The international asset manager expects AI will reach a point in the near future where it can autonomously manage investments within certain parameters set by fund managers.
Robeco is investing heavily in its adoption of artificial intelligence (AI) across both its asset management division and the broader firm, with the asset manager confident that the technology can drive better outcomes across its funds.
Mike Chen, head of Next Gen Research at Robeco, said that while AI has been used mostly as an efficiency tool for tasks such as summarisation and drafting communications so far, it has not been widely applied to investment decisions and actions as of yet.
“There have been a few reasons for this. First of all, the market is very complicated and it’s difficult to predict what will happen. The market is essentially a learning system itself as it changes over time,” said Chen.
“People are also not that comfortable with letting AI run for it.”
However, with AI evolving quickly, Chen said the funds management industry is likely to see the rise of agentic AI in the future, where AI is able to make decisions and take actions not just reactively but autonomously and proactively.
“This is essentially the AI agentic vision where AI [evolves] into autonomous agents. I think that autonomy is really important because at the moment you have to prompt it in order to get an answer,” said Chen.
In the future, Chen said fund managers would be able to direct AI to manage investments on its own within certain boundaries they have set.
“[For example], you could say you can trade anything under $1 million in aggregate. So, if the AI wants to take some sort of portfolio action within that limit it can just go ahead. It would then tell me afterwards why it had decided to do that so that there is a paper trail for auditing,” he said.
“I think this is where the industry is heading.”
In order for AI to be able to operate with a higher level of autonomy, Chen said it will need to be given tools to be able to interact with the digital and physical world.
“At Robeco, we are already endowing our AI with certain functions and access to databases so that it can run basic code and process information it retrieves from the database and then come back with answers,” he said.
Chen also believes that the acceleration of AI will lead to quantitative investing and fundamental investing becoming increasingly integrated.
“Historically, quant and fundamental have been two rather separate disciplines. So, we may see an interesting phenomenon where they merge together because the siren song of what AI promises is too powerful for fundamental managers to ignore,” he said.
“For example, we have an AI that can help us detect emerging market themes and we have actually created a product called dynamic theme machine (DTM). Every quarter it monitors all the market conversations and identifies all the themes that are either going very strong or are emerging right now in the market.
“AI can provide a list of those themes and investment managers can then apply their own fundamental process to the themes it has found.
“So, I do see those two disciplines merging together.”
For fund managers who are looking to increase their adoption of AI, Chen said it is important that they invest a lot in systems and also invest in the right skills.
“You need to have robust infrastructure. All computing can be rented, but you still need to set up the tech stack, the program pipeline, and the data pipeline,” he said.
“You also need people who know how to do this, and those people aren’t cheap. Talent is a scarce and expensive resource from a financial management firm’s perspective, and you need to have the conviction to know that the investment is actually worthwhile.”
Due to these reasons, Chen said there are still varying levels of adoption of AI across the funds management space.
He said that investment managers should not be intimidated by AI, particularly given the improvements in accessibility in recent years.
“I would encourage others to experiment and play with it. Everybody has to start somewhere and I believe we’re only at the very early part of this revolution,” he said.
“Today’s AI is the worst it’s ever going to be, it can only get better and more sophisticated moving forward. That may be unsettling for some, but I think it’s an exciting future we’re going into.”
