How fund managers are using AI

AI feels like it’s everywhere at the moment, penetrating our day-to-day lives more and more as industries and people are confronted with how this technology will change the way we live and work.

The asset management industry is no exception: it’s not just attempting to invest in companies benefitting from AI advancement, it’s also competing in how it uses the technology to deliver a higher level of returns.

Specific hiring

Ben Rogoff, manager of the Polar Capital Technology trust (PCT), has been vocal about how he goes about investing in AI as a theme, looking across the “ecosystem” as he describes it, targeting both the companies enabling the technology and those applying it to “generate tangible economic value”.

This strategy has contributed to the trust making almost 230% over three years, ahead of its Technology & Technology Innovation sector peers (213%), according to data from the Association of Investment Companies.

But identifying those AI ‘winners’, and general stock picks, has become a process infused with AI in and of itself.

Rogoff’s fellow manager, Alastair Unwin, says the 12-strong investment team had “incorporated a lot of AI into our own process” allowing them to “reach more companies and parts of the market”. They say it's ‘supercharged’ their process.  

The trust made a series of hires to bolster its AI investment capabilities, with one person hired to “specifically drive AI through their investment process”.

In their team, AI is not picking stocks, but helps drive new ideas, which Unwin says are now “overflowing”. AI also reduces the admin burden such as writing up earnings calls after results to allow the team to focus on higher value-added work.  

The PCT team conducts over 1,000 company meetings each year and Rogoff exalted this human touch at a recent AJ Bell Retirement Money Show event.

The team uses Claude to “do an analysts job”, building a bull and bear case on a stock “in minutes”, Rogoff said.

‘Summarising’ is so last year

One of the major things AI is hailed for is making it easier to get a quick summary of long or complex topics.

The world is awash with anecdotes of people using it to give them a bullet point summary of hefty legal documents and contracts, presidential speeches and news stories.

In markets, it’s often used to summarise investor calls, allowing analysts get coherent information of more companies than they could individually cover without the tech, just like Rogoff’s aforementioned example.

George Gatch, CEO of JP Morgan Asset Management (JPMAM), says that currently, one of their research analysts will cover 30-40 companies, but in ten years time, this will double or even triple.

“The machine does the digging, and then the human does the decision-making and applies judgment to the process,” he says.

But with the widespread use of Large Language Modes (LLM) like ChatGPT by any casual or seasoned investor, this has now become the baseline AI use case of active fund managers.

“Summarising an earnings call is useful, but it doesn’t provide any edge. That’s very 2025 now”, says Anuj Arora, Head of Emerging Markets and Asia Pacific Equity Group at JPMAM.

“Everybody's doing it. The retail investor does it. Our competitors do. But just because you've got ChatGPT doesn't mean you're a smarter investor suddenly,” he adds.

Building their own AI tools and models

To push beyond this level, JPM has invested heavily in the build out of its own AI tool, Spectrum, a proprietary data-driven investment platform. JPMorgan Chase & Co. spends $20 billion on technology alone and has 1,600 engineers within its asset management business solely dedicated to building AI tools.

Spectrum is now used to automate 75% of its intraday trading, the automation of which saved clients $4 billion in transaction costs last year.

Their fund managers use SpectrumIQ. This takes in 7,000 broker research calls a day, covering 90,000 securities, 11 times more than what it could handle just a couple of years ago.

This allows managers like Arora to take a run of the mill earnings call on his top holding – MercardoLibre – and run a multi-dimensional analysis across his current and historic convictions. Through this, he can see if the stock is operating in line with his investment case, if his analysts are going above or below market consensus in their reports, if he needs to consider rebalancing his position and his own historic biases all within a couple of tube stops on his morning commute.

Previously, all the topline information existed in different systems, like Bloomberg terminals for macro news, emails for fundamental research, and risk management was its own siloed system . Before, none of these interacted, but now, Arora describes it as a “mosaic of data”.

Spotting risks

One thing Arora can now be alerted to through SpectrumIQ is if he’s built up an otherwise unnoticed risk exposure he doesn’t want in the fund.

Often when fund managers like a stock and buy it, they look for similar characteristics in other options, but these positions can unintentionally add up to one bigger one risk or theme.

“Five years ago, when my clients asked me how my funds were exposed to Covid-19 or inflation, I didn’t really have a very good response,” Arora says.

“Now when they ask me what's your exposure to AI, or the Iran war, or defense spending, I give them a number. And I can also tell them the stocks that are contributing to that number. And when our clients see this, they know that we are not just tracking the risk, we're ahead of it.”

Fellow asset manager Schroders also created its own generative AI assistant, Genie, and last year, it unveiled ContextAI to help its analysts and fund managers with sustainability-focused research, utilising LLMs  to “create an AI environment that is uniquely ours”, says Pablo Riveroll, global head of equities research at the firm and a fund manager on its Latin America and Global Emerging Markets desk.

“2026 is shaping up to be the year when AI moves from a productivity tool to investment insight,” he says, particularly around risk.

Like JPMAM’s Arora, Riveroll sees 2026 as the year AI moves from being purely a productivity tool to one for investment insight.

He says AI is now in its ‘second phase’ in the firm, helping them organise thinking, track convictions, understand exposures, and learn from the outcomes.

“This kind of thesis-level transparency, augmented by AI’s ability to continuously process information, represents a genuine step change in how we understand and manage risk,” Arora said.