CIOs rethink alpha in the age of AI

What’s occupying CIOs’ minds in superannuation?

AI and quant strategies as well as investment internalisation and impact-focused outcomes were among topics challenging CIOs at a recent Super Review roundtable, in partnership with Northern Trust Asset Management.

Artificial intelligence and quant strategies as well as investment internalisation were among topics on CIOs’ minds at a recent Super Review roundtable, in partnership with Northern Trust Asset Management (NTAM).

At the roundtable in Melbourne, chief investment officers and investment leaders discussed how AI is already helping portfolio managers identify alpha, optimise risk, and process vast amounts of previously untapped data.

Far from being a gimmick, AI is becoming a practical tool that uncovers hidden opportunities and supports smarter, long-term investment decisions. Hosted in partnership with NTAM, the roundtable explored the challenges and opportunities for Australia’s biggest super funds as they embrace AI in their investment processes.

Michael Hunstad, NTAM president, highlighted AI’s growing role in active management, emphasising its potential to transform how large institutional investors operate at scale in increasingly complex markets.

Patrick Nicoll, head of asset allocation at MLC and Insignia, asked how AI differentiates between producing factual information and generating actionable insight, particularly in markets influenced by human behavior, and whether AI can reach that level of insight.

Hunstad said AI and human insight must be combined. AI excels at identifying poor companies, while fundamental analysts highlight the best. Together, they can enhance security selection, though the ideal human-machine synthesis is still developing.

“There has to be a synthesis of human insight and AI at some point… it’s impossible for a fundamental analyst, no matter how good they are, to sit down and say, ‘I know everything there is to know, even about one company, much less 20 or 30 or 40 that’s in their portfolio,’” Hunstad told the roundtable.

“Our own fundamental analysts are using AI tools and helping them kind of make their analysis. Going forward, you need to garner a lot of insight, even for a human analyst to make good security selection decisions. Then you run into the issue of ‘Is the human really necessary, or can the machine do all of it?’ – I think that there’s benefits to both.”

“AI is really good at differentiating between better performing companies and those not doing as well. Fundamental analysts tend to point out the best of the best. If you can bring those two concepts together, I think you’re going to get a happy medium. But we’re not quite there yet.”

For TelstraSuper CIO Kate Misic, she queried how NTAM balances AI-driven models with traditional quantitative principles, ensuring AI identifies patterns with valid economic or financial rationale rather than arbitrary correlations.

Hunstad explained models should never make decisions independently. Instead, AI enhances quantitative strategies based on factors with a clear economic rationale. Some audit trail detail may be lost, but AI helps quantify complex influences like intellectual property or technology momentum.

UniSuper’s head of ESG, Lou Capparelli, asked, “what makes something an AI investment process that isn’t just a quant-based investment process dressed up in a new buzzword?”

Hunstad said AI goes beyond traditional quant by uncovering complex patterns and true drivers of market returns that humans couldn’t find alone.

“One of the big innovations of AI is, instead of just saying, ‘I’ve got these variables that influence returns or macro variables,’ I can actually extract themes from that data that are the true drivers of market returns or growth or inflation… You can have a lot more information feeding into the model than you could ever have done in the past. There’s marketing to it. Don’t get me wrong. The gap in terms of what’s hype and what’s real with AI is getting smaller all the time, and its potency is constantly increasing.”

With only about 45 per cent of institutional assets actively managed, retail investors increasingly drive price discovery via ETFs and emotional trading. While drawdowns can be deeper, rebounds are faster, creating alpha opportunities. Overall, markets are becoming less efficient.

“If you look back 20 years ago, most institutional assets were actively managed – about 95 per cent – so you had a lot of analysts thinking about cash flows, valuations, and profit margins. Institutional asset owners have gone from active management to passive management,” he said.

Matthew Tan, head of asset allocation at LGT Wealth Management, asked about workforce development and staff learning in the age of AI.

“Two years ago, Northern Trust decided all new hires in the investment function must be technologically skilled, able to code and work with data tools like Python or MATLAB,” Hunstad said. “Existing staff are being upskilled, as handling massive data volumes is essential to adding value, even though top talent is expensive.”

Hunstad explained that as super funds and sovereign wealth funds grow, traditional active managers face capacity limits, leaving large portfolios effectively ‘index-like’ despite paying active fees. To generate scalable alpha, these institutions increasingly turn to systematic and quantitative strategies.

“Part of the challenge of traditional active management is that capacity constraint, but as these funds get so big, you have to look for other sources, and systematic strategies tend to be great source for that,” he said.

“Our quant clients are the largest institutions in the world. We recently entered the Saudi Arabian market with PIF and NDF as clients, huge sovereign wealth funds, looking for scalable sources of alpha, and that’s exactly what they’re trying to achieve through quantitative implementation. I think scale really changes the dynamic of how you look for active exposure in your portfolio.”

The senior investment leaders were also asked whether it becomes harder to achieve alpha as a fund grows.

CFS’ head of cash and derivatives John Iles noted that many large Australian funds respond to growth by hiring staff, as they can afford the cost of internalisation.

“I’m not so sure that many organisations are investing in multi-year, multi-million dollar technology projects. Perhaps, if we really want to internalise asset management, I think possibly that ought to be the place to spend scale dollars, not just on FTE headcount,” Iles said.

“I think that’s one of the challenges for boards and CIOs over the next 10 years. It’s great that you’ve got scale – how do you spend that scale?”

TelstraSuper’s Misic added that as an investor’s size and capabilities change, the sources of alpha must also evolve.

“I do think that your source of alpha needs to change with your size and other endowments that you have as an investor. Size is one of them, but it also might be your innate capabilities, your internal capabilities,” she said.

“I don’t actually believe that there is a size at which alpha is not achievable, but that your sources of alpha and what you’re going to do to try to add alpha needs to change depending on your scale and your capability. Whether you’re looking at your people capability or system capability, I think is an important question.”

Capparelli noted UniSuper is one of the few funds that has largely internalised its asset management, observing that internalisation works for some funds but not all.

Hunstad said many funds start internalisation, but few complete it, and some eventually revert to outsourcing. Economics – high data, technology, trading, and talent costs – make internalisation challenging, especially for passive strategies. Some institutions pursue it for strategic reasons, but success requires careful planning, scale, and paying for top talent, particularly on the active side.

Participants noted that internalising private markets like infrastructure, property, private equity, and credit is more complex than listed assets due to lumpiness and diversification challenges. Hybrid models, co-investments, or outsourcing are often more practical.

Direct deals can work for some funds, but access to top managers via fund-of-funds remains valuable. Building internal IP and research capabilities is possible but takes time, often starting in silos before evolving into a cohesive ecosystem.

The discussion also highlighted a shift from scoring and ranking methodologies toward impact-focused outcomes. ESG integration now embeds sustainability across investment processes rather than just labeling products.

Active ownership, exclusions, and forward-looking sustainability considerations drive future returns. While some funds offer explicitly labeled sustainable options, ESG practices increasingly apply across entire portfolios, focusing on meaningful outcomes rather than just metrics.

Australian funds were praised for leadership and collaboration in ESG, balancing competitiveness with shared responsibility.

“One of the things that I love about the industry that we work in … I do think Australia stands proud in its record of integrating ESG into its processes and taking it very seriously and being quite vocal on the world stage and really calling out when we think that it’s a little bit of greenwashing etc,” Misic said.

“I think that we do have a lot to be proud of across the industry. Many of us at this table will be some of the longest-standing UN PRI signatories in the world, and we do take it really seriously.”