The future of investing: How AI is reshaping the industry

Hosted together with Intervalor and attended by around 50 Nordic institutional investors, the discussions moved beyond the headline question of whether asset managers “use AI” and focused instead on how AI is governed, where it adds value, and where human judgment remains indispensable. We used the opportunity for an in-depth interview with our speakers.

Let’s start with a reality check: how far along is the industry really when it comes to AI adoption?

Thomas: In the real world, we don’t yet see many truly agentic workflows – most of what we see is still prototyping. Only very few processes have been fully automated and streamlined.

Furthermore, more output doesn’t necessarily translate into more value: Just because everyone can produce more (longer reports, more presentations etc.) does not mean that these additional activities add value. As a firm, you have to make sure AI does not become an end in itself.

That distinction matters because the survey suggests that many investors and asset owners are already living with AI in their organisations. The practical question is therefore shifting from access to AI towards discipline in its use.

With AI becoming more embedded in decision-making, how should investors think about trust?

Florian: Trust comes from a credible and transparent process, and from results. You should always be cautious about relying on AI narratives: There will be more and more stories about fantastic AI processes. But investors should ask: what does it actually do, and how much does it contribute?

Thomas: Investors should focus on the process rather than on the tools. Ultimately, any AI-enhanced investment approach needs to demonstrate that its results match its promises.

For institutional investors, this points to a more concrete line of questioning for managers: where is AI used, what decision does it influence, what controls sit around it, and how is its contribution measured?

How does working with AI agents change day-to-day research?

Laura: I find them fundamentally different. It’s a completely novel mode of work for me. One of the key advantages is that they’re very good at doing boring, repetitive tasks. They never get tired. At the same time, working with agents provides unprecedented transparency: I can actually watch the thought process and see what happened at each step, especially if something goes wrong.

But there are also clear limitations: Agents tend to be confidently incorrect rather than stopping and saying, ‘I don’t know.’ That makes human oversight essential: We need checkpoints and reviews built into the process.

An interesting parallel emerges when working with both human interns and machines: Both interns and agents are sensitive to vagueness. If I cannot clearly state the task, probably neither of them will provide the right solution. The quality of the result increasingly depends on how precisely humans define the work in the first place.

What new skills are becoming essential in this evolving setup?

Laura: The ability to translate a defined problem into a clear instruction becomes a key differentiator. I am now spending less time writing code and more time structuring the task in my head.

In asset management, that skill is more than a productivity technique. It shapes how research questions are framed, how assumptions are made explicit, and how errors can be identified before they travel further into the investment process.

So where does AI meaningfully improve the investment process today?

Laura: In research, it can significantly reduce time spent on coding or repetitive analyses and help accelerate the discovery of new ideas. This is consistent with broader observations: AI is particularly powerful when it comes to processing large volumes of unstructured data and enhancing signal generation.

The immediate opportunity is therefore less about replacing the investment professional and more about freeing up time, widening the research funnel and improving the speed at which new hypotheses can be tested.

Where should investors remain cautious?

Florian: There’s always the danger that the past is too heavily represented. It is important to think beyond historical data. You have to consider different scenarios: what happens if there is a war, a recession, or another external shock?

Another critical risk is losing oversight: If you don’t understand what’s happening, you lose control. For that reason, strong foundations in statistics and programming remain essential.

This is one of the strongest cautions for investors. AI systems can be powerful pattern recognisers, but markets are shaped by events that do not always resemble the past. Scenario thinking, stress testing and human challenge therefore remain central to any AI-enhanced investment process.

How will investment teams evolve over the next five years?

Thomas: Overall, teams will become smaller, because many preparatory tasks can be automated. At the same time, the role of humans shifts significantly: It becomes important to understand how to supervise these agents and how to keep control. The portfolio manager increasingly becomes a supervisor of AI-driven processes rather than a purely manual decision-maker. As quants, we have a natural advantage here, because we are already very experienced in supervising data-driven processes.

Conclusion: augmentation, not replacement

The rather cautious view of our experts aligns with the results of the investor survey during the event: AI tools are already widely used, but full automation remains rare. Among respondents, 40 out of 47 said their company already uses ChatGPT and/or Copilot without restriction, while only 2 out of 47 said it uses AI agents or AI systems that handle entire workflows or processes. When asked about the biggest challenges in the financial sector, respondents ranked quality and reliability of results first, ahead of human oversight, governance and regulation.

Investors did not overwhelmingly expect the portfolio manager to disappear soon. The more immediate change is subtler: the human role moves up the control chain. For asset owners and consultants, this raises a practical challenge: asset managers should be able to explain not only their AI ambitions, but their governance, escalation points and human accountability.

The lesson from the Nordics series is that AI is neither a silver bullet nor just hype. It is becoming part of the investment process, but its value depends on process quality, human oversight and the ability to challenge what the model produces.

AI is a powerful tool that can:

• significantly increase efficiency
• unlock new sources of data and insight
• accelerate research processes

But success depends on something more fundamental: understanding the process, maintaining control, and knowing where human judgment still matters most.