As Australian banks lean further into AI, the asset manager has identified the sector as being largely insulated from disruption.
In assessing ASX 200 companies against the ongoing impact of AI, Fidelity International is targeting long-duration winners and has found some sectors may effectively act as AI fortresses.
At a recent roundtable, portfolio manager Sam Heithersay said the firm has built an 11-part framework to gauge companies’ ability to withstand AI disruption.
The work began in September, when the firm concluded AI posed a real threat to many of its portfolio companies’ business models. It comes after other firms, including Morningstar, have applied an AI-risk lens to their ASX coverage, leading to downgrades in parts of the tech sector.
At Fidelity, each analyst reviewed their full coverage, assessing competitive moats such as capital intensity and human trust, alongside the “adoption friction” that could limit companies’ ability to deploy AI effectively. These included factors such as unionisation, which could limit operational efficiencies, as well as data security and trust concerns that create friction in adopting AI.
It also reviewed how companies had sold off over the past six months, creating some pockets of value. The next step is assessing whether companies are agile enough to adapt and evolve, with the work still ongoing through CEO and board-level meetings.
Though Fidelity examined companies from a bottom-up perspective, Heithersay said some “adoption friction” factors were sector-wide rather than company-specific. It also found that entire sectors could be AI-proof.
“We identified sectors and companies that we would call a ‘fortress’, in the sense that it’s very hard to understand how AI will necessarily disrupt their business models. They share a lot of the same characteristics that you might expect: capital intensity, regulatory moats, and the kind of primary concerns around safety and trust.”
While Heithersay said utilities unsurprisingly screened well on these factors, portfolio manager Zara Lyons added that banks also score strongly on the same measures.
“Obviously, they’ve got strong capital requirements imposed upon them by regulators. And in addition, I suspect that the regulators themselves will be slow to move with AI adoption, so that, to some extent, creates a fortress-like designation for them. They also may benefit from being able to adopt AI within their organisations,” Lyons elaborated.
She said that in meetings with banks, discussions have focused on how many software engineering hours could be reduced through AI tools used in coding. There is also potential for the technology to be applied to areas such as anti-money laundering (AML) work in the future.
It comes as AI has been a major theme for Australia’s big banks this year, with several rounds of job cuts from Commonwealth Bank (CBA) and others partly attributed to efficiency gains from AI, with most affected roles in technology teams. ANZ has also recently created an AI chief role, signalling the strategic importance of the technology for the bank and positioning itself to compete with peers in the space.
Meanwhile, in fraud and AML, CBA announced late last week that it has developed an AI agent that identifies emerging fraud and helps strengthen defences against scam patterns.
Lyons said she believes that banks will be working closely with regulators to keep people in the loop around these developments: “For obvious sensitivities around the workforce as well as not letting the machines run off with our financial system.”
But the view was not uniform, with Fidelity’s head of investments Paul Taylor adding that no sector is purely positive or negative in the AI transition. He said banks, for example, traditionally benefit from customer inertia, with people often not moving their money in search of better interest rates.
“There’s an argument that you could just say to an AI bot, ‘move my money to the highest interest rate’, so you can get between banks’ competitors. So now all of a sudden, banks can’t benefit from apathy. If that pays off, they don’t benefit,” Taylor said.
He concluded there are multiple paths if AI enables more open banking, but cautioned against getting ahead of developments.
Comparing AI to the dotcom era, Taylor noted that banks were initially seen as “losers” from the internet, but ultimately emerged as some of the biggest winners.
“It allowed them to move from what was a very expensive process of people coming into a branch and doing the transaction to everybody now doing the transactions on their phones or on their computers, and that massively reduced the bank’s cost of doing business.
“People might not like the banks, but they trust the banks, and a lot of the time the internet platforms weren’t at that same trust level, so people didn’t move.”
