From Buzzword to Business Impact: Embedding AI Across Portfolio Companies | CapLink Group European Summit 2026

[MUSIC] D-Neil, Executive Director at ECI. We are a modern intelligence service provider to the alternative investment space, predominantly focused on private equity, providing a raft of solutions around cybersecurity, managed services, and the buzzword of today AI and value creation within our clients, within our own business, and helping our clients within their own port goes. And I am joined by Valerie, who lets introduce herself. Thank you. I'm Vene Heligat. I lead the tech data and digital at Astog in our operations team. And if you don't know Astog, we're a private equity firm managing in three main sectors, healthcare, business services, and software. Excellent. So we've just finished, or Valerie has just finished a panel on GENTAKI and value creation within port value companies. And I feel that you have a lot more to say and a lot more to give and educate us on your experiences and what you're doing within your own business. I can talk a little bit holistically about what ECI are doing, what we're seeing across our own client base with regards to creating value with respect to AI. But I think to start Valerie, if you wouldn't mind kicking us off with some examples of what you're seeing, what you're doing, maybe at the higher level and then drilling into it, so maybe some specifics. Sure. So we have two major streams regarding AI and the streams that I'm going to talk you through are cross-portfolio. So we do it for at least some of them for like 25 of our companies and the other ones for about 15. So one of the first stream that we work on is leveraging AI to increase efficiency of our port value companies, focusing on good market customer operations and support functions. So I have a very structured playbook and I would say it's low risk, extremely codified and we take company per company, we look at what their balance sheet looks like, where are the cost buckets and we help them identify the most promising use cases and we provide support, tons on support and guidance to execute and capture the value. And this usually we try to get the first use case of the ground in less than two months and then usually it's a 12 months journey. And of course if we find on our way some critical business processes that we can automate, we also do that. So that's the first stream, it's more applying AI to our current operations. Can I ask how many companies you applied at to how many of your port goes as a design of a book? Currently 12 have started their journey with us. Some of them were already there doing some kicking off their journey by themselves. Yeah, yeah, okay. Between courage 12 over the last two years, 12 companies to be honest. And I think the second part is that we hear an awful lot about repeatability and sort of an individual use cases and there's no one silvable, no one size fits all. Have you managed to conquer that? Have you cracked that challenge where you've been able to create something within those 12 port codes that is repeatable, something that you can productize if I call it a word, if I productize an AI solution? So I don't know if we managed to crack, but at least we're very investing a lot of effort to work on it. Right. The way we structure our actions around that is trying to find a use case. Usually it's in go-to-market or customer operations. That is a pain for all the companies that we have in our portfolio, or at least a large majority of them. And then the way we tackle these, you know, how can we make something repeatable, is through a very collaborative approach with our portfolio companies. Our community of practice is extremely dynamic. We have over 200 participants, so digital tech leaders from our portfolio companies. We talk to them every week. They talk with foster peer-to-peer knowledge sharing, etc. And so they get together to solve their issues. So typically, if one of the companies looking into how can we automate sales outreach, they're going to reach out. We create some share resources. So they know who to reach out, say, "Hey, how did you do this?" And then we get them to speak together, so we gather them during meetups, and we could define the learnings. It's like a self-healing, self-learning, self-education community, if you will. Exactly. So we contribute. We create the space for our companies to share their learnings. We contribute with our expertise, or we ask an external partner to step in. And then we create the space for them to grow together. Because it would be very, I mean, it would be a big lack of humidity to say that we know what to do exactly. Okay. And these particular sources, we're going to move on to the second use case that you mentioned, surely. These companies, is there a sub-sector of your port codes that this is applicable to? Is there a market segment that this is being driven by? No. No, because typically, if we look at go-to-market and customer operations, it's a bit less in manufacturing, but business services and software, and large manufacturers, it's applicable. Okay. Who doesn't want to have higher conversion rates for its RFPs? Or a quick go-to-market strategies, right? Absolutely. And better access to data, better results on the data that we have. Better qualified leads. Absolutely. Training the sales also. I think, to echo, I mean, it's certainly an area that we, internally for us at the ECI, it's certainly an area that we are using AI is responding to DDQs, responding to RFPs, or pre-qualification questionnaires, being able to access data, and then repeat the responses, supposedly, using time and human capital to write the same answers, which is what I mean, unfortunately, or fortunately, which is what I've had to do for the last 20 years in my career. Having the ability to cut that down now is formidable in terms of our approach to be able to respond quickly to clients. And, concisely, I think, the quality and cleanliness of the data in the background is massively important. The governance around that data, and I think the checks and balances that still need to go into our paramount, but being able to get that data quickly, get the answers to the questions quickly, and then go through your checks and balances as a final step is hugely powerful for any businesses. So moving on to the second use case. The second topic that we're addressing is AI in product development lifecycle for software or tech heavy companies. Why? Because we saw over the last two years that we're starting to have two kind of groups. The companies that are using AI to do the same thing as what they were doing before, but they're doing it just faster. So typically shorter cycle times for their product, but it creates very marginal incremental gains. And then we have a second group of companies that started tackling the hard stuff, really rewiring the way they operate, putting AI at the core. So it means rethinking their operating model and the way that engineering and product teams and business teams are going to work together. It means embedding AI at the core of their product through building agent workflows inside their product, and also having a different plans on how they structure their product problem app. And for these companies, what are our biggest challenges to put at scale, what they have started to demonstrate, and what we're helping them demonstrate. Meaning they start working in a very futuristic way. I mean, futuristic. If future layer will probably happen in 12 months, but then illustrating and demonstrating what does it mean to build an operate software in a modern way. And while they do this, we nurture the transformation plan that will enable us to scale these best practices to the rest of the organization. Because it has many impacts on. What do you think? So I've got two questions around this. Are we seeing a common driver? So we heard on the panel that you were just on about use case deployment or seeding the field and just giving people tools or whether we're kind of driving from the top down or the bottom up in terms of whether it's, you know, there are the younger generation that are driving there, the use of AI or is it the CEO. So two questions. Is there a common denominator that's driving the deployment for AI within these firms? That's question number one. And question number two, I guess, is what's going to be the main barriers for further investment in AI? Because I think is, and again, coming back to the panel that we just heard, we talk about we kind of get to a so what state. And we deploy the tools and we go through development and we back the teams and you've got tiger teams that kind of go out and we back those guys. And you kind of get to a point where you say, so what, you know, what is the ROI? What is the point of, you know, what is the point of the development? of ROI, what are the benefits that we're seeing, how do we monetize that, or how do we productize that, what do we need to do from an investment perspective to move forwards with that, and who signs off on it? Do you sign off on that as a GP? Do your port coves have to sanction internally the investment themselves? Do you give them the investment to make these changes in this technology landscape? So, two loaded questions there, if that's okay. So the driver for software tech heavy company is very simple, right? It's a question of sustainability. Can do we believe that we can continue delivering software the same way that we're delivering software six months ago? It's obviously no, right? That's why we started the program and of 24. So the driver is just, I mean, to be able to continue to have a competitive team of engineers and product experts. It's also a question of attractivity, right? Now the best talent on the planet, they are not interested by delivering software the old way. It's like if you're asking today to a business analyst who uses cloud every day, you know what? You don't have the right to use neither cloud, neither a charge IPT in this company. Welcome to Excel. He'll never sign for contract. So we need to be in front runner to attract the best talents. That's the driver. In terms of ROI, etc. I like to flip it the other way, the question the other way around. What is the value of a company who is not a tech company who is not leveraging AI in its PDLC? Right? So zero or close to zero. Yeah. So it's not a question of should we go, should we? It's the question of how fast can we go to keep up with the market and to keep up with the available technology that is popping up every day? So it's really this muscle that we want to build and that we've started very successfully building in the 25. How do we manage to get our tech companies to be fully at ease with these new technologies? And how do they evolve? They ways of working. And how do we put this at scale? I think I'm starting to. So I think we've covered off two sort of two work streams that you've described, which is which is fantastic. I'd like to understand a little bit about the cultural changes within your business from the value creation team to adopt AI within the portfolio companies. I'm just like, when did you make that decision? Why did you make that decision? Like who made the first move? And what were the reasons behind that? And then a secondary to that initial point is what tangible successes have you seen from making those changes over whatever period of time that was? Okay, so it all started. Many, many, many movies you've got. The operations team at the stock is rather new, right? The keys joined as a bit of operations a bit over a year ago and I was here, I joined my stock two years ago. And I think, you know, a stock is a very pragmatic, impact-driven P. So we tested something mid-24 with one of our portfolio companies, we reached out. I said, "Hey, I have this problem. Is there a way you can help us?" And so we did. It was our first GNII initiative generated incredible impact. And that's how it started. So mid-24 still very, very, very young, right? Still very, very, very young. And now having impact across 12 additional portfolios in one work stream and how many others in the other work stream in the software work stream? Well, software 14 companies. And overall AI in business functions is about 12. That's amazing. That's amazing. So the community helps, right? Because, of course, you cannot, I mean, we're a small team, we're four in digital and tech. So you cannot manage this by yourself. So we have vetted partners that help us on the ground. And we rely a lot on the communities, our champions. You know, they help our other talents develop. We do a lot of events, in person events, online events. So would you say you do a lot of collaboration events within your port? Yes. So the teams within your port goes. Gov and buy, you know, gov and buy the, you know, the four members of staff that you have internally asked to talk. And then you, so you've touched upon here. I didn't, I didn't want to turn this into a pitch and it won't be. But you mentioned partners. So can you tell me, can you tell me the importance and the value that the partners bring to the overall picture or individual projects or those sort of what does that give you access to scale? Okay. Right. So if we need a typically three weeks ago, we kicked off five pods. The pods that I was talking about for software companies. So five pods in three locations. Each pod requires two, three FTEs on the ground. That's where we, that's when we pull in partners. We also have partners on very specific use cases. Right. Right. So for example, automating a configuration solution for a broker company. Well, we look for a partner who already did this because he'll be able to bring his learning. Okay. So repeatability across the part of network as well as scale. And I guess you have the ability to collaborate with your, your teams and your your your tiger teams within the poor coast to ratify sort of what's being driven through to close. What do you see for your portfolio company development AI development within portfolio companies adoption for the next? And let's not push this out so far, right? Because I think we heard, we heard in the panel that, you know, the internet was adopted over a 15 year period and it was like it took 15 years for this, this thing to kind of become a thing, which is obviously rules everybody's life now. What do you see over the course of the next 12 to 18 months with respect AI given the rate of change with technology, given the rate of adoption with AI and I mean, I can attest to this from from a client perspective and from ECI's perspective that, you know, six months ago, it was it was a buzz word, you know, now we're engaged in over a hundred projects with our clients globally, you know, talking about various different use cases, various different deployment readiness programs, you know, security assessments around AI, specifically for ASTWG and your port coast, what are you seeing for the next 12 months? So, across our sectors, in the next 12 months, all our C3 tool will be completely fluent in terms of AI, being able to understand the difference between an AI driven workflow and an agentic workflow, for example, and being very clear sighted on what it means for the organization. I think they will all have the beginning of a transformation plan, meaning they will have been able to prioritize the domains on which they will be the high the most impacted and they will have had time to structure plan. Software companies, it's a different story, they will be way ahead, but I think if we manage to get to this, it will be good because if our C suite and our COs, you know, understand knows exactly what domains will be impacted and have a plan, it means that they will have run some experiments and they will have selected their most promising use cases and be ready for scale. I think in less than 12 months, I mean, it can be a bit of a challenge because again, you cannot capture any kind of impact until it lands in the PNL, so it's not by saving 10 minutes of 5,000 people that you're going to generate any kind of impact. Sure. So we need to do the next step. Next step and push that out. I think I would second that and what I would like to see and echoing what you said, I would love the industry to be able to sweat down or put into the melting pot all of these different use cases and come up with a subset of use cases, a subset of meaningful use cases that we can and I keep coming back to it to productise. I don't think AI has been productised. I don't think you can productise AI. I think it's something that needs to be a working progress over the next sort of 12, 18 months to really find out what is creating most value, what is repeatable, what is common. I don't think we've got a commonality across the different use cases at the moment. I'd love to see some commonality. I'd love to see us in a position where we can productise 5, 10, 20 different use cases at the moment. It's 100. So we bring that down and then we really start focusing on and polishing and developing those use cases. for mass adoption in the market. Valorate, it's been an absolute pleasure. Thank you for your time. Thank you for your insights. It's been a real eye-opening conversation for me to understand what you guys are doing and what the market's doing within your portfolio companies. Thank you. It was a pleasure.