Launching the Future: Ash Pembroke on AI Strategy, Scale, and Value Creation in Private Equity

If you can build a tool set that can make your fastest horses faster, I think that's a much more compelling and clearer value proposition than trying to bring sort of the bottom 30% of the workforce to parity with the middle 50. Welcome to AI and PE, the future of value creation. The show where we cut through the noise to hear directly from experts, practitioners and members of the PE community on how AI is transforming portfolio companies, accelerating growth and rewriting the playbook of private equity. I'm your host Kyle Romer, managing director and head of US Data and Analytics here at Accordion. In today's episode, we're joined by Ash Pembroke, managing director of AI strategy and value creation at Accordion. She's got some great advice on how to approach your AI strategy from accelerating your strongest performers to pushing your leadership team to identify your biggest business needs, where AI, data and tech deliver the highest ROI. Here's my conversation with Accordion's Ash Pembroke. I thought I'd just start with, what's the biggest wild moment you've had so far in AI in your career? Okay, so I was with a coworker at a former firm and we had challenged ourselves to grow without increasing head count and some functions where you would have assumed that you needed to grow head count in order to scale things like sales, sales, ops, engineering. And I saw a moderate firm with a really good process hygiene and basic GNI integrations between Slack, Salesforce, some custom development essentially scale their team like we're talking like 25 to 50 to $100 million without materially increasing the size of that team, they invest in the caliber, but not necessarily the supporting personnel. And I think that was the first aha moment for me where I had seen all these tools come together in a function to drive sort of a big ticket item for the leadership team and accomplish a goal and I was able to see that and be a part of it. And so it converted me to the ability for AI to help mid market company scale and grow really quickly. So I became really passionate about it. I think we're constantly talking today about our service businesses getting getting up ended, disrupted our folks catching up that are maybe up starts to more established players. I'm finding it harder and harder to understand how many people work at this company? How quickly are they turning around proposals? It feels like smaller shops are able to show up in a material way at scale that I don't think they could have before and even if their hit rate stays the same on lead conversion, the at bats that they get are a lot higher. And I think that is a really interesting and challenging thing to consider. What was an example with them like either service delivery or go to market where you're like you really saw the impact where you didn't have to think about SGNA scaling the same way it might have historically. I'll give you another example that I think is really concrete. So I've been working with a customer who is a roll up of its HVAC lots of plumbing and electrical so they do all kinds of things. Labour turnovers a huge issue in those businesses and it's not just turnover just bodies. It's turnover of less experienced people that miss estimate jobs go out and mess up workers at cascading effect when experience leaves. And so one of the most compelling uses I've seen of like AI in the field as an example is if you have an app that's trained on say every version of HVAC equipment that you could possibly get from any OEM in all the really inexperienced technician has to do is snap a picture. It automatically catalogs the system that that technician is looking at. It can say like hey, these are these are knowledgeable estimates. It gets somebody who's new at their job and maybe not super efficient that much more efficient faster. And then the accuracy of every other downstream decision starts to get better. In the enterprise, I mean, I've got some some lovely mentors at large OEMs right now that are eating technical debt for breakfast in their automotive platforms in their legacy code bases in their hardware structure. And it's public, it's in the news. But if you would have asked me, you know, a few years ago what it would have cost to drive an enterprise transformation program, one of these really large companies, you would have seen price tags hundreds of millions, right? And these sort of Sisyphean timelines. And now that's getting much crisper with the help of Genai and code code development and creation. And I just think it's it's becoming material. And the error rates are going down and so as a former ledite in the last probably 16 months I've become much more of a believer. I fully agree. I think the idea of saving a single headcount in a department like we should not be talking about that. Frankly, every single person in every single department should be leveraging some of these tools to make them more more efficient, more effective. Let's take take a consulting business like ours. You have this interface with the customer, someone enters a lead into a CRM. Then they log meeting notes, then they do this. You sort of are context switching across multiple applications. In the past, if you did a time study, you might throw RPA at that or some other, you know, random, daisy chain of solutions. I have seen almost all of that be automated away with custom integrations with Slack and Salesforce and whatever repository is your contract master. Like I can spin up a lead in deal cloud or Salesforce. I can memorialize a conversation into notes. I can kick off those notes into bullet points for a proposal. I can do all of that within the console like Slack. And there are lots of AI native businesses that are doing that. But there's this interesting dilemma that I consistently face, which is, are you automating away the substance and the expertise? And I'm just curious, like, if you think about pitching a client as an example, if you're in professional services or in service industry, are you losing some of that? Or do you feel like the models are good enough now where it's providing the right level? I have a much more, maybe, compassionate evaluation of how AI helps me. And here's what that is. How long would it take me to explain what the customer said and what I think the shape of the proposal would be to say like a 30 year associate? How long would that take? And how many iteration cycles would we go through? And I think this is what people are grappling with across the economy to take it really big for a second. The structure of how we, as knowledge workers, create our teams and craft it. I think a really, really well tuned version of an enterprise GPT with some of these different agents that can plug into creating PowerPoint or proposals, they can get you probably as far as that first draft. Is it perfect? No, but it wouldn't have been perfect either way because so, so much of our work is caught up in like data entry, drudgery or translating a meeting into notes or notes into this or putting that into a CRM. Every team experiences this swivel chair problem. And I think it's actually one of our least favorite parts of our job. If we're honest with ourselves, but it's so much of it today. And what if it just didn't have to be? I like the general contract of like, if it's painful and manual and boring to do, gosh, shouldn't we be thinking about AI for something like that? So I want to, I want to ship gears a little bit. So we spend a lot of time with PE backed CFOs and private equity sponsors. And if you're one of those sponsors, whether you're an investor or an operator and you're prioritizing, you're going to put in a half million dollars for a million dollars in the CapEx budget for the next year and 26, like what are you putting it towards? Like imagine you're a service business or imagine you're in industrials. Like how do you think about that? Because that is one of the most common questions we get today is where you're putting your dollars towards and how do I get a return on that? I am hearing a lot of conversations about we have to be AI enabled at our exit. What does that mean for us? What have we invested in so that we can tell this value story? And there's a couple areas that I think AI opens up. So one, traditionally, there's been this kind of like, reticence to operationally integrate certain functions. If you're trying to get, say, three or four product companies to sell together, you different sales forces, not the software, like humans doing things, and you're trying to rebrand and create this catalog of services and it's very additive and you have some pricing strategy baked in there. It takes like a year or two to get people walking and talking in the way that you're going to start to recognize that value. I've seen AI enabled like learning management solutions. There are ways to scale people's change readiness and also the change to new operational processes. I think that's a huge opportunity that's sort of untapped. Yeah, I think that's an important takeaway here. I think AI, it's built on data. So there's a significant amount of us that have sort of looked at the value creation plan and we've said, hey, you know what, I'm not going to, I don't think we should invest a million dollars in a data warehouse of any variety at this moment, right? We've got a group A over here has this data mark. We have this for maybe sales, operations is over here. We're all fine in our own silo and then we miss that first timeline for exit or the value creation plan changes and we need to instead of just like growth, we need to maybe incentivize people to purchase more or change the dynamics of like lifetime customer value. How do you get more utility out of that without that data? And that's just like getting your understanding of what the value could be. It's not the execution of it using AI, which does require some data being in the same place. What I think is compelling about that challenge now is that if you were to send me and say like ash, you're going to have to go harmonize all this data, go talk to all these people, figure out the data cataloging of design this EDW with AI enabled teams actually doing data engineering and requirements gathering. That's months, not yours on the enterprise scale. It should be. We should also be challenging ourselves when we invest in data to do it faster and more elegantly than we have in the past. You need to do that stuff like you need to do it yesterday, certainly prioritize it. But just be clear, you're not going to get a ton of value out of it because that's your first kind of entree into into AI. It's the second and third use cases and examples that are you're going to take a little bit bigger rest, you're going to redesign some things that are core to the business and or some core processes. And then I think, you know, the last category I would say is I'm very challenging. I was just on with a group of CFOs this morning. I said, I think you guys should go back and look at your last three years of strategy and look at any place where you said cost too much. We can't move into that adjacent market. We're not qualified, cost a lot of money, I'd have to spend up another team here, go back and look at those strategies and assume every step of that business case is now enabled and just like the denominator is changing, right? And so I think we just have to go back and reevaluate some of those strategies. And so I think going back to the drawing board with some of those teams, reevaluating what it will cost to either acquire or expand or grow in this new way, those are really interesting opportunities right now and I see teams pulling away who are thinking creatively about that and going and executing against that because that's growth. That's not just cost takeout, right? It's not just using AI to say like, oh, I'm going to remove one head count out of each, you know, shared service division. Those are small potatoes in the sometimes in the scale of what the sponsors want to accomplish. And so I think challenging ourselves right now to think really big, even if it feels a little bit weird, we're used to always focusing on tactical value creation. We know the playbook, we know how to execute. I think challenging ourselves in this moment is a really good thought experiment if nothing else. Go back to your value creation plan, like really go back, really push it through the paces of where the technology is today and what you might do differently to your point to expand product service, market, et cetera. I think it's a massive opportunity and probably under underutilized right now with CFOs and management teams. What's like a like a fundamental thing that that CFOs could be doing right now with AI that you feel like is being overlooked? I had a long conversation with a leader who said he spends most of the month trying to close the books. And then when he does close the books, he spends another week trying to compile all the data sets so we can create a board deck and he said there's not a single week that I'm not closing the books, preparing for that board meeting or working on the weekend at commentary. And then of course, you know, why did this number go up and down, pushing that all the way back down through the plumbing of the organization and pulling those in size back up. And if I could just share one thing, it's that if it involves like the routine commentary generation about data that just every month, you have a new row, it goes up. It goes down, start using AI, not just to create your content, but to auto summerize some of this stuff for you because I have yet to see a V1 of a deck that goes to a CFO, not actually be totally rewritten by that individual like on a Saturday morning, right? That always happens. So if you can let things be a little imperfect from an AI tool and let that start to fast track whether that's accelerating the close, whether that's auto generating commentary, whether that's using off the shelf AI tools to do computer vision to extract text off of PDFs that your team has for competitive intelligence. Just think about all of the ways in your life that your team is just really kind of struggling with time management because of all this process driven stuff. So just go back and see if you can get the pieces that you have to do every month, optimize so that you can breathe. And if you can breathe, you can start to do really fun stuff like go answer secondary and tertiary questions related to those metrics or say, I want to test this in this market next month because I've seen it three quarters in a row and I'm tired of it. Like that higher order of thinking, I think, is difficult to do when you're struggling with small teams just to perform the functions to keep the business operating. Let's move into something that is just that all over headlines in the last several weeks is this MIT study on the 95% of AI projects pilots are failing, advice to a CFO, a CTO, a PE back business, like, how do I just incrementally improve my success rate of taking on an AI project? Like what are some of those things that will lead to a more successful outcome versus every other technology implementation thing they might have tried in the past? I think one of the reasons why things are failing is that we are trying to take a tool that does something really well and it's getting people to parity and we're applying it broadly to the workforce and we're saying, wow, this is really doing a great job. Meanwhile, your top 25% of performers are like, this is not useful to me. I could do as much work in my day as the bottom 25% by three. I'm just a much more efficient person. And so one of the things I would challenge CFOs to do when you pilot this stuff, there is going to be there's going to be this like adoption group and they're going to raise their hand and they're going to be like, I want to do this. I want to do this. There's going to be another group and they're going to be like kind of crusty and not as much fun and they're going to be like, this is stupid. I don't want to use it. I'm better than this. I would encourage you to figure out if you can build a tool set that can make your fastest horses faster, I think that's a much more compelling and clearer value proposition than trying to bring sort of the bottom 30% of the workforce to parity with the middle 50. And so if there's anything you can do to mobilize your highest performers around this at a very tactical level and empower them to design solutions, that's where I see big differences. Could we make it real for the audience like, what does that mean in this context of AI work? Here's a really tactical example of that. At a former firm, we had tried rolling out some supporting functionality to route tickets. It was contact deflection business case, right? This was for a medium size SaaS company and they were, you know, they had kind of maybe gone to the lowest common denominator in their offshore shared service center. They had not done really good documentation and they had a lot of turnover in the, you know, probably the newest 30% of their CSRs and they found a material increase in the productivity of those CSRs. What they did it do is say, if I take the most productive group of my people, I give them more autonomy, right? So more decision making power within the context of their role. So more data to make decisions kind of like the four seasons model or any luxury hospitality company that sort of gives an employee like a certain amount of cash every day. Think about that in terms of capacity for autonomous decisions and customer service reps. If I do that and I couple that with productivity tools to get my highest performing people out of drudgery, I might not need the bottom 30% of the pyramid. And so when I see these pilots that have these wonderful metrics wash out in adoption, it's usually that scaling moment and it's because a lot of the workforce kind of didn't need it. And they were grasping its flaws about what the biggest problem is. And the thing that's the easiest to automate first is probably not your biggest problem. Certainly an exciting, exciting time to move past that kind of first level level use case or tooling to something more transformative, I think. The other side of that is if you are in a situation where the problem that you need to solve cannot be solved by humans. I'm talking about like thinking boldly about what AI does well. So in that same example, what we found was this SaaS company didn't need to automate the sort of first three to five steps of an interaction with some of their lower level CSRs. What they really, really needed was a customer success onboarding agent that could take a new company that had bought the SaaS software. And then quickly go through, have you checked this? Have you done that? Or configuration settings, here's your SLA. Think about all the steps and customer onboarding. That was what got them the biggest lift and it also standardized their data as they onboarded customers. So that entire thing created efficiency across that value chain, but it took some time because I think the leadership wanted the quick win and they wanted the safe use case. But there's so much. And for I'd say for a while now, certainly the last 12 to 18 months where there's all these small efficiency plays and people are taking very, very small risk. And as a result, classically they're getting very, very small returns on some of these things. And I think the safe use cases right now are getting commoditized into tooling. So use the tooling, turn it on, use it. But then maybe go to phase two and I think that's where you're seeing a lot of silent but transformative pilots because there are people don't want to talk about what they're doing right now because they're building true IP. These have been like really, really interesting examples, like real examples that are in the market. Whether it's for CFOs or not. And I'd love to just get a sense from you, like what are the ingredients that frankly piggyback businesses should be looking for in a partner to deliver some real value around AI? I would challenge everyone to think about solving problems with small teams because you don't need an army of people to write documentation, to transcribe meeting notes, to turn meetings into a grid, like on Excel. Like think about all those artifacts that help you get to decision. I feel more involved in than ever that a small group of people with these interdisciplinary skill sweats and these tools can actually execute really strategic change in an organization without over engineering the size and scope of it. And so, you know, as an introvert myself, I'm like, well, we don't need 20 people at this party. What about three or four? And that makes me, that makes me happening on the inside. It's good for the budget. All things are good. It's not so dissimilar to the concept of, you know, teams no larger than two pizzas, right? But like now, now I think it's, you know, you could argue, gosh, that should really only be a one pizza team, a three to five person team delivering some massive, massive impact I think in an organization. But I keep point to that, though, is just multidisciplinary skills to be able to tackle tackle some of these problem statements. Yeah, I would say hiring people that think about patterns and repeatable things and that's really how you get scale with like consolidated AI use cases. So patterns, scientific methods. So how would I know this was true, not just conjecture, but also like the spirit of, I hate pretty sure, but this is what I don't know and I'll figure out how to get it. And then I think people with that owner mentality. And so I think what's challenging, you know, even to us as we recruit is I'm essentially asking people to go out and look at their job and automate away drudgery. And I think that that first step makes people feel a little bit insecure about their role. And then it also makes them move laterally across different functions, right? Because they can't just do one thing really well. They have to do a lot of things alongside that. I think when you do that, you're pressure testing like traditional job descriptions. And so somebody that's comfortable with ambiguity to a certain extent, I think is also really important. If you've got folks like that in your organization, I'd be pulling them out and empowering them. That's it for this episode of AI and PE, the future of value creation. Be sure to follow the show on your podcast app of choice so you don't miss our next conversation on how AI is reshaping private equity. Follow according on LinkedIn to see more insights on value creation and private equity. And if you want to learn more about what we do, please visit accordion.com. [BLANK_AUDIO]