Artificial intelligence is rapidly becoming embedded in investment management. From research and due diligence to sustainability analysis and portfolio monitoring, investment teams are discovering new ways to process information faster, automate repetitive tasks and improve efficiency.
The conversation about AI often centres on what the technology can do. How much time can it save? How much can it improve productivity? Which parts of the investment process can be automated? These are important questions, but they are not the most interesting ones.
The more significant question is what happens when AI becomes commonplace. Because while firms may currently be focused on gaining an advantage through technology, the reality is any competitive edge created by AI is unlikely to remain exclusive for long. Eventually, access to sophisticated tools will become widespread, capabilities will converge and the technology itself will cease to be a differentiator.
At that point, the focus shifts from the tool to the people using it.
See also: How AI is transforming our understanding of biodiversity risk
This was one of the most striking themes to emerge from a recent discussion with Ashley Oerth of Invesco, Jenn Hui Tan of Fidelity International and Fraser Lundie of Aviva Investors at the Investors ACT 2026 Conference.
While all three were enthusiastic about AI’s potential to improve investment processes, they were equally clear that technology does not eliminate the need for human judgement. If anything, it increases its importance.
For active managers, this presents both an opportunity and a challenge.
For decades, investment management has competed on the ability to gather, analyse and interpret information, but now AI is changing that dynamic. Tasks that previously took hours or even days can now be completed in minutes, while large volumes of structured and unstructured data can be assessed almost instantly. Further, investment professionals are gaining access to tools that would have seemed unimaginable only a few years ago.
Yet greater access to information does not automatically lead to better decisions.
Indeed, there is a risk that the opposite occurs as if investment firms increasingly rely on similar models, trained on similar datasets and designed to identify similar patterns, then investment decision-making itself may become more homogeneous.
As Fidelity’s Hui Tan pointed out during the discussion, there is a possibility that investors begin approaching risks and opportunities through increasingly similar lenses, creating the potential for greater overcrowding in markets.
That concern was echoed by Lundie, who suggested AI could accelerate consensus positioning by enabling investors to arrive at the same conclusions more quickly than ever before. While markets have always experienced periods of crowded trades, the speed at which those positions develop may increase significantly as AI adoption grows.
If that proves to be the case, then one of the industry’s most valuable assets will not be technology, but the ability to think differently from competitors.
Challenge and perspective
This is where culture becomes critically important. The investment firms best positioned for the future are unlikely to be those that simply deploy the most sophisticated AI systems. They are more likely to be the firms that create environments where those systems are challenged, interrogated and complemented by diverse perspectives.
Technology can provide answers but culture determines whether people ask the right questions.
Strong cultures encourage intellectual curiosity, constructive disagreement and genuine debate. They create an environment where individuals feel comfortable challenging consensus views and questioning assumptions, even when the data appears compelling.
These characteristics have always been important in investment management, but they become even more valuable in a world where AI-generated analysis can appear authoritative and convincing.
In fact, one of the more encouraging observations from the discussion was that AI may create more space for precisely these kinds of conversations. By reducing the time spent on manual data gathering and routine analytical tasks, investment professionals can devote more energy to interpretation, debate and decision-making.
Lundie described how discussions within his own team have evolved, with greater emphasis now being placed on testing assumptions and exploring less obvious perspectives.
That shift feels significant. The true value of an investment team has never been its ability to collect information, and increasingly, AI can do that. The value lies in determining what information matters, understanding its limitations and applying judgement in situations where certainty does not exist.
The same principle applies to governance: while AI will undoubtedly play a growing role in supporting investment decisions, accountability cannot be delegated to an algorithm. As Tan noted, fiduciary responsibility remains the clear boundary. AI can inform decisions, but responsibility for those decisions ultimately remains with people.
This distinction matters because trust remains at the heart of investment management. Clients do not entrust their assets to algorithms. They entrust them to investment firms and the individuals within them. Technology may support those professionals, but it does not replace the responsibility they carry.
The industry is therefore entering a period where technological capability and human capability will need to develop side by side. Firms that focus exclusively on the former risk overlooking what may ultimately prove to be their most durable source of competitive advantage.
As AI becomes more accessible and more powerful, the differentiating factor is unlikely to be who has the best model. It will be who has built the culture capable of using those tools most effectively, challenging them most rigorously and applying them with the greatest degree of judgement.
In a future shaped increasingly by artificial intelligence, culture may become one of the few advantages that cannot be automated.
