Chapter 2
AI-powered investing and trading
Summary
AI is transforming industries across the board - and fixed income is no exception. Firms are increasingly integrating generative AI into key investment processes.
According to our research, the top use case is the creation of synthetic data to augment training datasets (51%), followed by credit risk assessment and bond valuation (43%).
Other prominent applications include automating client reporting and communication, market sentiment analysis, and automated trade execution.
Adoption of AI varies across functions. Technology and infrastructure teams are leading deployment efforts (52%), followed by risk management and compliance (38%), and trading and execution desks (37%).
This trend indicates a clear prioritization of AI in both core operational areas and functions tied to risk and trading.
Despite strong momentum, heads of trading expressed notable concerns about AI use in fixed-income markets.
Key issues include the risk of underestimating liquidity challenges, inflexibility in adapting to regulatory changes, and misclassification of structured products.
Additionally, concerns around bias from reliance on historical data and the opacity of 'black box' models highlight the ongoing need for human oversight, transparency, and expert judgment in AI adoption.
Are you currently using Generative AI solutions in your workflow operations, and if so, in what areas is it currently bringing the most value? (Respondents were asked to select three options)
Creating synthetic data to augment training datasets
Credit risk assessment and bond valuation
Automating client reporting and communication
Market sentiment analysis and news interpretation
Automated trade execution and order routing
Analyzing large unstructured data sets, like earning call transcripts
Portfolio optimization and scenario analysis
Generating trading signals and identifying arbitrage opportunities
What teams are currently using AI solutions most effectively? (Respondents were asked to select three options)
Technology and infrastructure teams
Risk management and compliance teams
Trading and execution desks
Credit analysis and research teams
Quantitative research and modelling teams
Portfolio management and strategy teams
Client reporting and investor relations teams
Data analytics and science teams
Algorithmic trading development teams
Cross-functional teams integrating AI across departments
"AI works best on the research side, where the data actually exists. So, data analytics, portfolio management and strategy teams - that's where AI solutions can be used most effectively. Looking at the underlying holdings within a portfolio and “similar to” credit analysis, looking at the actual numbers and using generative AI to synthesis that data quickly, and in a usable format is where the technology can bring the most value right now."
Neal Rayner, Head of U.S. Fixed Income Trading, Janus Henderson
"If AI is being utilized with traditional human analysis and intuition, I think that it's very additive. If you're just relying on AI, I do agree with some of these statements here. Again, we're still very early, and if the idea of a generative AI tool is that it learns as it goes, we're talking about a very young tool, and I'd rather it be something that’s been through many different cycles and not just looking at past cycles. This is particularly true when you consider that data that was available 10 years ago is not sufficient to analyze in conjunction with the data that we have today."
Neal Rayner, Head of U.S. Fixed Income Trading, Janus Henderson
