Fund managers now use artificial intelligence in at least six distinct parts.
The majority of fund managers now use artificial intelligence in at least six distinct parts of their investment processes, Ethan Spiegel, senior investment analyst at Zenith Investment Partners, says.
“We’ve moved beyond whether managers are using AI to how they’re using it, and it is more varied than most people expect,” Spiegel says.
The six ways fund managers are applying AI in practice are:
- improving operational efficiencies
- information sourcing and fact-finding
- data verification and challenging ideas
- industry research
- data processing, and
- coding.
The most common application, improving operational efficiencies, involves automating routine tasks such as drafting emails, writing internal memos and handling administrative work. This use case cuts across all manager types and often serves as the gateway to wider adoption.
“Once a team sees AI take an hour off their week on admin, they start asking what else it can do, and that’s usually how the adoption begins,” Spiegel says.
Another widespread use is information sourcing and fact-finding, where analysts rely on AI to quickly summarise large volumes of reports and stay on top of news flow for individual stocks and sectors. Spiegel says this doesn’t replace direct engagement with company management, suppliers or competitors, which remains central to fundamental investing.
Analysts are also turning to AI to stress-test their investment theses and uncover risks they may have overlooked. While the technology most often confirms existing views, it regularly highlights factors that had not been properly weighted.
“This is a great way of pressure-testing ideas,” Spiegel says.
“Most of the time, AI agrees with you, which is reassuring in itself, but it sometimes flags something that hadn’t been weighted properly, and that’s where it can be extremely beneficial.”
For industry research, AI can help analysts rapidly get up to speed on niche sectors and unfamiliar market structures, and is especially useful for research teams covering a broad range of asset classes.
Bradley Antman, investment analyst at Zenith, says the technology has heavily reduced the time it takes for research, allowing fund managers more time to focus on more pressing priorities.
“AI has compressed this from days to hours, which means the meeting itself can go straight to the questions that actually matter,” Antman says.
Among quantitative managers, data processing is a key use case. These firms work with vast amounts of alternative data, including credit card transactions and foot traffic numbers, to inform their investment theses. AI is used to clean that data by removing outliers and errors before it feeds into their models.
The final use case, coding, has long been a feature of quant management, though has not previously been executed by AI. Internally developed AI tools are now accelerating model building and refinement. However, Antman says that coding had become so accessible it was now a baseline expectation rather than a competitive edge.
“Coding was arguably the first place quant managers used something AI-like, long before anyone called it that. What’s changed is that it’s no longer a point of difference between managers,” Antman says.
“That said, what AI is not doing is selecting stocks. Humans can interpret probabilistic outputs, build relationships with company management, and make investment calls – traits that are becoming more valued, not less.
“AI has access to all the information that exists today, but it doesn’t have an understanding of what might come tomorrow, which is where a manager’s relationships and deep experience drive value for investors.”