From finding ‘needles in haystacks’, to identifying the most niche of strategies, AI’s prospects look bright and gen AI could evolve to be a cherished colleague.
Participants
- Brian Slattery, head of Northern Europe, Clearwater Analytics
- Michaela Campbell, managing director, Hayfin Capital
- Louise Jack, COO, Local Pension Partnership Investments
- Victor Mayer, managing director, Pantheon
- Emily Pollock, client director, Schroders Capital
As long as it doesn’t replace you, you may come to see AI as the perfect colleague. AI can out-compete you in many areas – but at least it puts all this effort into helping you. It hunts down information in the blink of an eye and does your menial tasks. And AI won’t complain if you take all the credit.
Whether you welcome AI or not depends on if you are what Emily Pollock, co-founder and client director of the private assets solutions business at Schroders Capital, calls a “power user”. Speaking about the acceptance of AI and its integration into everyday business practices, she said Schroders had invested a lot in building tools to leverage AI and wider technology. However, once firms have done this, they could find it a challenge to get people using these tools consistently.
“Firms will likely find that as younger, tech-savvy individuals enter the workforce, there will be a natural progression towards broader acceptance of AI. In the meantime, there will be varying levels of enthusiasm and expertise regarding AI.
“The ‘experts’ and ‘power-users’ are few but crucial, while the ‘enthusiasts’ are more numerous. But all are vital for early adoption,” she said.
Louise Jack, chief operating officer at LPPI, thinks about users as either “enthusiasts” or “explorers”.
“Firms will likely find that as younger, tech-savvy individuals enter the workforce, there will be a natural progression towards broader acceptance of AI”
Emily Pollock, Schroders Capital
“The experts and power-users are probably a tech team. There are not many of them but they are experts in AI. The enthusiasts are larger in number and love it. They really want to try AI, while the explorers will simply move on to something else if they find out quickly that AI doesn’t seem to work for them.”
Trained models
So, does it work for our panel? Michaela Campbell, a managing director in the direct lending team at Hayfin Capital, said: “The industry is relatively early in this journey, but the pace of improvement and adoption will only accelerate. The models will need to be trained so that the output is more useful to people, but I do think that the pace of acceleration is getting faster and the models are certainly getting better and better.”
She added: “The data we receive is expanding and it’s multi-lingual. I think eventually AI will not only be able to read information from PDF documents, but the AI will actually make some sense of it.”
Campbell emphasised the challenge of handling unstructured data – data that is not organised in a pre-defined manner – from various sources. She highlighted the industry’s ongoing efforts to solve this problem through generative AI models.
“The data we deliver to asset owners or investors often comes from numerous unstructured sources. Solving this is something the whole industry is trying to deal with,” Campbell said.
“Eventually AI will not only read information from PDF documents, but the AI will actually make some sense of it”
Michaela Campbell, Hayfin Capital
She expressed optimism about the accelerating pace towards solutions, predicting that data would soon be gathered and analysed much more efficiently. “It’s going to be a competitive differentiator to get this data, store it, analyse it, and distribute it faster and in a more standardised format,” she said.
Brian Slattery, SVP and head of Northern Europe at Clearwater Analytics, said technology would facilitate the organisation of both structured and unstructured data to work in lockstep and that AI’s role was pivotal for this.
“Without technology, I don’t think there is a chance at all trying to solve this data problem that affects the whole industry. Utilising AI helps communities to have a unified point of view as close to real-time as practical,” he said.
Needle in a haystack
Slattery also questioned the persistence of manual tasks in the digital age, arguing that AI could automate many processes.
“AI can even stabilise a team by retaining critical information that never leaves the organisation. There’s no risk of departures”
Victor Mayer, Pantheon
“What happens when there is a data point that needs exploring – such as a big spike in book yield, or a capital call that doesn’t quite make sense – and you need to investigate it but you don’t know where the information is? It’s like finding a needle in a haystack and a lot of times the portfolio manager would have to go and look for it, wading through large volumes of data, even across different jurisdictions.
“Today, an AI bot can take a specific prompt in plain English and respond with information in seconds.”
The needle in a haystack can also be an investment strategy so niche that it can get missed by human brains. Victor Mayer, of Pantheon, said a particular strength is AI’s ability to handle large data sets that form the basis of investment decisions. Most firms will have 12 to 15 investments in a fund, he said. But in the process of finding those investments, they will have reviewed and filtered perhaps a hundred to 150 companies.
“A universal theme we observe across people we work with is that the biggest upside is for investment execution. For example, instead of having five analysts mapping the market to find investment opportunities, AI can act like in-house eyes, finding investment themes that are niches within niches. This efficiency in data handling allows for a more streamlined approach to identifying investment opportunities and executing strategies.”
LPPI’s Louise Jack suggested AI’s potential is far from fully realised in the asset management world and that interest in AI, together with generative AI’s effectiveness, will grow as it becomes more embedded in business processes.
“Today, an AI bot can take a specific prompt in plain English and respond with information in seconds”
Brian Slattery, Clearwater Analytics
She added that LPPI had trialled generative AI and its usefulness had been demonstrated for summarising minutes from meetings, as an example.
“We are now challenging AI to carry out more complex tasks, such as analysis of spreadsheet information detailing historical performance, for example. It is challenging to make the AI analysis relevant and nuanced enough. Analysis of investment data is complex. The use case is not as simple as understanding sales data on how many widgets were sold in Arizona.”
She added: “You can start to see the power of this when software providers build AI into their existing software and natural language processing gives us the ability to ask questions of the data, saving a lot of time. Ask for a document containing a specific piece of information and it is retrieved very quickly.”
However, Jack sounded a practical note, emphasising the importance of data security and the need for thorough checks before adopting third-party tools. “Before adopting any AI tool, it’s crucial to understand what is happening to your data and where the boundaries are,” she said.
Perfect employee
As well as being a near perfect colleague, for business owners it could also be the perfect employee.
Pantheon’s Victor Mayer said a well-integrated AI system is like a “permanent brain” for an organisation, mitigating the risks associated with employee departures.
“It becomes an extension of your team that will help you scale the business without requiring the same expansion in headcount. It can even stabilise a team by retaining critical information that never leaves the organisation. There’s no risk of departures, no risk of people leaving with all that knowledge.”
The consensus is that AI holds immense potential for transforming private-market processes. However, its success depends on widespread adoption, proper data management, and continuous improvement of AI models.
