Game Changers in Generative AI w/ Hg’s David Toms

Previously on Dry Powder, David Toms, the head of research at HG, took us inside of their early experiments with Gen.A.I. It was one of the defining moments where you realized that everything you thought a computer could do had just changed. Today on the show, we'll see how Gen.A.I. could be a game changer for HG's business model. I'll ask David how they're fundamentally rethinking the way they price and package software. We had one of the largest software as a service companies in the world, a C-suite speaker from there who stood up and said, "Per user per month pricing is dead." Software has always been priced on a usage model, you know, how many seats and how intensively you're using it. But if we can actually measure the value that a customer is going to get from the software, you can price it really differently so that the user gets a huge return on investment and they're still willing to pay a very good price for the software. The end game has to be, well, what are we actually doing for you? What are you prepared to pay for us to solve that business problem? What are you prepared to pay for us to ensure you don't have to run another payroll manually ever again? We'll also explore how they scale new products and features across their portfolio and how they spur a friendly competition between management teams. I'm Hugh MacArthur, Chairman of Bains Global Private Equity Practice, and this is Tri-Powder. So let's talk about scaling AI and how does AI fit into your model of investing in some of Europe's leading software businesses. And I think I know what you're going to say here, but I'm going to ask you if you think that AI could really be a game changer when it comes to things like product releases across different countries and all types of interconnectivity in the businesses that you invest in. Yes, it's a game changer on so many levels. Another specific use case we're seeing where it's really quite powerful is refactoring old code. It turns out that because the large language models are trained on the internet and there's a lot of old code out there on the internet, they're pretty good at reading and understanding old code, whether it's COBOL or any other of the sort of more mature languages, if we should call them that, whereas the number of developers for those tools is going down every year as they retire. That DVD is not retiring, OpenAI is not retiring, so refactoring old code, it is very effective at that. And that starts to change some of the dynamics of how you look at your product portfolio and what you do with the product portfolio. And you can look at some products that say, well, those might have been something that would otherwise have been sunset, because if you can refactor them at half the price, then you can move them to a new world. Does that change how you price and package software and is what worked in the pre-Genai world the same as what's going to work in the post-Genai world, if I can use that term? Yes. This is something we devote a lot of thought to. Our software leadership event, this is going back about just over a year now, we had one of the largest software as a service companies in the world, a C-suite speaker from there who stood up and said, per user per month pricing is dead. Not now, not next year, but it's dead in the long run. Soon. What does it get replaced by? Initially, some kind of usage consumption model, but again, the real game changer is when it's results-based payment. So I'm actually paying for the software to solve the real problem. If you look at the history of software for the last 60 years, it's essentially been a battle between software companies trying to convince users of the value and users looking at the risk of implementing the cost of implementing the software. And when you have the old world of perpetual licenses, that was requiring a huge capital investment from the customer for a very uncertain return. So typically large companies would have extremely stringent ROI requirements and very high ROI numbers required because they knew how often the software had been oversold and under-implemented. So the perpetual license model, you could never get full value from the software because there was a massive risk premium. The customer just needs the risk premium, so they can only ever pay 20% of true value or whatever it would be because they don't know it'll work. And then we move to software as a service and subscription pricing, which are wrapped in with it. And you've taken away the implementation pain from the customer and you've also reduced the risk because you can roll it out to 50 people and if they like it, you can roll it out to 5,000. So you've started to move the balance of perceived value towards real value. So the customer now doesn't need such a big discount because there's not such a big risk. But even so, they don't know what the actual value to their people is. They don't get the software in and up and running, but you've told them that your wonderful new payroll package will save them eight people in their HR department who can go and be deployed on training and more interesting things, but they don't really know that until they've got it up and running. And by then, you've already struck a price and agreed a deal. As you move to a usage-based world, you say, well, it doesn't really matter. Here's the software, pay for it as you use it, and the more you use it, the more value you'll get. And you won't mind paying because every time you use it, you're getting value. So that's, I think, where we're moving to, but the end game has to be, well, what are we actually doing for you? What are you prepared to pay for us to solve that business problem? What are you prepared to pay for us to ensure you don't have to run another payroll manually ever again? I love that entire notion of results-based pricing. That just sounds so incredibly compelling, and as the industry moves toward any notion of getting that right, you can just see the benefits and the scaling getting to escape velocity and that little application wave that you described earlier becoming a massive tsunami of value. Yes, but I've got to put a caveat on this. You have to get it right, and this is an area where we do spend a lot of time with the portfolio on pricing and packaging of software because if you get your pricing of this wrong, the risk is you end up giving a lot more value to the customer. I'll use a very live example. There's a large U.S. payroll software company about a year ago released a new product, and the new product was materially better than its old product. Its revenue went down when it released the new product, because what happened is this company was previously charging its users per payroll run or per payslip processed. The thing is, with this old software, there were a few mistakes creeping in, so it wasn't running one payslip per employee per month. It was actually running instead of 1.2, and if you were charging per each of these you process, well, you're making 1.2 of revenue when you should really only be making one. The new software they introduced was dramatically more accurate and gave the employee an opportunity to get in there first, see what the provisional payslip was, make any changes to their hours, their overtime, whatever else. The new version of the software, every payslip was perfect. That meant you went from 1.2 to 1, and you were charging per each of these you were running, so all of a sudden you were running fewer payslips. You had introduced a better customer proposition, and you were making less money from it. Because you're pricing in old terms and not the new terms, you're not pricing to value, you're pricing to volume. Correct. So, getting that right is going to be crucial. So that tsunami of value metaphor, you've got to be really careful that the tsunami's not crashing on you. Yes, you can give it all to the customer if you're not careful. Right. Totally makes sense. So, how do you decide, David, when you've got a new feature that really is ready to scale and you're ready to push it out there? Can you go as fast as you'd like without breaking things like this pricing example that we just talked about? The beauty of operating in Europe and the beauty of operating in really quite defined markets is you get to do the same things again and again with little tweaks. So we're not trying to roll out a feature or a capability to tens of millions of users all at once. In general, across the portfolio, we have the chance to experiment on particular products in particular, GOs, whatever it might be. So, going back to the example I used of changing a chart of accounts to a list of tasks. That was one product in one GO within one of our businesses. It was probably 1% of revenue, maybe even less that we did that on. So, you can see if it works in Finland before you go to Denmark and the rest of the Nordics in Germany and all the other geographies. Exactly. And then we can talk within the portfolio about who's doing what and how does it work. We've got about 50 companies that all do very similar things. The opportunity for one of them to learn from another, you know, I've done this here and it's had this effect. Another one says, well, we've done this over here and it's had a slightly different effect. They can learn how these things are rolling out. Now, David, there's been some discussion about pure play GNI providers and new entrants taking over from systems of record providers. How do you see that balance of power evolving and could there be negative impact somewhere in there? You could put the scenarios on a spectrum here. You've got your sort of pure play GNI scenario, which says you can do everything in the world with generative AI and you won't need any systems of record software. That's kind of the most extreme case. You don't hear many people pushing that case, but you hear a little bit of it. And then at the opposite extreme, you've got the GNI can never do anything involving numbers because it just doesn't get things right often enough. It hallucinates too often. You can't have it hallucinating your payroll. And the answer is somewhere in the middle and it moves around depending on the use case. I have an example I show, which doesn't work so well on a podcast, I'm afraid, but an image produced by generative AI, which is largely a proper image of a bike. But generative AI is filled in the bottom of the wheels and it looks like a great image. If you zoom in enough, you can see that the angle of the spokes changes where GNI took over to fill in the bottom of the wheels. So it looks to you and I like a bike, but you look closely and you realize if you try to build a bike wheel like that, it would collapse because the tension has to be exactly perfect in every one of the spoke and the angles have to be perfect. And it's just not done it like that. But what it can do is you can say, well, I want a bike to look like this. I think what we're going to see from generative AI is a world where it calls up the software. So it says, I need a bike wheel design piece of software or I need some accounting software to solve this problem. And at that point, the systems of record software is still at least as important as it ever was because it's the only thing that can calculate your payroll, your taxes correctly, file them correctly with the Danish government or however it might be. So the system of record software still does that perfectly. The AI could potentially be acting as the interface layer. That means you're improving ease of use. In every situation we have ever seen in the technology world, when you improve ease of use, you increase uptake. Because you get people like me using it once you improve ease of use. But let me ask you one last question that's been on my mind, which is, what do you do to encourage knowledge sharing between the portfolio companies and what types of formats have you found to be most effective in promoting that sharing? We find the single most effective thing is to rely on people's natural desire to help each other coupled with their natural desire to compete. And so if we want to effect change, let's go back a couple of years. Interest rates are rising, the value of cash has changed. All of a sudden, DSOs becomes more important to a software company. You don't want to give your customer longer payment terms because cash now has a value. So we thought to the portfolio, yeah, we need you to re-prioritize DSOs a little bit. They're never bad in a software company. But if you can take five days off across the portfolio, that's good for the cash position. And so what we did was we didn't go along and say, you need to do this with some prescriptive approach to it. We took the data we have on all the portfolio companies and we produced a ranking. And we sent this round to the CEOs and the CFOs and we said, this is the ranking of the whole portfolio. And here's where your company sits in it. We didn't give everybody's name. We didn't show the whole thing to everyone. We said, here's the overall ranking and we told each company where they sat in it. And the first question that they asked from this is, how do I get better? And then the second question they ask is, who do I need to go and talk to to help me get better? And if you're one of those top 10 companies, you're feeling pretty good about yourself and you're, yeah, I'm happy to talk to them. I'm happy to tell them how great I am and I'm happy to tell them what we do is so great in our business. But we find that showing the companies where they are in any form of benchmarking, because we've got 50 very similar businesses, the benchmarks are reasonably appropriate between them all. So you show them where they are in a benchmarking, you tell them where they sit and you give them the names of some people who are good and they solve the problem for you. You don't have to go in there with your big stick beating them and saying, you need to improve your DSOs. You need to do XYZ because you're the guy from the private equity company. You're not the guy who's kind of in the arena fighting the battle. They want to talk to other people in the arena fighting the battle. That's amazing. I love that entire ranking system and I don't know how you resist the temptation to put everybody at slightly below average in order to stimulate the competitive fires to get better. They talk to each other. They'll come back and say, I've just been to eight other companies and we're all below average. It's above average. Can't quite do it that way. That's not the right way to do it anyhow. David, I really want to thank you for coming on the show today. It's been an incredible conversation. I love pragmatic approaches to things like AI where you could kind of talk about theory and fluff for a long period of time and not get anywhere. I certainly learned a tremendous amount. I'm sure our audience has as well about really practical applications of AI in the investment world and what you're doing and how the specific use cases you're developing are really accelerating the overall development and deployment of AI in the companies that you own. So thank you again and I really appreciate your time. Thank you. Thanks a lot for having me. I really appreciate it. Thank you. I'm Hemak Arthur. Thank you for listening.