Eric Heliene, head of buy-side trading desk, Groupama Asset Management
In the buy-side trading world, technological innovation is a perpetual driver of transformation, and large language models (LLMs) are emerging as key players in this ongoing evolution. These models, with their ability to analyse vast volumes of unstructured data, are opening new perspectives for understanding market dynamics. The extraction of sentiment and trend indicators from diverse sources – financial news, expert analyses, discussion forums – could revolutionise how buy-side trading desks approach investment decisions.
These real-time insights allow for quicker reactivity and adaptation to market fluctuations, providing a valuable competitive edge. Furthermore, LLMs promise to transform the execution of trading orders. By anticipating and assessing the potential market impact, these models could optimise timing and pricing strategies, thus minimising costs and maximising transaction efficiency. This ability to dynamically adapt to ever-changing market conditions signifies a significant step forward towards more strategic and thoughtful trading management. Next year could witness this revolution, where LLMs are no longer just analytical tools, but integral strategic partners in the buy-side trading world.
Sylvain Thieullent, chief executive officer, Horizon
The world of trading has become increasingly automated, relying on quantitative analysis, predictive models, and advanced algos to increase both speed and efficiency. It is without question AI is going to continue to dominate headlines in 2024. The data required to meaningfully apply AI is becoming cleaner, and the technology tools smarter, and I’d argue this marks the beginning of a completely new way to trade.
One of the use-cases set to take off next year is around liquidity management, utilising AI-based tools to seek out market inefficiencies and generate more profitable trading opportunities. Recognising when the market’s liquidity is changing is essential for traders, investors, and financial institutions, and AI can be used to enhance the way in which we identify these fluctuations. Applying this to trading strategies, market participants will have an even more optimised way of executing orders for their clients.
Lisa Schirf, global head of data and analytics, Tradeweb
Artificial Intelligence (AI) being front-of-mind for next year is no surprise, but it does continue to present significant opportunities in applying data science to develop more transparent, intelligent and efficient ways to trade. AI models are useful in looking at historical information and estimating what could happen in the future based on a set of parameters. It can synthesise all of the information that goes into an optimised execution decision by helping to predict the best possible parameters, such as timing, dealers chosen and number of dealers selected, based on a set of given requirements. We expect to see the technology increase low-touch trading, with trades being executed without human intervention. We’re also looking at the technology’s capabilities in liquidity predictions.
We expect the use-cases in electronic trading to only keep on growing next year, particularly in the management of this proliferation of available data we’re seeing, from solutions for predictive prices for dealer selection and count, to different trading protocols, to handling less liquid securities. As we see the demand from the market increase, we’re already building model-based prices with advanced AI techniques that allow us to predict the price of a security based off of public and Tradeweb proprietary data sources. For municipal bonds, this means leveraging proprietary data science to calculate both intraday and end-of-day prices for these bonds.