“First of all to be totally transparent, I used ChatGPT as well, to get some of the questions and validate the questions which we’ll ask. We also asked ChatGPT who would be the perfect panel for this discussion.” So began Shaun Lee at State Street, moderating an industry panel on 29 November.
End goal
Globally, the end goal of using AI in the private markets fund industry centred around improving efficiency, decision making and overall performance.
Lee told the panellists: “To really have a streamlined client experience, I think, we’re looking into different tools to try to make that a smoother process for them, but also not kill the workforce…. we could onboard investors in such a clean, scalable fashion, if we can figure this.”
Cara Browne at EQT said: “Managing the number of minutes and investment committees, board meetings, etc. That could be an area where we would see loads of efficiency gain instead of taking me meeting minutes. And then all of us having to review them however many times before, we sign off on them or ratify them with the board.”
Helge Baur at Allianz Capital Partners said: “The next bigger evolutionary step will [be] can AI really add value to our decision making? For example, investment colleagues always ask when can we start to use it in terms of data support, as in making decisions, giving us better decisions and so on? That’s a tough one, right? Because we still start with major efficiency gains. And then we need to see, how can we do it? Especially because we also have a need to look into whether it’s a data set, how does that influence decisions and so on? So from a regulatory perspective, which boundaries do we give to AI? And this discussion needs to take place on different levels. We defined for us that none of the communication to clients, for example, goes out directly out of the machine, everything is reviewed. It’s the same on all sides, it may change in the future. Let’s see.”
Replacing humans?
Some employees express concerns about potential job displacement due to AI, particularly in industries where automation can replace routine tasks. Replacing humans is a complex and frequently debated topic.
In his opinion, Vijay Snakaran at State Street, said “AI is a support function, the way we deliver on these operational about efficiencies is by deploying AI as an agent or an assistant, not as a replacement of the person in the loop.”
Nick Laird at Verumex commented that “it’s a progress question, of course, everyone is afraid or will be replaced, and so on. I think we will simply see a different kind of skill set that we will need going forward. And we see that already in Silicon Valley, [where they’ve started] to hire prompt engineers. So engineers which just think about what is the perfect prompt so that I get close to a 100% [reliable and usable] answer from my machine. I think that’s the way it will go going forward, that we just [think about] how can I get the maximum out of the machine? And yes, we will never replace human beings in terms of that they will act based on” data and analysis produced by an AI tool.
Future expectations
Baur said: “I would phrase it rather general so everything that is very text and language intensive. So these are use cases which we try to look at this might be board meetings, taking minutes, summarising it and so on.” A preliminary screening of legal documents was another example, as was using ChatGPT to help staff prioritise which files to handle first.
Lee wrapped up the panel saying, “Interestingly enough, I did check ChatGPT with your questions, and it didn’t come up with half the answers you came up there with, [showing] there’s still room for human interaction.”