How investing in AI can power the future of private equity

Welcome to this insightful episode on the role of artificial intelligence (AI) in transforming private equity. Join Luke Pais, EY Asia-Pacific Private Equity Leader, along with guests Petter Weiderholm, Head of Digital for Asia-Pacific at EQT, and Vikram Khanna, EY-Parthenon Asean Digital Leader, as they explore how AI is revolutionizing deal sourcing, portfolio management and risk assessment, giving firms an edge in today’s competitive market.

Key takeaways: 

• There is a shift in private equity towards data-driven approaches through AI integration, enhancing operational efficiency and decision-making. 
• AI optimizes investment sourcing and due diligence by expediting research and assessing large data volumes. 
• Responsible AI implementation is important to address challenges like output unpredictability and the need for human oversight while focusing on strategic execution to maximize AI's potential.

Luke Pais (Luke):

Welcome to Money Multiple. Artificial intelligence (AI) is transforming private equity (PE) by enhancing how firms analyze, invest and manage their portfolios. Traditionally reliant on data and intuition, private equity firms are now leveraging AI to process vast data sets, identify trends and predict outcomes with speed and precision. 

AI-driven due diligence enables faster, more comprehensive assessments of potential investments by analyzing financials, operational data and market trends, supporting better, quicker decisions. In managing portfolio companies, AI helps optimize operations by spotting inefficiencies and forecasting demand patterns, boosting performance and value creation. Additionally, AI is reshaping risk management and ESG evaluation, giving firms a sharper view on sustainability factors. That’s the hypothesis. 

To help us unpack where we are on this journey, I’m joined by Petter Weiderholm, head of EQT Digital for Asia-Pacific, and Vikram Khanna, EY-Parthenon’s Asean Digital Leader. Petter, Vikram, welcome to the program.

Petter Weiderholm (Petter): 

Thank you very much. 

Luke:

Petter, as a starting point, you should probably give a little bit of a background on EQT and what you do specifically within EQT.

Petter:

Sure, happy to do so. So, EQT is a private equity business started in Sweden and comes from the Wallenberg family. This is important to understand. The Wallenberg family is the oldest industrial family in Europe, and the family business has been around since the 1850s. The family business is currently on the sixth generation. And what they have been doing, I joke and say that they own everything in Sweden except Spotify and Ikea, and it’s sort of true. They own companies like Ericsson, Electrolux, AstraZeneca, ABB and Nasdaq. It’s really incredible what they’ve been able to do. 
 
They started EQT together with the founder, Conni Jonsson, who worked at Investor, to become the transformative private equity business that used industrialists from their network. Instead of having financial engineers on the board, they said, “We can help future-proof these companies by using our industrial network,” and they did so. And that has been a success story. We’re now celebrating 30 years. It was founded in 1994, and we have now grown to be the third largest private equity with something like US$250 billion in asset under management. And in 2022, we merged with BPEA (Barings’ Private Equity) in Asia. So we are now truly global. 
 
And what I do specifically here is working with the center of excellence, which is called the EQT Digital. So, on the investment side, we support all our deals and our funds’ portfolio companies with digital transformation. And this is from the Wallenberg view and culture of making sure to future-proof our companies. We believe that to future-proof a company, you need to be a digital leader in your industry and your market. So, that’s the background here. And that’s also why AI is so interesting, because how do you future-proof your company if you have no idea about how AI will impact you? 

Luke:

Thanks. That’s fascinating. It’s more than 150 years of evolution. Vikram, do you want to talk a little bit about your role at EY Digital? 

Vikram Khanna (Vikram):

Sure. It’s a real pleasure being here today, speaking on this topic that is becoming super relevant for the PE industry. So, I’m a partner at EY-Parthenon. I work with clients actually across industries and focus on driving value through digital innovation and increasingly AI, which cuts across digital and innovation.

So, the reason this topic is so interesting is we did a survey earlier this year. EY runs a survey called PE Pulse, and it showed that 75% of PE-backed companies agreed that they must act now on GenAI to avoid giving competitors a strategic edge. So, this has translated into a number of conversations. And this cuts across general partners (GPs), limited partners (LPs), and portfolio companies (PortCos). Almost all of them are engaged in active initiatives or discussions on this topic.

Luke: 

Thanks, Vikram. So, maybe, Petter, let me start with you. I think looking at the topic of AI and generative AI (GenAI) from a PE lens, it’s a little bit different to a traditional corporate because you have the firm level where you’re actively raising funds and investing and managing a whole bunch of investors. And then you have the portfolio level where you’re trying to drive value in the company. So, maybe I’ll start with an overarching question, which is, as you look at the state of AI and GenAI and as we stand, how do you see this impacting the private equity space in the near to medium term?

Petter:

I think there are differences depending on the time horizon you say. But, when it comes to how we use it, and I think that this is important to understand that we started to build out EQT tech and our own data play, which we call the Motherbrain, already in 2016. Then, I joined the company as the chief information officer (CIO) and was responsible to drive our internal deal transformation together with the chief data officer (CDO). At EQT then, this was the first time that we started to use big data in our core process, which was the lifecycle process and the sourcing aspect. And it was used for ventures to see what potential new companies should we go out and meet and potentially invest in.

If you think about it, that’s a needle-in-a-haystack problem. Maybe you have 5 or 10 investment professionals in your venture firm, and every week there are 150 companies starting that might be relevant. So, which ones should we spend time with? You can’t go and meet them all and assess all of them. And what type of signals can you actually use to assess that? And it turns out that you can use a lot of data from the internet, whether it’s website traffic, LinkedIn profiles or all sorts of data, and try to assess if this is a company that probably is worth spending time with. 
 
So, we started to invest in that, and that turned out to be quite successful. And what we’ve done now internally is expand on this so now all our business lines use this for assessing and driving our deal management process really. Of course, there’s a difference between business lines. If you work with larger funds and you buy billion-dollar companies, you know what companies are your targets that you have other types of use cases.

But what has happened over these years is that this has become part of the culture and how we think it’s natural for us to use data. And I think that’s one of those sort of maturity problems that you will see with adopting AI because it’s not something that you just throw money at. This is a part of how you organize yourself, your culture, and the way you work internally.

If you look 10 years into the future, what does a well-run private equity company look like? And if you don’t have any sort of usage of AI or any sort of usage of data, of course, you’re not going to exist. You’re going to make a lot of mistakes. You’re not going to be there to compete where you need to compete. So, I think that has a huge impact long term to actually invest in that now. I think we’re lucky that we started early because we’re doing that pretty well.

Luke Pais:

Yeah, that’s really interesting. And I think in many ways, EQT may be further ahead of some other PE companies. As I said earlier, what we’re seeing is whether it’s LPs or GPs, at both ends of the spectrum, we are seeing this need and desire to use data more effectively. 20% of our tasks —our human tasks — can be done a thousand times better by AI, and those are the areas we should focus in. And I think this wealth of data and different forms of data is an ideal place to attack that problem. And I think we’ll see a lot more of that.

So I’d like to expand a little bit more on the investment selection. I think private equity has always been a data-driven game, but with a lot of judgment, intuition and vision applied to that, how is AI changing that as you look for new investments? That’s the first part of it. And then as you evaluate these investments that you identify, I guess today a lot of your time and judgment is also spent on whether AI will ultimately impact or disrupt or maybe even enhance some of these businesses. Can you expand a little bit about how you go about it in EQT?

Petter: 

When it comes to the role AI plays early on in the sourcing, I think number one is the efficiency that you get now. First of all, we’re a thematic investor. It’s not like we are opportunistically going after anything. Rather, we invest in health care, technology and service businesses, and various sorts and subsectors there. And typically, what you need to do is to map out a sector and a subsector, and start to understand what are the actionable targets that fit into your hypothesis of where the world is going. In order to do that, before AI, you can imagine the amount of manual research that is required, and you can buy this research from advisors, or you get from investment banks, or you do yourself with a team that start to work at that.

What happens now is that with the help of AI and our own ability to adopt and use these things, combine the existing ChatGPT and so on with the data that we have internally and the experience we have, and combine that so that you much quicker can assess a situation. You can assess a subsector much quicker and much faster. And I think that helps gain conviction. It helps us finding the needle-in-the-haystack also. That’s one of the areas I see.

And when it comes to performing the actual due diligence, I think that’s a separate topic we can talk a little bit about. Maybe, Vikram, you want to chime in on this one?

Vikram:

Absolutely. I agree. What we’ve seen previously is this heavy reliance on manual data analysis, financial modeling, expertise of investment teams to make decisions, and that’s still very relevant. But I think instead of sifting through the spreadsheets with the introduction of AI, this landscape is changing fast. We have natural language processing, which always existed in machine learning. But with the push combining that with GenAI, we are seeing the ability to screen deals much faster because you can take structured and unstructured data, and at least form an initial view and be made aware of initial risks.

Also from a sector perspective, we are also starting to consider how a company could be disrupted by AI and what opportunities exist. And this applies across sectors.

Petter:

Yeah. I think one thing here on performing due diligence (DD) is the learning that we’ve done from our side going out to advisor and asking, we are much faster getting to a much bigger depth and can ask more pointed questions, which impacts the quality and the type of work that the advisors are doing, whether it’s looking at what’s the time of this market or making product assessments or whatever it is that we ask advisors to do. We will have come much further into that process on our side before going. And I think that puts different requirements on the devices we use.

Luke:

And Petter, maybe just to follow up on that in terms of as you’re evaluating deals at an early stage, today as you’re presenting these deals to your investment committee (IC) for initial approval, is the potential disruption to the industry or to the business by AI a big factor now in that decision, and have you had instances, for example, where you might have walked away from a deal because you see that playing out?

Petter:

Yeah, yeah, absolutely. The reason we do DD is to figure out the right companies to buy, but also to figure out what not to buy. And it’s important to separate between the sector and the specific company. And if a company is in a specific sector, say in the health care sector and it’s a medical technology type of business, you can’t change [those attributes]. The specific sector has inherent risks and opportunities and regulatory regimens and all those things.

What we can control is, how do we respond to that? How do we deploy capital to build AI capabilities, and what is then important to look for in that company? What are the key things we need to focus on during a DD to address that? So, that is very important for us. So, we build a view of different sectors, like which ones likely have positive tailwinds and which are going to be challenged.

Then have we walked away from deals? Yes, absolutely. And we walked away from deals late because of too much uncertainty on valuation that you need to perhaps change business model or that there will be disruptions between the value chains, people going upstream, downstream and that sort of movement. But yeah, that’s how we think about it. It’s an absolute key question for all the investment committee decisions.

Luke:

Thanks, Petter. Both of you actually talked about the diligence process and how AI is being incorporated into the diligence process. Can you unpack for us a bit? Diligence obviously is a—as you get to detail diligence on a company—a very time-bound exercise. There’s always typically a limitation on information that you have access to. Keeping all of that in mind, and you probably also commissioned many different streams of diligence, how has AI impacted or enhanced this, and what are some of the practices that you’ve currently incorporated into this process?

Petter:

So for one, it defines to an extent, what is it that we need to look for in the due diligence? What’s the scope of the due diligence? As an example, if you believe there are tailwinds in this industry and you look at this particular company now, it’s very important to spend time understanding the data sets that the company has that potentially may be useful. What are the rights to use those data sets to improve their products or services or so on? Then what’s the capability in the company? Do they have the technical platforms, landscape and talent to address this, or do we need to build that and deploy capital into building out those capabilities?

When it comes to the actual process, I think to a large extent, it is used for individual productivity. My life typically looks like this: “Hey, here’s a 250-page consultancy report for this company. What do you think?” And the answer to that question—of course, you can read 250 pages and try to come up with that, or you can use tools to summarize, generate what to focus on, which of course is helpful.

It also introduces an elemental risk. What if you miss something? But the thing is you’ve got to trust the process of the full diligence and eventually find the way. You got to extract the key questions. And where I feel pretty comfortable with using large language models (LLMs)—they are fantastic at helping you ask the right questions if you ask them to generate the answer. And you can have hallucinations, reliability problem, robustness problems and so on, right? But using these tools for us to understand what are the key strategic questions for this company, I think that’s something that you can use across from digital work streams or other work streams in a typical DD.

Vikram:

Super interesting points there. I think from our perspective as advisors on due diligence, the expectations have changed, and you alluded to this, because there’s a first line of diligence you can do yourselves pretty quickly. One thing that we are becoming aware of, and it’s very clear that we internally need to use these tools to get to that first answer very quickly using proprietary datasets and external data, and then dive in deeper; the ability to get to a deeper diligence is now much more possible with these tools. If you think about it, just like during the diligence process, you’d have a bunch of documents and even financial models and legal documents. These tools are really helpful in showing red flags in forecasting. And making full use of that allows the analysts and the teams working on it to get to the insights faster and have more of those insights. We recognize that. And we are investing quite heavily internally in actually building up tools that allow us to do this.

Luke:

So, from what I’m hearing, today the major use case of the application is around significantly enhancing the quality and efficiency of the existing workflows versus changing the way diligence is necessarily approached. Would that be a fair summary?

Petter:

I think it’s also changed what is possible to do. So, for those who haven’t been part of these type of processes, typically you have a couple of weeks to assess the company, and you have a limited amount of questions you can ask, and limited amount of interaction with the management team. So, you can’t analyze as much as you want. But what is now possible, and this is not just only due to AI, but due to tech enablement. For example, understanding geospatial data relating to a particular company. Maybe we look at the buying a business. We own veterinary businesses, we own fitness center businesses, these sorts of things. You need to understand the penetration of market, where are the competitors from a geographic perspective, as an example. So, this is now possible to do within a short period of time and with limited data sets, and data sets are available to buy. And similar things around price comparisons or understanding sentiment of employees, getting data around that, or sentiment of customers on a per-site basis or per product line. That gives you much, much more insight than what a typical meeting with a management team. It really helps you understand what are the key questions we need to dive deep on, where are the softness in the management team’s plan, and where are the things that we actually see more potential than a management team, because we actually now have a information advantage based on the tools and the capabilities we built.

So, that is part of that. You can call AI, you can call advanced analytics. As an example, we use LLMs quite frequently to scrape data in websites in other languages, and you get the translator and you get a good enough understanding of what’s going on. You can do analysis based on that. These are new capabilities that technology, broadly speaking, has brought to the table from a DD perspective.

Luke:

So better questions, of course, productivity and efficiency gains, but then also far deeper insights and maybe conclusions that you would not normally otherwise have had access to.

Petter:

Yeah. Or conclusions that would have taken too long time to do manually. Of course, you can ask someone to Google and do all these things, and it takes them two weeks and probably is mind-numbing, boring. And now we can automate that and get it done in a matter of hours, which I think gives you much more data, much better insights.

Luke:

Thanks, Petter. So, I wanted to shift back a little bit now to the period when you own the asset. So, the deal is done, you’ve taken control of the company, private equity has done a good job of significantly enhancing the value of an asset over its holding lifecycle, typically five years or maybe a little longer than that. Can you talk a little bit about how GenAI is impacting the way you approach value creation in your portfolio companies?

Petter:

First of all, the way we work with this is not that AI is like a separate topic from everything else. When we think about this, we look actually about the full digital maturity of a company, about their strategy—do they fundamentally understand how their customers use the digital touchpoints, what jobs do they do for them?

As an example on a strategy side, how are they using digital for the product and service for their operations, for their marketing and sales, and for the business model pricing? So, that’s where we make money, to save money. In order to do that, you need enablers in terms of your people, organization, ways of working, partners, technology as a cornerstone, or centerpiece in the way we think about it. Do you have the infrastructure in place? Do you have the business applications, your automation capabilities, and so on? Cybersecurity is critical here. And then we also specifically address the advanced analytics and AI capabilities. So, do you have data governance? Do you have the data platforms? Do you have the advanced analytics and AI capabilities in your team to do this?

So, when we look at a value creation plan for a company, and we tend to be fairly successfully increasing the revenue of our companies, we’re not that great on operational efficiency typically. So, we look at that—what are we doing here? What can we do in terms of, for example, dynamic pricing? What can we do in terms of how we use digital marketing? And what data do we need for this, and what data platforms do we need for that? So, that’s one area. How are we changing our product to actually be able to sell more, go into a new market or a new segment, something like that?

And then we need to make sure that we have the right enablers in place to execute this AI because it’s very easy, specifically from the outside. You buy companies, you need to be very humble about what you know and what you don’t know. So, you need to partner with the management team and figure out what can we do here? And the problem with some of the GenAI things is that it’s very, very easy to generate the pilot. You get some data, clean that up and you do a pilot, and it looks really good. The step from taking that from pilot to in-production, that is very hard.

Typical things has been, maybe before AI, okay, can we build a website, build an minimum viable product (MVP) and go live with them? There’s a little bit of complexities around that. It takes some time. But the step from being an MVP to being full live is not that hard because you’ve crossed a lot of the technology challenges and risks of doing the MVP.

But with AI, that’s sort of the reverse. It’s very easy to make a very sexy pilot, but it’s very hard to make that go live. So, how do we work with this? We think there’s three main buckets here. One is the internal efficiency: and this is actually fairly accessible to almost all companies. What can we do better? Like we already discussed here, I use it for my personal efficiency to help summarize and read large reports, right? And many can do this as well.

Then you have, what can we do in terms of better serve our customer needs? Can we automate something? Can we do something there? This is typically easy to do a pilot for and hard to go live at scale because of some of the challenge with that.

And then you have, of course, the one which is the new innovation. Can we do something brand new here based on AI, which is typically what we put in our upside case?

So, those are the top three buckets. The first one we can bet on. Second one we’re willing to invest in and learn. And this is the approach that we’re taking now because, at this early stage of AI, the right strategy, since there’s so many uncertainties around this, how will it develop, what will it do, and so on—the right approach in those situations is to learn as fast as possible. So, we have an AI challenge in our portfolio companies, and then they’re going to pitch back to us, and we will select the winner of the best AI use case, and they will win glory, honor, and probably something more to promote that sort of fast learning. And that is for that sort of second bucket that I talked about.

Luke:

Vikram, you work with a lot of the large conglomerates in the region. So, is it a similar approach? Is it a different approach that you typically see?

Vikram:

I think the approaches to value creation, whether a PE-owned company or a large conglomerate, are not that different. The big thing here is to use it responsibly -- to have certain capabilities -- to address risks. To be clear about that, there’s obviously a huge amount of change management required, but there’s really opportunities, as Petter said, that come up. The strategy on how to invest in AI needs to be clear. There are clearly opportunities where we’ve seen companies, both PE-owned companies and other companies, where a new market or a complete repositioning of the company is possible just by what AI can do in terms of automation, in terms of speed, and the ability then to address a wider market and a different customer pool. So, even during the diligence process to uncover some of those opportunities is interesting and then to follow through.

The others, obviously internal efficiencies, you’ve got to be clear about on what a company should or should not invest in. Because this technology is moving so fast, there’s a risk of obsolescence. Like if you invest in one LLM and something else comes along or very quickly makes it obsolete, you’ve got to think about that. So, if a company were to think about writing parts of an ERP, a lot of these companies are already building in AI capabilities into their packages, so that’s not recommended. But if there are other areas that provide you a real advantage, then absolutely lower cost by a significant degree, then it makes a lot of sense.

The stage we’re at is there’s a lot of experimentation. We’ll probably start seeing production cases. And as Petter said, moving from pilot to production is a completely different ballgame. And we’ve witnessed that and have been part of that journey. You have to think about scale. You have to think about price. And all of those considerations need to be taken account of. And we’re starting to see this. And over the next few months, I think we’ll see a lot more cases going into production.

Luke:

Thanks, Vikram. So, I also wanted to talk a little bit about the risks and the guardrails that you put in place as you do this because, clearly, technology is developing at a very rapid rate. On the other hand, when we look at the talent within the corporations, you get a whole spectrum of talent. Many of them may not be geared up to embrace this in a full way or to understand it and to use it effectively. Can you talk a little bit about some of the risks that you see and how do you essentially put in guardrails to make sure you get the best, but manage the downside?

Petter:

I think this is really interesting to talk about, actually. One thing that we do is we go through our entire portfolio. As I said, we have a few defined sectors that we invest in. So, we build out our view on a specific sector risk and opportunities. Then you have to look at each specific company; what is their potential in there, and we’d had a pretty interesting model. We think about businesses that are more with atoms than with electrons are a bit less likely for downside risk. If you are a manufacturing business or working with atoms and real things or an energy company, we have an infrastructure fund. So, on number of energy businesses, as an example, energy transition businesses, the role of AI is probably not going to disrupt your key business, it’s going to be helpful.

If you’re in the electron business—say in the internet media—there is likely much more risk for disruptions, changes in value chains, new innovations, someone coming and doing what you do at a much, much lower price. So, that’s sort of one way of thinking about it.

But you have a lot of opportunities, of course. To do this, we talked about before on productivity improvements, for example, for software engineers, what’s the opportunity to go after? And I think there’s a lot of those type of opportunities that if you do it right. If you do it wrong or ignore it, then it quickly turns from opportunity to risk with that sort of foundational philosophy that we have at the firm to future-proof our business. And we believe you need to be a digital leader in your industry and market. So you need to have that adoption of these new technologies. And that is a key risk.

Even if it’s possible to do, how do you get an organization to actually adopt a complete new way of working, going from manual to tech-enabled? That is hard. That was hard in the ’90s. That was hard in the 2000s, 2010s, and so on. Whether it was cloud transformations or mobile adoption, or now AI adoption, it is something that requires much more than technology. It requires governance, change management and these sorts of things. That, at least, is how I think about it.

Vikram:

I think companies we are seeing are going to take two approaches, and I think you need to do both. One is rapid, and the other is more calibrated. And we are seeing this even with the very large companies. They’re starting to put out new models, put out new applications very quickly, and then caveat it very heavily.

And I think the important part here is we still rely on judgment from humans. As long as the control lies within humans, I think it’s something that you can move at pace. On things that are more complicated, you can be much more calibrated. Obviously, the regulations and policies are moving on this too. So, I think we are still aware of from where agents are going to be doing things automatically with no human in the loop. And we were talking earlier for the more creative type of work, to make things really engaging and fun, you still absolutely need humans. And I think that’s going to evolve over time. So, I think the general message here is the ability to experiment and do things rapidly and to learn quickly is going to be important.

Petter:

Yeah, I think in terms of guardrails and issues here, LLMs, as they are very impressive, but they also have these limitations on hallucinations. There was actually a great report coming out just a week ago or so by some researchers where they changed small parts of data and the input prompt, and that gave a large variation on output from correct to incorrect because they added some minor variations of wording and so on. And of course, that I don’t see a near future where you have that sort of fully autonomous things that is just based on LLMs. I think LLMs is probably one key architecture element, but there needs to be a number of other things in place.

So, humans in the loop, end of the day. There was this airline that released an LLM and that LLM came up with a new cancellation policy on its own, that cost that company lots of money. And these are risks that you put yourself to if you don’t think this through properly. So, I think having some real serious thought about what do you do internally, what’s okay to fail with, where are errors, not catastrophic or really bad, and where are they really, and trying to understand that to mitigate those risks with these new elements. Because historically, you look at computing, and the thing that has one key quality of it has been—it is predictable, it’s repeatable, it’s robust. You write an algorithm, and you will have it do what it says. A problem is if you have a bug, it’s probably a poor implementation or a poor understanding of constraints or poor description of requirements. But here it’s not. It’s something else that doesn’t have the same thing that we expect from a computer system. So, I think humans in the loop for quite some time going forward.

Luke:

Thanks for that. I like the way you framed it in terms of atoms versus electrons.

Petter:

I think it is pretty helpful, right?

Luke:

Yes, it’s a good way to look at it. And just on that note, as you exit businesses today, are you seeing companies that have AI embedded in some aspects of their operations? Are you seeing that potentially being a value enhancer in terms of exit multiples? Are we at that point yet or is this a little too early in the journey?

Petter:

I think it’s impossible to single out something like that. Generally speaking, high-growing, fast-growing, high-margin companies with market tailwinds will have a higher multiple. How do you get to that? Well, part of that is, of course, causality. The ones that are using AI—will they get the higher multiple—or will the ones that are actually really good companies the ones that are going to adopt it? And why they are adopting it because they are good. It’s not they are good because they adopted it. So, chicken and the egg for that. I think it’s hard to say something about that.

Luke:

Is that though a very important part of the storytelling around the business?

Petter:

No, I think if you look at what has generated returns for us specifically, we are an investor in good companies and make them great with typically revenue generation, help them sell more. That’s the absolutely largest part of our value creation. Then you have operational efficiency, which is another part. And then of course you have multiple expansion. But the thing is you don’t really control that, do you? Because that is the market. The market has typically the land for the multiples.

We will see over time, I’m sure, but I think doing this attribution of a specific adoption of a specific thing, I think that’s likely a mistake because you need to understand the causality. And it’s a hard problem to single out.

Luke:

Yep. But I think that’s absolutely a fair point. So, on that note, thank you so much. It’s been a great conversation. Maybe before I let you go, I can ask each of you to make one final comment on how you see the impact of AI, GenAI in the next few years ahead. Petter, you first.

Petter:

I think in the short term, in industries that where it’s unclear, where’s the impact going? You’ll have uncertainty, and uncertainty is not great for investors. You want to be able to predict how things are going to go. So, that’s for certain industries within the electron space. If I put it that way, I think it will be interesting to see where things play out.

This is really, really exciting times because now we’re all learning and we’ll see a number of massive successes going from pilot to going live with really, really groundbreaking stuff. And I think we’ll also probably see a couple of failures here because that’s expected. This is innovation that we’re doing at the moment—of course, not everything works out. So, I think in the next couple of years, things will clarify much more where things are going and how we can use it, what works in different sectors.

If I look at the PE industry, I think this is also, to a very large extent, the ability of the firm to actually use this technology and lead the way and show, “Hey, this is how we use this to be as efficient as we can. What are you doing there, portfolio companies?” And leading that way is what we should continue to do. If you’re a PE that doesn’t do it, you’ll have trouble down the road.

Luke:

Vikram?

Vikram:

I can’t predict where this will go, but I think it’s safe to say that we will see the level of intelligence embedded in most things that we do. And the hope is that the things that are tedious and very manual to do, computers can do way better than us, AI can do way better than us, that’s what they’ll focus on, and we’ll be able to think about things more broadly. Because ultimately, even if you look at LLMs, the creative output, and I’ve seen a lot of images being generated and other things, but it relies on what was generated before. It’s a variation of that. So, I think this will provide an opportunity for us to think even more creatively, hopefully.

I think for PE companies and portfolio companies, we will see new levels of efficiency. And as Petter said, I think there’ll be successes, there will be failures. But the key is the intersection of AI with other technologies like robotics and other things which are coming to market at the same time makes this very powerful, and so we’ll see where that goes.

Luke:

Petter, Vikram, thank you very much for joining us.

Petter:

Thank you. 

Vikram:

Thank you. 

 

You have been listening to Money Multiple. If you like what you’ve heard, subscribe to our show on Apple Podcasts, Spotify or Amazon Music.