When implemented effectively, AI-powered solutions can automate repetitive tasks, enhance integration across disparate systems and uncover valuable insights to boost alpha performance. Market practitioners at the Equities Leaders Summit in Miami this week explored what AI means for the trading desk, and how it can help firms to parse information from fragmented data points to improve trading decisions.
A key challenge in the adoption of AI at the coalface of financial markets is trust. ”We are trying to find practical use cases for AI that aren’t a waste of time,” says Medan Gabbay, chief revenue officer at execution management system (EMS) provider Quod Financial. ”You can’t use it to invest, because your brokers won’t accept it.”
‘An army of interns’
One of the obvious benefits of AI is the ability to automate manual, time-consuming processes – and that is where progress is being made. “The biggest success that we have had using AI has been in automated testing,” says Gabby. “It doesn’t sound sexy, but we can build test scripts using natural language, and we’ve gone from making 50 tests a day to 1,000 tests today. Now we can deploy upgrades every month. You want new features? We’ll test it. We’ve got 15,000 to 25,000 automated tests running monthly, all of it enabled by AI processes.”
AI is not coming for your job. However, somebody who’s properly using AI is probably going to take your job
This automation not only reduces operational costs and frees up resources for higher-value work, but it also helps to mitigate the risk of human error. As Quod Financial’s Gabbay explains: ”The goal is to find these use cases that actually benefit your business. It’s not about AI suddenly building a use case for you.”
Adam Sussman, partner at asset manager Stack, was also bullish about the opportunities to save time and money through AI, likening it to “an army of interns” with the ability to save substantial time and effort during the research process. “You can go on to any of the LLM [large language models] and have them write a Python script that will download all the 606 reports, put it into a data format and do the analysis for you. Whereas four years ago, you’d either be spending a lot of time on Excel, or needed someone smart enough to know Python. Now, you can just copy and paste.”
Golden-use cases
Another area where AI is transforming financial firms is in the realm of data integration and analysis. Manus McGuire, of data analytics platform KX, notes: ”We have a front-row seat. We’re definitely very much in the trenches with our customers.
Action points
– Explore opportunities to leverage AI for improving data quality and monitoring processes;
– Investigate how to enable more organic, bottom-up AI adoption within the organisation, rather than a centralised, top-down approach;
– Identify specific, repeatable tasks that can be automated or enhanced through AI, rather than focusing on one-off projects;
– Ensure that the underlying systems and data infrastructure are flexible and enable easy integration of AI-powered solutions.
“When you’re layering and repairing those unstructured data sets on top of your structured data, that’s when you can really drive out what we call golden-use cases. These are where you’re not necessarily looking at efficiency or productivity gains – you’re looking at ways you can improve your product or service so that you can charge more money for it.”
AI’s potential to enhance investment performance is the ultimate dream, as AI-powered systems hold the promise of identifying signals and patterns that human analysts may miss, enabling firms to gain a competitive edge.
But the key question is what exactly is being replaced. Stack’s Sussman says: “If it’s a one off, you probably don’t want to use AI, because by the time you figure out how to ask the right question, you probably could have just done it yourself. But if you have to do this task 500 times a year, it’s probably worth getting AI to do it for you.”
Learning how to ask the right questions is key. Jesse Forster, head of equity market structure at Coalition Greenwich, comments: “AI is not coming for your job. However, somebody who’s properly using AI is probably going to take your job.”
Make your mark
AI adoption that solves genuine problems remains the key challenge. McGuire at KX points out that investment banks would tend to set up a horizontal team, such as an AI Swat team that assists different desks, in a centralised model. Hedge funds and buy-side firms can take a more organic approach, leveraging individual experience in AI or ML to grow it organically.
Gabbay says: “It’s the kind of ground-up implementation that needs to come from individuals, and then it’s the business’s job to support that and build it out. But in general, banks are terrible at sponsoring stuff – they’re great at crushing innovation. Because innovation isn’t what our industry is built around. Our industry is built around not changing.”
But on the buy side, there are more opportunities for individuals to make their mark, he argues. “Sell-side technology is starting to enable that, adopt it and refine it.”
While adoption of AI in financial services is still in its early days, the evidence is clear: firms that strategically leverage this transformative technology can achieve significant operational efficiencies, enhance integration and data utilisation, and, ultimately, aim to generate those ambitioned superior investment returns.
