?
Looking ahead, I expect AI to continue improving both the speed and quality of investment analysis. The objective is not simply to conduct more research, but to free up time for deeper analysis, better investment ideas and more informed decision-making. The greatest opportunity may be improving the quality and consistency of investment research rather than pure automation.
In portfolio construction and risk management, AI has the potential to help investment teams assess a much broader range of scenarios, constraints and potential outcomes than would be practical through traditional methods alone. For portfolio managers advising insurance clients, this could provide valuable insights into how portfolios behave under different market, regulatory and capital scenarios.
Agentic AI is another area with significant potential. Although its adoption remains at a relatively early stage, it could eventually automate many routine and time-consuming tasks associated with research, data gathering and monitoring. This would allow investment professionals to devote more time to investment analysis, client engagement and strategic thinking.
While the long-term impact of AI remains uncertain, I expect its role within the investment process to continue expanding as the technology matures and firms gain greater confidence in its practical applications.