Practical uses for AI in the investment management industry
BY Yaelle Gang | April 1, 2019
While the term artificial intelligence was first coined in 1956 and has come in and out of vogue, it’s likely here to stay this time, said Paulo Salomao, managing director of asset management and pensions at Accenture, speaking at the CFA Society Toronto’s annual spring pension conference on March 28.
An AI winter isn’t on the horizon, he said, referring to the amount of computing power, the abundance of data, the amount of research on AI and the level of investment in AI.
While AI is here, it’s at different levels of maturity and visibility across the investment management industry.
When ideas develop, the visibility is higher and people are excited, but the outcomes aren’t there, said Salomao. Yet when real value starts to be exploited, the visibility dips as the AI matures. “At that point, people are a little more cautious of the amount of noise they can make because they start to believe they can actually profit from this.”
As techniques propagate and are adopted, visibility starts to rise again and continues to rise as the techniques are optimized and take up broadens, he added.
Using AI for public markets security selection is still in the idea development stage, for example. “We talk a lot about using AI to augment the role of humans, of traders, when you’re making investment decisions,” said Salomao. However, he showed an example of a fund doing this that hasn’t performed well. “This is not to say that it’s not going to happen. It’s just to say that there’s a lot of hype, but it’s really not mature.”
Moving along the maturity spectrum, real value is starting to be exploited from using AI in investment risk management, he noted. And areas where techniques are starting to propagate include investment data analysis and reporting, consolidation of portfolio reports and the use of advanced sentiment analytics for public disclosures, said Salomao.
When it comes to consolidation, for example, minority investors can’t necessarily dictate what a portfolio company report looks like. “There’s a lot of people who spend time putting together the reports, extracting financial information, trying to understand what’s changed versus last year’s report or [the] first quarter report, and so AI can help with that. It can release time for your people to actually go do things that are higher value-add, like thinking about what to do with that information.”
Equity research optimization is one example in the category of techniques now being optimized, he noted, referring to Alphasense as a specific example of an AI-based platform that aggregates different sources of equity research, including internal data. “You can actually read through the paragraph before you pay for that page of your equity research report and you can make sure that it’s actually relevant to what you’re trying to accomplish.”
This can reduce the price of equity research, added Salomao. “I think this is quite powerful, but the powerful thing isn’t necessarily only the platform, it’s the transformation of the investment business model that comes behind it.”
Some areas where the industry is more mature when it comes to AI include using robotic process automation for middle and back offices or virtual agents for employee support, he noted.
“I don’t think AI is a panacea, but I actually think that it can unlock large value in some very specific areas of our industry. And so I would encourage you not to stick your heads in the sand and ignore it, but actually proactively explore it.”
In the investment management industry, there’s value in first-mover advantage, said Salomao. “We make money off of information asymmetry. That’s effectively how this works. So AI can help you get an edge in terms of information asymmetry.”
And building a strong data foundation is key. “You need to be able to nurture your data and the quality of your artificial intelligence insights will only be as good as the quality of the data that you have behind it.”