--- title: "Transcript: AI Across Your Portfolio: A Playbook for PE Firms" webinar: "AI Across Your Portfolio: A Playbook for PE Firms" date: 2026-07-01T14:00:00Z duration: "1 hour" speakers: ["Charlie Cowan"] topics: ["Evaluate AI readiness and opportunity across portfolio companies", "The PE plugin: a Claude tool for scoring investments on AI potential", "A playbook for organising AI adoption at portfolio scale", "Why AI adoption is becoming a valuation driver and how to position for exit", "Getting started: run your first portfolio AI assessment this week"] recording: https://youtu.be/yFp9JUXrv5k canonical: https://kowalah.com/resources/wednesday-webinars/ai-across-your-portfolio-a-playbook-for-pe-firms/transcript source: Kowalah Wednesday Webinars --- # Transcript: AI Across Your Portfolio: A Playbook for PE Firms **Recorded** 1 July 2026 **Duration** 1 hour **Speakers** Charlie Cowan ## About this session PE firms have a portfolio-wide AI opportunity. This session covers how to evaluate AI readiness across your investments and organise adoption across 10, 20, or 50 portfolio companies. ## Topics covered - Evaluate AI readiness and opportunity across portfolio companies - The PE plugin: a Claude tool for scoring investments on AI potential - A playbook for organising AI adoption at portfolio scale - Why AI adoption is becoming a valuation driver and how to position for exit - Getting started: run your first portfolio AI assessment this week ## Recording https://youtu.be/yFp9JUXrv5k ## Transcript **Charlie Cowan:** Okay, hello? I am just gonna share my screen, and here we go. We are recording as well, so let me present, and hopefully you can see my full screen! Well, welcome to today's Kowalah Wednesday webinar. You join me from a client site where I'm in one of their phone booths, where we're running a global training program that kicked off this week, here in the UK, Ireland, and the US. But absolutely thrilled to join you for today's Wednesday webinar. If this is your first Kowalah Wednesday webinar, we get together every Wednesday, and we spend an hour talking about something that is top of mind in the world of AI and business, and that might be something that we are learning or seeing internally here at Kowalah. It might be something that we're seeing in our clients, or generally what's going on in the industry. **Charlie Cowan:** And it gives us some time when we can take stock and pause and take a look at what's going on and share stuff that is useful and valuable for you to take back to your business. We will go through our normal flow today. We'll spend probably 30, 35 minutes talking about the main subject, which is AI across a portfolio for private equity companies. And then we'll touch on our regular slot of AI in the News, where we'll spend a little bit of time to look at what's been going on over the last 7 days, and there's been quite a lot, so we've got a bit to cover there. And then there's an opportunity for you that are watching live to ask your questions. **Charlie Cowan:** So we are in a Zoom event, and so what you will find, depending on what device you're on, either at the top or bottom of your screen, you will find both a chat button, where you can chat, and then you can ask questions via that, and then there's also a Q&A button, where you can more formally ask a question, and then I can answer it. Now, as I'm in a WeWork booth today, I have only got the one screen, so I'll do my best to keep an eye on any questions. But if not, my colleague, Caitlin, that is on the line as well, will do a good job at pointing me in the right direction. Now, what I would say is that if you have got questions about anything as we are going through the subject, then feel free to ask. **Charlie Cowan:** Don't feel you need to wait until the Q&A section at the end. And of course, if you're watching this on YouTube, feel free to ask your questions in the comments, and we'll come back to you via that as well. So, to today's main subject, AI across your portfolio, a playbook for private equity firms. We're doing quite a bit of work with either the portfolio companies at private equity firms, or we've been working with a few PE firms themselves directly. And so this has been something that's kind of top of mind for us over the past few months. But, and this was a couple of months ago when I took this screenshot, I thought this would be a useful way to kick it off about why this is such an important topic. **Charlie Cowan:** This was a tweet from someone that I saw on X, formerly known as Twitter, where the guy's just saying, I've just been speaking to a couple of friends who work on the operational side of some large PE firms. And he's saying that every PE firm is now scrambling to find and recruit AI talent that can implement the newest tools into their portfolio companies. I won't read the rest of it, but I do want to point out this section at the end. For the portcos that have already piloted this, the efficiency gains are absolutely huge. Some are well within the range of 20% plus headcount reduction potential for back office functions. **Charlie Cowan:** Now, of course, we're going to talk in a minute about the business model, the value creation model for PE firms, but of course, reducing costs, reducing payroll, and therefore headcount, is a key element of that, and that drives that efficiency, which drives the exit valuation. And we're going to come back to that in a little minute. But I had posted this in a private WhatsApp group that we have for some of our clients. And people piled on, saying this is exactly what they're hearing, either because they are a portfolio company, and the PE firms are driving this, or for people that are connected with PE firms who are validating this. It is a hot topic, and AI directly influences the business model for a PE firm. So let's look at a couple of other data points, just to set the scene here. **Charlie Cowan:** This was from some research from a firm called Accordion, that happened just earlier this year. And this was some research into CFOs in portfolio companies of private equity companies. And the question here was, is AI important to your strategy? And, the answer was 97%, yeah, it's important. AI is really, really important. Very few saying it's not important to our strategy. But then, counteract that, within the same bit of research, are you measuring, and are you seeing any measurable impact from that AI strategy? And only a third were saying that they could actually measure, or had any measurable impact. So 97% say it's very, very important. Two-thirds say we've got no way of measuring or seeing any impact from that. So that's a challenge. Next up, bit of research from Accenture at the end of last year. So, again, focused on portfolio companies in private equity firms. So, do you have an AI strategy? **Charlie Cowan:** So, 60% of those portfolio firms saying, yes, we have got a defined AI strategy. And then the next question, or follow-on question was, and are you tracking the EBIT returns of that AI strategy? I.e., can you tie a dollar or pound, or euro number to that AI strategy? And 85% of them saying, I'm not even tracking it. So we've got a strategy, but we've got no way of linking that back to the actual profitability of the business, the earnings, and therefore, as a private equity firm, the potential exit valuation of the business. So there's obviously a disconnect between the intent and the results that these companies are seeing. So let's tie this into the overall business model of private equity, and this value creation equation. **Charlie Cowan:** So as a private equity firm, we want to increase the valuation of the businesses that we invest in, and we can do that through a number of different ways. So one, obviously, is by increasing the EBITDA of the company. Can we grow the earnings of this business? And that can be through increasing the revenue, so more cash is coming into the business, or through reducing the costs of the business. So, we're going to take out costs in the back office, we're going to take out headcount, and this is going to increase the earnings that sit between those two. Secondly, there's expanding the multiple. So, when we sell this business, we might have bought it at 8 times earnings, and we're going to sell it for 12 times earnings. Well, how would we do that? Well, that might be because we've moved from being a traditional services business to being more tech-enabled. **Charlie Cowan:** It might be we're a traditional manufacturing, we've been able to inject some element of technology into that business, so we can expand the multiple that we see. And then the third can be by paying down some of that debt that may have been applied to the business and loaded up onto it when the PE firm made that investment, that acquisition. So can we take some of the operating cash flow and use that to pay down the debt, so that when we do sell, more of that money goes back to equity and less to pay off the debt? And so what's interesting about this value creation equation and these three elements is that AI can help across all three of these. And that's why so many PE firms are leaning into this. We can drive up the revenues, we're going to talk about this in a minute. We can drive down some of the costs. **Charlie Cowan:** We can evolve the actual business model of the portco, so we can expand the multiple. And then we can use some of those additional earnings that are coming through to pay down some of the debt. So it is a win-win-win for the PE firm over the life cycle of the investment. And because all of those three things are tied together, we really are in the situation where every single pound of EBITDA can amplify that outcome. Because if that pound is coming through AI-enabled products and services, it's going to drive across all three of those levers. And so that's what we're going to spend a little bit of time looking at now, is how do we get these pounds, hold these pounds hostage so that we can accelerate that outcome for the PE firm? **Charlie Cowan:** Now, this graphic is one of my favorites, I don't know when it was created, I came across it two years ago, and we reference it in so many of our client engagements. The Gartner AI Opportunity Radar. And I'm just going to orientate you around it, because we're going to spend a bit of time talking about it over the next 10 minutes or so. So, with the radar, it has got two axes. So you've got the horizontal axis, which goes from left to right. So we're looking on the left at everyday AI use cases. So just the AI use cases that people are doing every single day. I'm writing a letter, I'm drafting something, I'm writing up some minutes, these kinds of things. On the right-hand side, you've got game-changing AI. **Charlie Cowan:** These are the things that are going to change the direction of your firm, maybe change the direction of your entire industry in terms of the way that you're applying AI. So that's your horizontal axis. Then you've got your vertical axis, and here you're going from the top, which is external, customer-facing AI use cases. So here we're thinking about external stakeholders. So, if you're the PE firm, this could be your portfolio companies, but if you are that portfolio company, then here we're thinking about our suppliers, we're thinking about our customers, we're thinking about candidates that want to join us, we're thinking about other investors, so it's all the people that are outside of our company. And then, down at the bottom end of the y-axis, you've got your internal operations. So these are your internal employees. And what this does really nicely is craft these four quadrants. **Charlie Cowan:** So, I'll just very quickly go around them, and then we're going to go into them in more detail. So, bottom left is everyday AI and internal, and so here we think about our back office. So, this is our internal functions, HR, finance, IT, ops, admin, these kinds of teams. At the top left, we've got the same, it's everyday AI, but now we're external customer-facing, so we're thinking of our sales and marketing teams, anyone in go-to-market, could be customer service, customer success, however you phrase that, anyone that's interacting with your customers. Now, bottom right is the guts of your business, the core operations, the core capabilities. And here I often think about Amazon. You order something on Amazon, it arrives the next day through their wonderful logistics network, it's not the right thing, so you take it down to the garage, you open up the cabinet, you put it in, you get automatically refunded. **Charlie Cowan:** Their whole end-to-end, vertically integrated sort of supply chain and operations, that is what sits in core capabilities. And then the top right is game-changing AI, but external, customer-facing. And here, we think about new products and services that are only available or possible because AI exists. So, are there specific data or markets that you can sell in? We're going to charge more because we've got AI injected into our business. So, this is the Gartner AI Opportunity Radar. Now, what you can do is just draw a diagonal line from the top left to the bottom right, and you can roughly say that anything that is on the bottom left of that line tends to be quite defensive. So this is about saving money. This is about efficiency. It's about productivity. Do we need the same number of people? **Charlie Cowan:** In that tweet that I showed you up at the start there about the guy saying that portfolio companies were seeing a 20% efficiency gains or reduction in headcount, this is on the defensive side of the line. But over on the top right, you could call this offensive or proactive, value-add, something like that. This is the opposite of defensive. This is creating stuff that is new and does not exist today, and is more value-add, rather than taking cost out of the business. And we're going to talk about this in a little bit more detail, but, what I really want to guide you towards is thinking more about the top right than just the top left. Because when you start to think about that exit multiple, it is the top right that is going to drive that multiple. The bottom left is there just really to help with the EBITDA. **Charlie Cowan:** You might save some costs in the short term, but it's the top right that is really going to drive the exit valuation and the multiple. So let's go into a bit more detail about the back office here. So we talked about finance, we talked about HR, legal, IT, these kinds of teams. And here are some of the questions that you might want to be asking with your portfolio companies. So, one is, do we need the same number of people at our current revenues? So, if our business is doing 500 million of revenue this year, do we need the same number of people to achieve that next year? And again, going back to that tweet at the start, this is where you might say, well, actually, we think we can do with 20% less of these people to be able to deliver the same amount of work. **Charlie Cowan:** A different question, which gets to the same point, is, well, can we grow while still maintaining the same number of people? So, if the goal here is to go from $500 million to $700 million, well, you might have had a people plan, a payroll plan, that means we've got to add on 20% more people in finance, HR, legal, in order to deal with all of this new revenue that's coming through. And I think this is a more interesting question, because this is more positive, and it gets us towards that top right, where we're thinking about how do we grow this business, but using the same group of people that we've got. Other questions that you might be asking is, now when I look across our office, when I look across our processes, when I look across our people, what jobs here really need a heartbeat? What jobs here really need a brain and a hand? **Charlie Cowan:** What needs the body? Or what is really a process, a skill, a workflow that we could build into an agent that our other humans can use? And as a PE firm, once we have seen that in one of our portfolio companies, what of these agents or skills can we templatize and share across the portfolio? Maybe there is an accounts payable agent that we have built that could be used across every single firm. Now, there's a phrase that you might have come across, and I think this is going to be really important. Whether you're in a portfolio company, whether you're in a PE firm, whether you're in any organization, is that there are three types of people in work. There are builders, there are sellers, and there are measurers. So the builders are the people that build your product. **Charlie Cowan:** So, if you are a product company, they are genuinely the people that are building the product, the people that are manufacturing, the people that are writing the code, they are the builders. Then you've got the sellers, the people that are out there speaking to customers, and they are selling this product. Now, that doesn't have to mean I'm a salesperson, I might be in marketing, but I'm there getting this product sold. And then there's a third category, which is the measurers. We don't build, we don't sell. We are here, the back office functions, to measure this thing. And I think the harsh reality is that measurers are going to be the ones that are going to see significant job disruption, because where we had financial analysts, we had IT operations, we had HR business partners. These are roles that are measurers. **Charlie Cowan:** They're not building, they're not selling, and these are the ones that I think are ripe for being seen as replaced by agents or skills. So, when you look across your teams, you might be thinking about that concept of builders, sellers, or measurers. If you are one of the people in one of these roles, I would be thinking, how do I evolve my role so that I'm considered a builder or a seller, and not just a measurer? So now let's go to the top left quadrant. So this is the front office. So we're still talking about everyday AI, and we're talking about external. So sales, marketing, customer support. So the very same questions, which I'm not going to go through again, these top four, the same number of people, can we go further without hiring more people, at what jobs, and so on. **Charlie Cowan:** But then these three other questions that I think are useful for you to talk about with your portfolio companies. So, given that we've got AI, what new channels can we open up? So I'm thinking here about how we might interact with customers in a different way, because we've got AI tools. How can we attract customers in a better way? Let's not just think about AI for ourselves, but think about our customers and prospects who are also using AI. They're sat in front of their own Claude, they're sat in front of their own ChatGPT, and they're asking, how do I estimate this? How do I find this? How do I solve this problem? How do we give them tools that can help them to interact with our people so that we can drive more pipeline that way. **Charlie Cowan:** And then, potentially, can we open up in new markets that we couldn't have done before? Just for example, as a European company, maybe you're just not selling in Japan because we don't have Japanese teams, we don't understand Japanese law, we don't understand how to sell to a customer there. Well, are there ways that we can handle that without needing to know the language ourselves? So lots of opportunities to accelerate our go-to-market teams, but importantly, to give our customers and prospects a better experience to buy from us. Now, we go down to the bottom right, to the core capabilities of our company. As I said, this is the guts of our business. How can we look in at what we do as a business to create products and services, to deliver those to our customers? **Charlie Cowan:** Can we improve these processes, reducing the time that it takes us to get an idea out into production and to bring that cash into our business. Can we be more innovative than our competitors, who are going slowly and steadily? We can go faster. And by compressing everything that we do in the guts of our business, can we reduce our need for working capital, which gives us more to spend on other areas of the business, or just reduce the amount that we need to hold? And then up to the top right, new products and services, new value propositions, things that we can only do now because of our use of AI. So what is it that we can sell that is only possible because of AI. Is this new data services? Is it new solutions? Is it new analysis? Is it new ways of communicating, new products, and so on? **Charlie Cowan:** Is there data that we sit on in our organization for our customers, aggregated across customers for different industries. Is there analysis that we've got over our customers' competitors? Anything that we are sitting on that we can now turn that into something that is valuable and we can sell to our customers as a new product? Are there new business models? Maybe we've always sold on a unit basis, or we've always sold fixed fee, or now we can maybe sell on a usage basis, or different kinds of models. Maybe we can sell higher levels of our services that are more premium, because we're going to add some AI into these tools. And all companies are figuring this out. How do we extract more value from our customers because of this. And this is a really interesting one. Were there segments of the market that you could not sell to before because it was not economical? **Charlie Cowan:** So you could think here about SMB, maybe. You've only sold to enterprise before, you couldn't sell to SMBs because it just, we couldn't staff them with sellers and support and so on. Well, maybe now, with AI, we can support those markets, and that opens up new routes to revenue that way. So those are the sort of four segments of the Gartner AI Opportunity Radar. And so let's have a look at some of the things that we see where businesses are going wrong. And here I'm thinking about the PE firms that might be advising their portfolio companies. Everyone loves a matrix, a four-box grid, so here's another one for you, which I'm sure you're familiar with. You might know about known knowns, known unknowns, and unknown knowns, and unknown unknowns. **Charlie Cowan:** It gets rather confusing, but you will know that there are some things that you know that you know, and there are some things on the bottom right. I know what I don't know. So, I know that I don't know that thing. And then on the top left, which is where you get a bit dangerous, I don't know what I know, so this is in my brain, but I'm not sure what it is. And then the real dangerous thing is I don't know what I don't know. And, this is interesting, because what I'm going to do now is I'm going to layer it over the AI Opportunity Radar. It's not one for one, but I want you to think about the bottom left and the top right here. So if you remember, for the AI Opportunity Radar, the bottom left is everyday AI and internal, our back office, our HR, our legal, our finance. **Charlie Cowan:** And the reason why I layer this on is, you know what those teams are. You know how many people are in them. You know what it costs you. And this is the immediate place that most firms gravitate to when it comes to AI. Just like that X post at the start, we can save 20% in some of these back office functions. I know what I know, I know what the payroll line is, and I can take 20% out. And so that's where people start. But like I said, in the AI Opportunity Radar, the real value creation is the top right of this, up towards the top right, and this is the unknown unknowns. I don't know what I don't know. **Charlie Cowan:** And so if I ask you in one of your portfolio companies, I don't know, a manufacturing company, what are the new products and services that are AI-enabled that can allow us to sell more and increase our value. Sell into new markets. You're gonna go, well, I don't know. I don't know what I don't know. I don't know enough about the market, I don't know enough about what AI could help with our portfolio company here, and so I don't really know. So what I will do now is just gravitate back down towards the bottom left. I know what I know, and I can easily take off 20% of that payroll line. So, this is an easy trap to fall into. I'll give you some ideas in a minute about how you can think more about this top right, but very much bear this in mind when you're speaking to portfolio companies about use cases. **Charlie Cowan:** How do we get them to think about the top right? A couple of other gotchas that we see, and going back to that Accenture stat at the start, I think a lot of this feeds into this. So, oh, we need to do AI. Everyone go and do some AI, sprinkle some AI magic over the place. We end up with just a lot of tactical IT projects, lots of individual initiatives in different business units, different areas. None of it is connected to the overall company strategy. Your company may follow OKRs, objectives and key results. These are the three main things that are going to drive our business value this year. And yet, lots of these initiatives are not very closely linked to them. And so, some work, some don't, some stay in a pilot, some don't, but it's not really driving a business strategy. **Charlie Cowan:** Link to that is, what is measured is the activity, not the money. So we've got these AI projects, everyone's being tasked on it, it's great, lots of stuff going on. Well, what are the numbers that you're measuring? Oh, well, we're tracking logins to ChatGPT. We're tracking the number of prompts in Claude. We're tracking the number of ideas that we are capturing. It's all good stuff that goes on a dashboard, but no one can tie this back to EBITDA. How much more did we sell? How much did we drive the selling price up on an average order size? How much did we save? And so, no hard links, no links to these hard numbers that someone can point back to. In many organizations, no governance. **Charlie Cowan:** So there's just a lot of these different ideas, but there's no one person who owns not just the AI strategy, but the governance of this. What is the AI center of excellence? What are the metrics? What are the numbers that we're driving, and how does this link back to the strategy? And to this lack of this oversight, either from the portfolio company themselves, or from the PE firm that's actually helping to drive or instill that governance across either one firm, or the overall portfolio. And then the final one that we see a lot of, not just in private equity portfolio companies, is just this gap between what individuals are doing and what the organization needs to do. And so, over the last year, there's been a lot of experimentation, even in very AI-forward companies. **Charlie Cowan:** We're AI-forward, we encourage people, go and get the tools, people who've got some budget, go and experiment, see what happens. And, the challenge is that that is not filtering up to driving this organizational change. And so, depending on how your business is structured, regions, business units, you want to be looking at those individual business units and regions, and saying, right, what are the processes, what are the metrics, what is driving the OKRs or the goals for this business unit? And how do we inject AI across that process so that it is the team, the division, the company that is building these AI tools and systems, and not just relying on individuals? So here are a few things that we're seeing that are working, that you can follow. **Charlie Cowan:** So, the first is, instead of just running off with these 25 sort of IT projects and going, let's just see what happens, let's come back to defining what the vision is. So, for an individual portfolio company, when we're thinking about AI in our organization, what is our vision? This comes back to a little bit about that opportunity radar, and how do we define, we want to win our market, and we're going to win that with new products, new services. It's not just about, well, let's cut out the back office and save some money. We might want to have a think about, at this business unit or division level, what kind of data are we sitting on? What kind of processes have we got that could be ripe for adding AI into it. Let's determine our current numbers. Now, this could be dollar, pound, euro numbers. **Charlie Cowan:** It could be a number, how long does it take something to happen? It could be a percent. But let's work out what the baseline is of these metrics, so that we can then come up with a plan to change those numbers. And of course, setting up some of that governance, that center of excellence. I mentioned a vision map. These we find are absolutely fantastic. So, it's a really nice, easy workshop. You can bring a number of people from across the firm, or the portfolio company. So this could be a CEO, it could be the CFO, it could be the CIO, it could be business unit leaders. And what we are asking people to do here is just to start to define, when we think about the culture of our business, when we think about our strategic objectives and our metrics, what is our vision for AI in this business? **Charlie Cowan:** This then comes out, often printed as a lovely A3, sort of branded, our AI vision map, and that can go up in the office, it can go up on the intranet, in your channels, and it defines all of the decisions that are going to come to you later about what AI programs we work on. This is the vision, to win our market, to provide better services to our customers, to sell more, whatever it might be. So that was step one. Step two is about then going and uncovering these high-impact use cases. So I'm going to show you this in a minute about how we run this in the Kowalah platform, but kind of a two-step process. So the first is launching what we call a discovery. So you can launch a discovery out to a subset of your employees. This could be just your leadership team, for example. **Charlie Cowan:** It could be, right, we're going to go out to our SLT, our senior leadership team, there's 30 people, there's 50 people, there's 100 people, and we're going to go and gather AI use cases for their business. They all then do this from within their own tool, whether that's ChatGPT, whether it's Claude, if you're using Slack or Teams, we've got a Kowalah agent. So they're having a discussion there, they're not facing a blank piece of paper, but what we then get is this whole list of AI use cases, what we call an opportunity. Having gathered those, you as a leadership team, are they going to score them on the ICE framework? So ICE is impact. So, what is the business impact of this use case? So, revenue, cost control, average order size, these kind of things. So, business impact. **Charlie Cowan:** How confident are we that if this AI system was in existence, we would have that impact? And then thirdly, what's the ease of implementation? Does this require lots of data, lots of integrations? Are our customers going to be accepting of it? And so, that allows you to take a very large list of AI use cases, and to filter out the ones that are most, most high impact. What we then do is map those against that radar. So you might have come up with 30 or 40 high-impact use cases. Well, let's lay them out across the Gartner AI Opportunity Radar, and then we can start to determine which are the ones that we really want to be focusing on. Step number three, and this is really for you as the PE firm rather than the portfolio company, is then to say, well, okay, we've been running this across one or two of our portfolio companies. **Charlie Cowan:** We've now got some tangible use cases, some ways of working. We've built some agents, some skills. So how do we now start to centralize that so that we can deploy that at scale and quickly across each of our portfolio companies? Now, we've been to a number of sort of portco away days that PE firms have been running, which has been lovely. I was at one just a couple of weeks ago, and the firm brought together three groups. So it was people teams in one room, finance teams in another room, tech teams in another room, all with this idea of sharing knowledge amongst each other, which is fantastic. But I think you can go further and say, not only are we going to get you in the room and help you and give you sort of coaching and advice, but actually, we've got templates here. **Charlie Cowan:** Here are agents ready to go, specific to our sort of investment thesis. And then the fourth part is building in this reporting from day one. Going back to some of these slides we talked about, or these reports we talked about at the start from Accenture, that people have got the strategy, but they can't link it to EBITDA. So whenever we're scoring one of these use cases, and we talk about the impact of that, we really think about impact in three ways. So one is what is the business impact, so the revenue impact, or the cost reduction impact. Second is productivity, or efficiency gains, and then the third is just adoption, people logging in and using the tool. And what we really want to guide the portfolio company towards is this top one. So what is the business impact? **Charlie Cowan:** So for this process, what are the numbers today, and how are we going to measure that so that over the course of the next 6 months, we can actually log that as an outcome, and we can go back to the board, come back to the PE firm to say, look, this was the number. We've moved that number, and we can evidence it. So that's kind of your four-step process. This all links back to driving those earnings up, because we're reporting it. Because we have implemented AI in that top right, we've got new products and services and value propositions that are AI-enabled. We can move ourselves from being a traditional manufacturing firm to being an AI-enabled manufacturing firm. We can drive the multiple, and of course, as we're running that, we're throwing off a bit more cash. We can pay down that debt, which helps as well. **Charlie Cowan:** So a little bit about how I can get started, and how we might be able to help you with that. So, one of the first things that I mentioned there was to launch a discovery. So, we can launch this, you can do this as the private equity firm. You could launch a discovery into one of your portfolio companies, or we can work with your portfolio company directly, and they can create that. We do a couple of things. So, one is we sort of capture who is it this discovery is going to go to. Is that 10 people? Is it 100? Is it 20? And then we capture all of their email addresses so that we can send it out to them. We then determine how are they going to respond? Is that via Slack? Is it via Teams? Have they got Claude? **Charlie Cowan:** Or are they just going to come into the Kowalah platform and do that and guide them through that process? Having captured all of these use cases, you get loads of really rich data, and in fact, let me just switch over to a little sort of demo environment here where I can show you this. So this is the discovery and the view that you would have as a customer. So, let me go here to the, so you're able to see, by all of the different departments how these use cases are coming in, and their average ICE score. What you're also able to do, because the ICE scoring has happened, you can start to take a look at them by, what this is showing you is the impact of these use cases, so obviously more impactful, more valuable to the business, over to the right. **Charlie Cowan:** The ease of implementation of that use case, so more easy to implement is important. And then the size of these circles is your confidence that you can implement it. And so you can have a look around here. Well, we've gathered 20 or 30 use cases, but actually, it's these ones in the top right that we should be focusing on, because they are the more impactful, easy to implement, and we're confident. So that's helpful for you as well. We can then also lay it out over the opportunity map. So, if you remember, game-changing AI, everyday AI, external and internal. And once again, we can start to say, right, well, actually, these are the things that we should be focusing on, rather than the ones down on the bottom left. **Charlie Cowan:** These might be useful in terms of the short-term efficiency gains around the payroll, but let's think about some of these things, and if we haven't got enough of these things, let's go and find them. So all of this happens within the Kowalah platform, and depending on how you want to run it, whether that's you as the PE firm or the portfolio company, this gives you, as the PE firm, huge visibility over what is happening and how you can drive this across other members of the team. Alright, so that brings us to the end of the core part of our subject today. I hope you found that valuable. Whether you are coming to it from the firm, the investment firm, or whether you're coming to it from the portfolio company, there's some ideas there for you to go away and start thinking about how you are gonna proceed. **Charlie Cowan:** Now, as we go into our usual segment of AI in the News, a few updates are coming out this week. If you were on last week's and the week before, you may have heard us talking about Fable 5, the newest, high-power model came from Anthropic. It was launched, and three days later, the American administration put an export ban on Fable 5, which meant that Anthropic had to remove the model from availability. As of today, it is just being relaunched, with a few controls that are available to allow people to use it. So just be aware, if you're Claude users across your company, across your team, you're going to see Fable 5 being returned. I won't go through the things that we talked about a couple of weeks ago, about pricing, and about the power and so on. But this is a whole new class that is being called the Fable Class of Models. **Charlie Cowan:** That are above Opus and ChatGPT 5.5 that have been previously the most powerful models. Gonna be hugely impactful for those working in pharmaceuticals, complex financial analysis, complex cybersecurity, these types of work. At the same time yesterday, Anthropic launched their newest Sonnet model, Sonnet 5. So, up until yesterday, we were on Sonnet 4.6, and so this is a bit of a leap, because instead of it being just 4.7, we've reached a new sort of round number of Sonnet 5. Now, Sonnet is the daily workhorse model. Any of you that are in Club Claude, and your companies are using Anthropic and Claude, Sonnet is what you should be encouraging your teams to use as their daily workhorse. It's much more efficient on tokens, and as you're going to see here, it's actually at near comparable power, but at a significantly lower cost than Opus. **Charlie Cowan:** So, these charts are always a little bit hard to decipher, but, I hope it shows you a couple of things. So, the grey, if I hover around here, the grey is what Sonnet 4.6 was before. So, at various levels of thinking, it was this amount. This shows you the average cost per task. So, the higher up it goes, the more it costs. Well, the higher up it goes, the better it is at working, and then the further to the right, the higher the cost goes. Now, with Sonnet 5, you can see that its pass rate has gone up, so even at very, at the low thinking, it's gone from right down here, up to what is, like, sort of, 75, 80% near. And, this is fairly comparable with some of these pass rates that are coming from Opus, which is the top model. **Charlie Cowan:** So, low and low, so not too dissimilar, but the cost per task is significantly less. This is a logarithmic scale, so something that might be sort of 40 or something, going down to near 20. So you're basically paying a lot less for the same amount of intelligence. So very much encourage your teams to be using Sonnet as their daily driver, and only go up to Opus, and very occasionally, up to Fable for really more complex tasks. Now, at the same time, this was maybe 3 or 4 days ago, June 26th, ChatGPT, OpenAI, launched their new series of what we would call the Fable Class models. So they've come out with 3 different new names, Sol, Terra, and Luna, which are in this GPT 5.6 series. **Charlie Cowan:** Now, unless you are spending all of your day in the world of AI, you are going to get very confused by all of these names. It was ChatGPT 5.5, but now we've got 5.6, with variations of Sol, Terra, and Luna. Very confusing. All you need to know is that Sol is the most powerful one. And this is this kind of Mythos or Fable style of model from Anthropic. Very, very complex, very, very powerful, and is being rolled out in a sort of limited release to try and stop having that Anthropic-style blockage by the U.S. government. Now, benchmarks are benchmarks. I say this every time a new release comes out. All you need to know is that benchmarks keep getting better. But let's take a look at this. So here, we've got Claude Opus 4.8, which up until recently, was Anthropic's most powerful one on this terminal bench, 78.9%. **Charlie Cowan:** ChatGPT 5.5 was 83.4%. Claude Fable 5, the one that's just been launched again, is 84.3%. And the Claude Mythos 5, the top one, 88% on this terminal bench sort of coding platform. And so what you've got here is Terra. Here, well, Luna over on the right here, ChatGPT 5.6, Luna, 82.5%, GPT 5.6 Terra up here, and then Sol, 88.8%, and then Ultra, with sort of Ultra thinking, 91.9%. So, all you need to know is they're just getting better and better and better, and whilst ChatGPT 5.6 is the more recent one, no doubt Anthropic will come out with another one, and then Gemini will come out with another one from Google. The numbers are getting better, the tools are getting more powerful, and as a human, we are no longer the limiting factor in these tools. So, that's all I wanted to cover in AI in the News this week. **Charlie Cowan:** New tools, some recurring tools, go and try to test them out. If you're an Anthropic user, then you're going to find Sonnet 5 in your account right now, and you would expect to see Fable 5 a little bit later. If you've got any questions, do feel free to ask them, but I haven't seen any that have come through in the chat, so I'll keep an eye on that afterwards. If you're watching on YouTube, thank you very much for joining us. You can put a comment into the YouTube channel, and we will come back to you with any questions and answers there. Thank you very much, and next week, we are going to be having another one of our fantastic Spotlight webinars, where I'll be interviewing Henry Chart, the CEO of CFS. **Charlie Cowan:** CFS is a construction manufacturer, and this really feeds into what we were talking about earlier, about transforming a traditional manufacturing business into an AI-enabled technology business. And we're going to talk about that journey into building out their new structural engineering platform built on the Anthropic platform. So, you can hit that QR code, or go to kowalah.com, and then look for webinars in the resources section, and I will chat to you next Wednesday. Goodbye. --- This transcript is published by Kowalah, a UK Anthropic Implementation Specialist. Webinar page: https://kowalah.com/resources/wednesday-webinars/ai-across-your-portfolio-a-playbook-for-pe-firms All transcripts: https://kowalah.com/resources/wednesday-webinars/transcripts