Alt Goes Mainstream: The Latest on Alternative Investments, WealthTech, & Private Markets
00:06:41 · Yann Magnan
[00:06:41] S1: So the first thing that really has struck my mind back probably like already twenty years ago is that I was a more junior valuation guy. I was tasked to work on the M and A assignment. We had to do evaluation of a business and to do that, we had to develop private business and we had to develop benchmarks to to public companies. I don't remember what the business was, to be honest with you, but I remember private business, we had a peer sample of private companies and it was like five, 6PM at night and we had to speak to the client the next day with some ideas and so on. And so I turned to my junior, I tell him, okay, can you assemble real quick benchmarks of these peer companies, revenue growth and EBITDA margin and various financial ratios and so on, and collect that information from brokers reports and so on. And the guy then, yeah, but it's going to take three days. I mean, come on, why? Why is that? I mean, you have to go through all this paperwork, all these amount of documents and find for relevant data and then plunge this data into a model and then cross check that everything is correct. Yeah. It's going to take two to three days. And I thought, mean, that's nonsense. How can that happen? Of course, that was some time ago when all the data sources like Cap IQ and FactSet and Affinity came to market with a way of looking at that data in an automated way. But still I've been thinking there should be a way for people that have to value these non listed assets that have to take a look as to how they compare to other type of companies to pull in the information in a much faster way, to pull in the intelligence in a much faster way, to go through the vast amount of data and documents that you have to go through to collect and pick this relevant information in a much faster way. And that's where we started thinking with your colleagues. I mean, AI should probably be helpful there, actually. And... Yeah. So that's one of the starting point.
00:08:32 · Michael Sidgmore
[00:08:32] S0: One question that comes up from that is, how much is the lack of uniformity a real challenge? There's lack of uniformity and standardization around how funds may value companies' assets and their valuations, how they look at their revenues, their valuation multiples, their methodologies. There's so much difference in the way that firms may apply their own valuation methodologies. Can you standardize and create uniformity within the valuation process? Or is it automation will get you part of the way there, and then there will still be a large manual component to this?
00:09:07 · Yann Magnan, Michael Sidgmore
[00:09:07] S1: My view is that you cannot automatize that work partially, but only partially. A significant component of any valuation process is actually the human judgment that goes into it. And you would have a different view on certain parts of a business that I would have. It is very important that we have these different views. The valuation process is more like a language that helps you articulate a view on the business and, in the end, form a view on the value of this business. But your human judgment is going to be very important. My human judgment is going to be very important. I don't think that we should automate that part of a process. Automating the calculation, I mean, that's, of course, that's the simple things that you have to do. Automating how you get or facilitating, I would say, other than automating, facilitating how you get to a faster human judgment, how you get faster of information that will help you form a view as to how this business is really comparable to that other one, which by the way, that does not exist. There's not one business that fully compares to another one. So how do you understand the nuances faster so that you can come up faster with the adjustments that you need to make in your valuation processes and in your valuation analysis so that in the end you come up with the proper construct of this solution. [00:10:24] S0: Do you think AI will eventually be smart enough where it will actually do a better job of the qualitative side of what you just described, the human judgment around valuation?
00:10:24 · Michael Sidgmore, Yann Magnan
[00:10:24] S0: Do you think AI will eventually be smart enough where it will actually do a better job of the qualitative side of what you just described, the human judgment around valuation? [00:10:34] S1: Possibly. But I will not bet on that. And I think we are years away from that happening if it happens. The work we are doing at seventy three Strings facilitating the valuation for the owners of the non listed assets. Ultimately, objective of this work, the objective for clients is to report their valuation to their LPs, to their investors. [00:10:58] S0: And do it faster.
00:10:58 · Michael Sidgmore, Yann Magnan
[00:10:58] S0: And do it faster. [00:10:59] S1: And do it faster. So do it faster, technology can help. The issue when you say you want to report something to a third party, and we're talking about money here, and the real money, and very significant money. The issue is that someone has to be accountable for that. You can't just spit out a number. You have auditors that are at play as well that will look into what things are being done. So can AI come up with a very good estimate of the value of an asset? Yes, probably. I would actually be ready to bet that there's already some very interesting hedge funds that are doing that quite effectively. But here is the thing. Again, the the world I live in, the issue is not just about doing the validation, is providing a certain level of comfort to the LPs that it's done properly and that someone is responsible and accountable for that. So the moment that you put an AI algorithm into it that just spits out a number, you're defeating the purpose. So what I'm saying, will AI get there? Potentially, but the key thing for AI to get there will be to find a way that the other part of the equation, the LPs and their auditors and the regulators as well, will be comfortable that the AI is coming up with the right answer. Because if they are not, who are they going to put the blame on if something goes wrong?
00:12:19 · Michael Sidgmore, Yann Magnan
[00:12:19] S0: I think that brings up a really interesting question related both to 73 strings, but also more broadly around technology being built in private markets. You rely on different stakeholders and technology solutions to be able to do the work that you do at 73. Where do you feel, at least right now as a business, your work starts and stops at 73? And then where do you include and partner with other service providers? [00:12:47] S1: Our work starts with the data or the documents that are sent from the portfolio companies, the borrowers to our clients. That's where we're picking up the work. We're picking up these documents. We have developed a set of technology that can read any type of document, PDF, PowerPoint, Excel, email. What am I missing? Well, probably a board deck, a quarterly report, an annual report, whatever. And it can read [00:13:11] S0: someone's mind. [00:13:11] S1: And and [00:13:12] S0: it can read someone's mind. [00:13:13] S1: Maybe at some point. And collect these documents, identify the relevant KPIs, collect the KPIs, standardize this information, create consistency. And then that information is used for evaluation. So from that perspective, what we need is, well, not much. Eventually, we need a Dropbox from the portfolio company, eventually that could accelerate certain things. But most often, we are seeing that the portfolio companies would send these documents by email to the investment teams, and it would be automatically forwarded to our platform, and it's done.