Artificial intelligence is already changing how people work across investment management. From analysing research and interrogating data to drafting communications and managing administrative tasks, the potential to improve productivity is considerable.
But while much of the debate has focused on which jobs AI might replace, two important questions are emerging: how should firms develop their people to work alongside it, and how can they demonstrate to clients that the technology is being used responsibly?
Research from Clearwater Analytics suggests senior asset management executives are relatively optimistic. Some 61% of those surveyed expect AI to increase their job security, compared with 28% who think it will reduce it.
That confidence is translating into action - seven in ten firms surveyed are providing AI literacy and awareness training, 49% are redesigning roles around human-AI collaboration and 63% are investing in internal mobility and redeployment.
Yet are firms doing enough to ensure employees develop the skills and judgement they will need as AI takes on more of their work? And how are they monitoring whether the appropriate safeguards are in place?
See also: In the age of AI, culture may become active management’s greatest differentiator
The risk of outsourcing judgement
The immediate benefits of AI are relatively easy to identify. It can process large volumes of information, summarise research and remove repetitive tasks. But productivity gains do not necessarily translate into better decisions.
At a recent Mediolanum International Funds conference, one example demonstrated how AI can be used not simply to produce investment research, but to challenge it.
Second Foundation, an AI-powered investment research platform developed by the firm, uses opposing analytical perspectives to test investment ideas, exposing weaknesses and assumptions that might otherwise go unchallenged.
The principle is important. AI can reinforce an existing view, but it can also be designed to question it.
That distinction matters for investment firms, where good decisions depend on challenging assumptions, considering alternative perspectives and recognising when something is wrong.
There is a risk that the more convincing AI-generated analysis becomes, the less inclined people are to interrogate it.
And if employees increasingly rely on AI to complete tasks that previously helped them develop expertise, where will the next generation of experienced decision-makers come from?
Can firms demonstrate good governance to clients?
For investment managers, there is another important consideration: how do they give clients confidence that AI is being used responsibly?
As AI becomes more embedded in investment research, data analysis and portfolio decision-making, firms need to consider not only whether the technology improves efficiency, but how they monitor its outputs and demonstrate appropriate oversight.
If AI-generated research contributes to an investment decision, can a manager explain how the information was verified, what assumptions were challenged and where human judgement came into play? And what evidence can firms provide to clients that these checks are taking place?
This raises questions about how AI governance is monitored and reported. Policies and internal controls are important, but firms also need to consider how they can demonstrate that those controls are working in practice.
For example, are AI-generated outputs being checked for accuracy and bias? Is there a clear record of where AI has influenced investment research or decision-making? And can firms demonstrate that the people reviewing those outputs have the expertise to challenge them?
These questions go beyond regulatory compliance. They are also about maintaining trust between investment managers and the clients whose money they manage.
Good AI governance is not simply about managing risk internally. Investment firms also need to demonstrate that the technology is being used appropriately, with clear oversight and accountability.
See also: From framework to practice: How EQ Investors is embedding culture into due diligence
Human skills still need investment
Furio Pietribiasi, chief executive of Mediolanum International Funds, sees AI as an opportunity to strengthen human capabilities rather than replace them.
“My assumption is human plus AI outperforms AI. But if you don't grow as a human, in competencies, in skills, then you are going to be the one that is outperformed by AI,” he says. Mediolanum has introduced an AI development programme across its workforce, while continuing to invest in training younger employees through its Investment Academy.
Pietribiasi is particularly concerned about firms reducing opportunities for junior employees because AI can already perform some of their traditional tasks.
“We need to invest in the youngest generation,” he argues.
If AI takes on the research, analysis and administrative work that has historically helped junior professionals learn their trade, organisations will need to rethink how that experience is developed.
Training people to use AI tools is important, but it is not the same as developing the expertise required to assess their output.
This has implications for governance too. Firms cannot expect employees to provide meaningful oversight if they have not developed the knowledge and confidence to question the technology.
Recruitment presents another challenge
The impact of AI is also being felt before employees even enter an organisation.
At a recent City Hive roundtable held with Simmons & Simmons, participants discussed how candidates can use AI to tailor CVs, applications and interview preparation to specific roles. Employers, meanwhile, are increasingly exploring AI tools to screen and assess those applications.
It risks creating an AI-versus-AI recruitment process in which candidates optimise their applications for technology that employers use to filter them.
AI could help employers consider a much wider pool of applicants, identify relevant skills and reduce administrative burdens.
But, as discussed at the event, if applications become increasingly difficult to differentiate, will recruiters place more emphasis on personal recommendations, existing relationships and professional networks?
And could that create additional barriers for people without those connections?
The challenge is not simply whether AI recruitment tools are accurate or compliant. It is whether their use supports the organisation's wider ambitions around fairness, inclusion and access to opportunity.
See also: AI in investment: Consensus positioning, fee pressure and spotlight on culture
Governance needs to look beyond the technology
These are not questions that can be answered through technology policies alone.
AI governance must consider the decisions being made, the behaviours being encouraged and the capabilities organisations want their people to develop.
That means thinking about where human oversight is essential, how employees can challenge AI-generated outputs and whether roles are being redesigned to make the best use of both human and technological capabilities.
It also means recognising that the risks of AI are not confined to what a system gets wrong. They include what happens to an organisation's culture when people become less confident in their own judgement, fewer opportunities exist to develop expertise or decisions become harder to challenge.
And for investment managers, there is an additional test: can they demonstrate to clients that their governance arrangements are effective, rather than simply pointing to policies that say they are?
AI offers investment firms an opportunity to improve productivity and decision-making. But automating more tasks is not necessarily the measure of success.
The real test is whether firms can demonstrate that AI is improving outcomes while retaining the expertise, judgement and accountability needed to manage the risks.
That is where good governance comes in.
Look out for part 2…