More information does not automatically lead to better investment decisions. As AI takes on more research, institutions need to know which parts of their investment process can still create an edge.
“Alpha will remain alpha but getting to alpha is a different discussion,” said Jessica Fhulu Mamelasigidi, who works in Investec’s credit investments team.
Mamelasigidi was speaking on ‘The future of allocation’ panel at Alpha Redefined, Investec’s inaugural Women’s Summit for Institutions in Cape Town. The discussion explored how investment teams can use AI without losing the independent thinking that underpins outperformance.
She argued that AI has raised the standard of analysis available to everyone. Work that once needed a large analyst team is becoming easier to replicate, making it “even harder for you to generate alpha”.
That sharpens a longstanding question for boards and allocation committees: how do institutions recognise investment skill and decide whose judgement to back?
Research beyond the consensus
For smaller investment firms, AI can help teams analyse more information without adding significantly more people. Perpetua co-founder Delphine Govender said this is reducing a disadvantage that once came with size.
“It’s allowing analysts and investment professionals to cover a lot more information than we historically could have,” she said.
Govender sees an opportunity to use that capacity on shares that receive little research attention in South Africa’s concentrated market. Automating data gathering can give investment professionals more time to investigate those businesses and form their own views.
But there is a risk. Teams using similar tools can reach similar conclusions. Govender’s team uses AI to identify the prevailing view, then tests its thinking against it.
“We actually use it more to suggest what the consensus might be saying,” she said.
Mamelasigidi said experience also shapes the questions investors ask and what they take from the analysis.
“The prompts that you’re putting in and I’m putting in are very different and therefore what comes out of that advanced analytics is very different. So, we may be using the same tools, but the outcomes are going to be very different,” she said.
Understanding the risk behind the return
Leanne Large, who works in distribution and loan syndication at Investec, asked how investors can differentiate their funds if they use the same models to price transactions.
She described an industry example where AI helped prepare credit papers and a credit committee then used AI to generate questions about them. She called the result “a war of AI”.
For Large, investment decisions still depend on understanding the credit and the contractual terms that apply when something goes wrong. This matters particularly in the specialised transactions her team matches with institutional investors.
“We can’t get away from the human element and understanding credit and what you actually buy,” she said. “You have to be able to really look at the information properly, understand it.”
Mamelasigidi said investors should apply the same scrutiny to the analysis behind a recommendation.
“You’ve got to be able to challenge that data. Start asking yourself where’s the source of this data? What are the formulas being used and why are these formulas being used?”
She uses a ‘devil’s advocate’ prompt to test an argument and challenge her own thinking. But she warned against accepting AI output without checking it.
“It still gets things wrong. It still hallucinates. It still references the incorrect laws or jurisdiction.”
The investment professional remains responsible for the decision. Mamelasigidi compared AI with Excel: an incorrect calculation does not excuse the person who relies on it. “You would take responsibility and ownership.”
Developing the next generation of investors
Govender questioned how firms will develop investment judgement if AI takes over the work through which junior analysts once learnt.
As a junior analyst, she gathered historical financial statements and entered the figures herself. Experienced investors can now automate that work, but new entrants still need to build the underlying knowledge.
“Those skills that might have been learned when you were in a junior stage of a career are sort of being leapfrogged,” she said.
Her concern is what happens as experienced professionals move on. “So, who will come through the system?”
Sindi Mabaso-Koyana, chairperson of MTN South Africa and co-founder of African Women Chartered Accountants, addressed access to experience in her keynote.
“If the same small group is repeatedly described as ready, we need to ask what development opportunities have been given to everyone else,” she said.
She urged senior leaders to give women demanding client work and exposure to investment committees.
Backing independent thinking
Mabaso-Koyana extended that challenge to the selection of fund managers.
“We say that we seek differentiated alpha but frequently allocate capital to what feels most familiar,” she said.
She highlighted the challenge emerging managers face when institutions require a track record before awarding a first meaningful mandate. Without an allocation, a capable manager may struggle to build the evidence an institution wants.
Govender also questioned whether the industry always rewards the qualities it says it values. A persuasive presentation or familiar professional background can influence perceptions of capability.
“Perhaps the industry is valuing confidence over competence,” she said.
Mabaso-Koyana said institutions should also examine how they build their own decision-making teams.
“Many institutions continue to construct their decision-making teams from people with remarkably similar experiences, professional networks and mental models,” she said.
Her point was not to treat women as a category of investment thinking. “Women are not an asset class,” she said. “Women do not all think alike.” Simply appointing one woman to an investment committee will not automatically generate alpha.
The investment edge, then, is unlikely to come from AI alone. It comes from how institutions use the technology, test its output, develop judgement and create room for genuinely different thinking.
2026 Women’s Summit for Institutions
Browse further in