Sept. 24, 2025

Margo Polishchuk: AI Won’t Replace Expert Networks, It Will Supercharge Them

Margo Polishchuk: AI Won’t Replace Expert Networks, It Will Supercharge Them
Margo Polishchuk: AI Won’t Replace Expert Networks, It Will Supercharge Them
AI Pathfinder for Private Equity Podcast
Margo Polishchuk: AI Won’t Replace Expert Networks, It Will Supercharge Them

Summary In this episode of the AI Pathfinder for Private Equity podcast, host Steve Budd speaks with Margo Polishchuk, co-founder and president of ProSapient, about the evolution of primary research in private equity. Margo shares her journey from the expert networking industry to founding ProSapient, emphasising the importance of customer-centric innovation and the integration of technology in research workflows. The conversation explores the impact of generative AI on the industry, the nec...

Summary

In this episode of the AI Pathfinder for Private Equity podcast, host Steve Budd speaks with Margo Polishchuk, co-founder and president of ProSapient, about the evolution of primary research in private equity. Margo shares her journey from the expert networking industry to founding ProSapient, emphasising the importance of customer-centric innovation and the integration of technology in research workflows. The conversation explores the impact of generative AI on the industry, the necessity of maintaining a human element in AI-driven solutions, and the evolving expectations of private equity firms. Margo also discusses best practices for firms to prepare for tech-enabled research.

Takeaways

  • Generative AI is reshaping research workflows in private equity
  • The integration of technology must maintain a human element
  • AI can enhance the speed and precision of expert intelligence
  • Firms need to adapt to evolving customer expectations

Chapters

00:00 Introduction to ProSapient and AI in Private Equity
01:07 Margo's Founder Story and Market Insights
03:46 Identifying Gaps in Expert Networking
10:14 Integrating AI into ProSapient's Model
13:04 The Role of Generative AI in Due Diligence
18:13 Customer Expectations and Market Dynamics
21:12 Preparing for ProSapient: Best Practices for PE Firms
23:15 Unlocking Unique Insights through Expert Conversations
25:00 Challenges in Building a Tech-Enabled Service
27:13 Future Directions and Geographic Expansion

Links

Margo's LinkedIn

ProSapient

AI Pathfinder Private Equity AI Strategy Meet-ups:

https://www.aipathfinder.co/private-equity-ai-events

Have a question or would like to suggest improvements, please contact steve@aipathfinder.co

SPEAKER_01

Hello and welcome to the AI Pathfinder for Private Equity podcast. This series is all about helping private equity firms move from AI experimentation to confident action. And one way I'm doing that is by speaking with founders who are building focused, high-value solutions across the PE workflow. Today's conversation is with someone who's taken a very deliberate approach to rethinking an essential part of the investment process primary research. I'm joined by Margot Polishuk, co-founder and president of ProSapient. Margot came from the expert networking industry and saw a chance to build something different, a platform that blends technology, scale, and speed without compromising on quality or trust. We'll dig into how ProSapient is using AI to streamline research workflows, what's changed in how PE firms engage with experts, and what's next in this relationship-driven space. Margot, welcome to the podcast.

SPEAKER_00

Hi, Steve. Thank you for having me.

SPEAKER_01

Let's get straight into it. I'd love to get a bit more of your founder story. If you could tell listeners a bit more about you, what you saw in the market that led to starting ProSapien.

SPEAKER_00

Sure, thank you. So my background is in expert networking. I worked at a different expert networking firm before founding ProSapient in 2017. I really enjoyed working in the industry. I love the industry. It's so dynamic, it's so fast-based. You're exposed to so many different topics, sectors, and subsectors. But I saw this industry that was really undertechnologized. And I saw the customers that were asking for expert intelligence that is precise and expert intelligence fast. And the offering, while useful, had very traditional processes attached to it, which meant that, you know, you know, getting hands on relevant expert intelligence took time for the customers. And also sometimes the matches to expertise weren't weren't exactly right. I actually tried to innovate at that other firm, but because there was so much demand for expert networking, the traditional firms didn't feel the need to do anything differently. And I thought, okay, well, that's their philosophy. But I think there is a way to do things differently by introducing technology into the into this sector. So I went my separate way. And we started, we Persapian has a philosophy of three pillars. The first pillar is customer. We always start with the customer. So from inception on day one, we started interviewing hundreds of users of expert networks to find out what solutions are already available to them, what do they love about the offering, and what's missing. And then we honed in on what's missing. The second pillar of philosophy is technology. So we built innovative technology, innovative tech features to address the needs of the customers. And then the third pillar is exceptional people. So we hired people that we thought were right for this culture so that we can provide a white gloved service.

SPEAKER_01

Thank you. I suppose I'd I'd want to go straight to what's missing. And look, this is your eight years plus into this journey. And so this isn't this isn't an outcome from the latest wave of AI. Um, and obviously it's come from you doing right in the middle of this, this let's call it problem to solve. So when you were talking to your experience and talking to those users of expert networks, what yeah, what was it that was missing?

SPEAKER_00

Yes, so at the time it was different to what it is today, as you correctly identified. At the time, Gen AI wasn't a thing. So back then, a couple of innovative things that we came up with were, well, first and foremost, clients wanted expert intelligence that is precise and they wanted it fast. So we had machine learning in the underbelly of our platform from the get-go so that we can identify uh and re-identify expert matches as fast as possible. And we had AI in the form that it was available there as well. And then we were the first expert network to introduce transcripts of expert calls. So that concept didn't exist before PerSapient. Now, of course, this is the industry standard. And then a year later, so that was year one, those were like the three key things we did. And then in the second year, we also introduced the expert messenger. So kind of uh the ability to interact with an expert on the PerSapient platform by just sort of through a chat function.

SPEAKER_01

Yeah, I can see, I can see how that that will uh has evolved potentially as well. I'm sure we'll we'll dig into that. Um so it wasn't necessarily that um anything was broken in the process, it was just that it could be enhanced greatly.

SPEAKER_00

Yeah, it was it was just look, it depends on the customer as well. PerSapient went after private equity firms and management consultancy firms working on behalf of the private equity firms when when completing the commercial due diligence. Um for for these types of customers, they want to move through that process really fast. So you can't take a traditional approach. You really need to think how do I get this customer to unique knowledge about this asset they're interested in yesterday. That wasn't what was happening in the industry. It was kind of very human-led. You know, let's have a look who's already on the database. Okay, there's a CEO that left six years ago. Nobody wants to speak to the CEO who left six years ago. Our customer wants to speak to the CEO who left six months ago, right? So you're gonna have to go into the market and you're gonna have to find a way to get that CEO onto the system really fast. So, how do you do this? You have to use technology and also you have to operate the culture in the organization, it has to be fast-paced. Right? You have to have that energy. You have to get this for the customer really fast so that they can have the edge on this deal versus their competition.

SPEAKER_01

Yeah. No, one thing that we're we're we're obviously coming up against now is is as we try to break traditional processes, people and culture comes into this, change management comes into this. Whilst I think that there's probably more appetite now, but but eight years or so eight years ago, was that something that, you know, to try and break a traditional process? Was that only really for those that were wanting to be innovative?

SPEAKER_00

Yeah, I look, I think um the benefit of starting your own company is that then you have a choice whom you hire and how you incentivize these people. So initially, what we frankly had to do is be so to really think about okay, we well initially we we we we were capital strapped like any kind of startup, right? So what do you have? You have equity. So we allocated pretty high proportion of equity to the first hires with an expectation that just like me and my co-founder, they will be working 24-7 with us and they will care the way we care. So that was kind of what we did. And of course, look, you also have to be selective. You mentioned, yeah, you know, I think before we started recording, you mentioned that um that certain people, there is there are people out there that do enjoy change, and you you can through the interviewing process identify people who are innovative thinkers, outside of the black box thinkers, want to kind of break the mold, want to create something new. So and you can select them, right? If if you're patient enough.

SPEAKER_01

Yeah, and that reflects on your um, I assume on your initial customers as well, is that you know that they had those similar sort of traits, is that uh curious enough and uh wanting to or seeing that that there was a a tech-enabled opportunity here to explore.

SPEAKER_00

Yeah, I think we there was an element of because as it happens, private equity investment professionals are entrepreneurs at heart. Because if you think about it, I mean they buy businesses, so they take keen interest in what different uh businesses do in different weird and wonderful subsectors. So when you come up with something new in the industry they feel familiarity with, we will receive actually really, really positively. So when you have two founders coming in, excited about this first platform that we built, you know, people felt excited with us. And they also felt that they were a part of contributing to it as well, because you know, as you remember, we kind of started with the customer. So they felt that their kind of input was then quickly taken on board and a feature popped up. And then they were naturally kind of they were naturally naturally felt, hey, that was their feature and they're gonna tell their colleagues about it and they're gonna try it.

SPEAKER_01

Yeah, I like that. I like that. And and look, I suppose officially a year into building my network and and in the research that I did, it was you know clear in that P space that you have to build credibility. And without that, it's it won't go too far. Obviously, there's too much at stake. But you know, a year in um and building that credibility, the access really changes and the and the the openness as well into how um you know how I can do things that's gonna help them, and as you say, that they're then part of the journey too. So I can see how that that works. Let's let's forward wine to today. It'd be great to hear how AI is fitting in in your model right now. And as I mentioned before, you know, considering that you started the business long before the recent generative AI wave.

SPEAKER_00

Yeah. So here I have to be a little bit careful without disclosing too much of the kind of confidential information around our innovations. I'll go relatively high level. Naturally per sapiented is a huge gen AI. We have a hefty generative AI product roadmap uh that we have partially already executed on and are excited to be seeing some of the first results of. Um, really, if if we were to break it down, but there's of course there's plenty we're doing uh in in the form of generative AI for productivity gains. So giving our associates tools to be faster, to be more precise in finding the right expert intelligence for our customers. Naturally, I think a lot of firms kind of are see, I think this is the area the productivity gains from Gen AI is the area that is already kind of nicely underway. So that's definitely an area synthesis of large amounts of data. Uh, again, that's something we uh we are doing. And again, I think other firms, um, not necessarily in our industry, but just generally speaking, are seeing productivity gains from applying AI to large sets of data or information of good quality to quickly uh kind of to be able to consume that information that's relevant really, really fast. And then we are building tools that our customers can actually touch and feel. So new products with Gen AI. That's a little bit more of an uncharted territory from what I'm seeing. So that's that's one. And another one is Gen AI is wonderful in in my mind, but it is also creating certain challenges. And one that we are seeing is it's creating a bit of a mess of kind of compliance and fraud. So we are building some tools to address that way ahead of time. And frankly, the more effective and powerful generative AI will become, the more of that issue people will see going forward.

SPEAKER_01

Yeah, yeah, I understand. Is there and I yeah, I'm like, I'm keen for you not you to you to share secrets, but just just imagining I I'm I'm there in a P firm and that there's a particular use case within this whole sort of workflow that you're helping to support. Could you just walk me through one aspect in a little bit more detail about how I might interact?

SPEAKER_00

With percipient. Yeah. So the the the majority of expert networking involvement happens at the stage of commercial due diligence when the customer has identified uh an asset of interest and they want to understand all the detail around this company, right? So they want to speak to uh the former of the a number of formers of the target company, they want to speak to competitors, they want to speak to loads and loads of customers, even run a survey of customers to just understand does this company have a USB? If they were to buy this company today, will they make their return in, you know, two, two, five years' time? So that's where the majority of expert networking kind of this is where we see the most of the expert networking involvement happen. However, there are other use cases as well. So, for example, at an early stage when a private equity investment professional is just exploring an asset, there are certain tools and offerings to get smart on a new asset that you're not certain about yet. Here again, private equity firms may want to come uh to you know, to may want to involve an expert network to help them get smart.

SPEAKER_01

Yeah, yeah. And are you able to take one element of that and just talk through then how ProSapien is, I suppose, making that far more efficient as a process?

SPEAKER_00

Yeah, um, so if we were to take the kind of the bread and butter of expert networking, commercial due diligence, we um we want with generative AI, there's generally kind of uh rebase what is going on. There is a rebasing of knowledge, if that's a bit a word. Kind of people can educate themselves on a company they're not familiar with pretty fast, especially with publicly available information. However, in private markets, I mean, there's just not enough information on private companies. So this is where our customers come to us. And here they want to move as fast as possible to consume kind of the general information and then to pivot into the unique uh information. So what we do, we use generative AI to escalate the general available information as fast as possible in moments in time. Then again, kind of with generic experts, within an hour, we were aiming to get this across the line for the customer. And then what we really focus on, hey, let's find our customer unique insights. And that can come either through by by finding unique and difficult to find niche expert or unlocking unique expertise that's locked in someone's mind.

SPEAKER_01

Yeah, yeah, okay, okay. Um, and when it comes to you mentioned compliance and and look, there's still going to be an element of human in the loop here. Yeah, what where where is that? Where is that human judgment still critical?

SPEAKER_00

That's an excellent question. And actually something that we talk about all the time. You know, we we with AI, and I think maybe a number of people you've spoken to already probably tell you the same thing. You kind of build something and then you deploy, you know, you ship that product. And then you look at it and you think, okay, well, what is the result? What is the accuracy? What is the speed? And then you pull back, you tweak, you redeploy. And then you go a number of times until you're very happy with the results you're getting out of the Gen AI tool that you built. It's it's achieving what is what it's looking to achieve. So at the minute, the majority of tools that we fully deployed are kind of internal and we have human in the loop on either side when liaising with any people, because we are, frankly, expert networks. They are people businesses. You're dealing with customers on one side and experts on the other side. And at the minute, I don't see huge appetite from people from deal purely with generative AI somewhere. So we have humans on on both sides at the minute to to talk to humans. But somewhere in between, we have quite a lot of gen AI that we use to basically speed things up and make things more accurate.

SPEAKER_01

Yeah, there's I mean, maybe we'll get to that near near the end in terms of what's next, but the opportunity to have almost um the amount of conversations that are had is that there could be some form of um, I suppose, um expert um that that isn't a human. But we'll maybe we'll get to that later on. Um with with with the the the um eight years that you've been in business, and obviously as we've discussed here, how generative AI is has added capabilities on top, um are you then seeing sort of a change in terms of the expectations of your customers in terms of what what you can do? Maybe that's in terms of turnaround times, etc. Are you sort of keeping up with the expectations or do you feel that they're yet to your customer base is yet to sort of really sort of take advantage of all the the tools that could be available?

SPEAKER_00

Yeah, actually, we're very lucky because our market is incredibly dynamic in terms of the shifting needs of the customer. This is why being in the market and talking to customers is so important. And this is why some of the traditional firms are struggling a bit because they're quite large and it's kind of difficult to innovate at pace. But we love that. So we love the ever-evolving kind of needs of customers because we're very edged and this is our ethos. We're a challenger. We love that. Let's, oh hey, there's a new need. Let's address it. How can we address it? And then we we kind of play around with different solutions until you know one solution lands, and we love being first to market with tools that solve the you know evolving needs.

SPEAKER_01

So, what are the patterns then you're seeing across the the more sort of digitally, digitally mature operationally sort of versus you know the operationally hands-on P firms? What is it that I suppose makes them more competitive or differentiated?

SPEAKER_00

So we actually so we're seeing a range in terms of how customers themselves are innovating. So that's a bit different. So their needs might be evolving really, really fast, but the pace at which they are innovating internally actually differs quite a bit. Earlier this year, we we ran a study to see the speed of a general eye adoption in consulting firms and private equity firms. And we saw that actually consulting firms overall are ahead, uh significantly ahead, and um private equity firms are relatively slow at adapting uh generative AI internally. Some firms are a bit more ahead, and other firms are, you know, are barely kind of exploring any generative AI tools, whether that being kind of external sort of third-party solutions or building something internally. You are we are seeing more of an emergence of um internal generative AI-oriented teams within private equity firms. So I suspect they will be doing more of that type of innovation going forward. But we are yet to see something happening at pace there.

SPEAKER_01

Yeah, no, I I yeah, I I well, you know, I can second that in the conversations that that I've had. Um they're taking a sort of um wanting to be within the pack, so they don't want to be falling behind. So maybe some want to be fast followers. There are some leaders out there as well, but yeah, generally it's just looking around. Are we still where we should be? Um, but obviously those that are starting to adopt are going to see you know uh greater returns you you'd imagine in the coming months and years. Let's just go back to this workflow around primary research. What is it that those firms could be doing right now, though? So rather than say jumping straight into with ProSapient, what is it that they could get themselves or at least be pro sapient ready, for example?

SPEAKER_00

There isn't, you know, we never sort of turn away a firm that hasn't built some sort of tech solutions internally. We have different tools available depending on the need of the customer. So if they don't have a get smart tool internally, we can offer that solution to them. But what we are seeing some of the sort of more tech forward firms doing is building uh sort of knowledge management platforms internally so that they can retain the knowledge that they gain and re escalate it. So, for example, going back to that, the transcripts of expert calls, you know, for some time, you know, they were just kind of lost in the ethos for the customers. So why not store these transcripts and then you know, you can. Apply Gen AI on top, and you can re-escalate this knowledge when you kick off a new project on a similar topic or on a related topic. So uh, and that's just one example. Of course, you can feed a lot more than just uh transcripts of expert calls into that. You could, you know, put uh diff different types of documents in old sanitized decks and so on. Of course, you have to be careful with compliance, so you may have to sanitize things, you may have to wait for some cool-off periods, but that's all probably you can navigate that, right? So I suppose what I'm saying, high level, you might want to think about how you store knowledge internally, and you might want to start applying AI to re-escalate that knowledge so that when you contact expert networks, whether that being ProSapient or another expert network, you can really focus on, you know, the knowledge that you don't yet have, or in the unique and freshest perspectives, right, on your company of interest.

SPEAKER_01

And have you seen different P companies draw different insights from those conversations? Like they have their own sort of flavor of the way that they use that. Is there anything that's proprietary in how they draw what they want to draw from those conversations, or even the questions they ask?

SPEAKER_00

Yeah. Well, that's an excellent question. And absolutely. Because if you think about it, so if we start with the kind of the point that there's just not enough information on private companies out there, right? So that's point one. Then okay, where do we find that information on private companies? Well, it's locked in people's heads, experts' heads who actually work in this industry. Okay, how do we unlock that information? You can ask a bunch of generic questions and never hear the right information. And you can ask very smart strategic questions and unlock a unique insight out of the same person that will give you a huge edge versus your competition. So, yes, absolutely. If you were to ask the right questions for the right expert, you could really get an edge.

SPEAKER_01

Yeah. And I suppose then that there's an opportunity with with with um with AI to build some sort of consistency around that as well. So, you know, it depends who's asking the questions, for example, and and the structure that you're um you're working towards. And and having in my uh previous career sort of a product manager is is you know, I've run a lot of research conversations, and if that's spread around lots of different people, you'll yeah, that's really the hard then to synthesize and pull that together in a consistent way. So I can see how that, you know, is a is a positive output. Let's just probably got a couple more questions for you. Um and as you know, you're reflecting on on the you know the founder's story up to now, what what do you think's been the most sort of challenging or surprising part of building the platform that you have in a space that's known so much for their sort of high-touch human-first nature?

SPEAKER_00

Okay, yeah, it's it's a it's a it's a good one. Um now, so something that is important to state up front is the people business. Expert networks require human intelligence. So I don't see at the minute how it could be a pure platform play. Because again, it can bring bringing it back to that point, you have customers who are humans, and you have experts who are humans. At the minute, they have no appetite in just talking to generative AI. And also, I would go a step further to say there is an element of mistrust towards AI at this stage, I think. So uh you will need to have a smart combination of human intelligence and artificial intelligence and also other technologies in the mix as well. It's not just about AI to get the customer to the right solution. So I don't believe in pure platform play. I believe in tech-enabled services in this industry.

SPEAKER_01

Do you and has that been a belief from the start, or has that been a learning curve for you? I mean, I know you obviously you did the job anyway, but did you see a different um path?

SPEAKER_00

That's been my belief from from the get-go. So my background in in in in in the prior firm, my last role was client-facing. So I am the customer first founder, and my belief has always been that you have to have a human in the mix. The types of customers we deal with want a white gloved service. There is very little appetite of clicking the buttons on some platform. The their desire is that you click the buttons for them and you have a dedicated team that understands what they do, what they're trying to achieve, their sector specialists, and that you will do everything you would provide that white glove service. Because frankly, if the customers wanted to click buttons, they would probably go about finding their own experts.

SPEAKER_01

Yeah, yeah, absolutely. Okay, we're not not heading towards um um that sort of s scenario, but but talking about paths and and not not necessarily revealing your roadmap. What is what's next then?

SPEAKER_00

Yeah, so again, have to be uh relatively high level here. Um and and look for us, um maybe maybe kind of stepping, taking kind of a step sideways from just the kind of the technology evolution and innovation. Um for us, a big uh area is geographical expansion. We are so our original first office was in London, and we started uh our growth journey in Europe, and now we're accelerating our growth in North America. That is a huge market for what we do, and we are grabbing market share really rapidly in that region, and we're very excited about the future there. There are also some other geographical regions where we see a lot of potential as well. So, geographical expansion is certainly on the cards. In terms of innovation, given the second pillar in the prosapient philosophy, technology-led innovation, naturally we have an innovation engine that is constantly turning. We are constantly kind of getting information out of the market to build new features, and we are constantly shipping product. We have a hefty generative AI roadmap naturally that we already delivering on, and we're really, really excited about generative AI and the future of generative AI.

SPEAKER_01

Is there, yes, it's very exciting, very exciting. Just to I suppose finally, and you're talking about expanding into new regions, is there any sort of cultural challenges to the way you're doing things? I mean, obviously you've the the human piece is still very integral, but is is there have you seen a uh any sort of differences of opinion on such a traditional process?

SPEAKER_00

From the customers or from the people higher?

SPEAKER_01

Yeah, from the from the from your customers, yeah.

SPEAKER_00

I think look, customers, different customers really, really there is a real kind of you won't hear the same answer from two different customers. And I think listening out for the different responses is part of what makes an effective challenger in the market because it's not one size fits all. And different firms have different ethers, different strategies, different approach, different tools available to them. So, you know, thinking about okay, what is the market I want to go after? And then let's go in and let's treat each customer as the unique kind of firm. Let's let's listen out to their opportunities and their challenges and let's join them on that journey and let's innovate with them. So that's that's kind of how how we approach this.

SPEAKER_01

Yeah, okay. Look, we've we have come to time, but I really enjoyed the conversation. So thank you, Marg, for your time. Uh, a really fascinating perspective on how primary research is evolving and why it can continue to play such a vital role in the in the deal life cycle that's tech enabled, but still very much uh human part of the loop there. For listeners wanting to learn more, you can check out prosapien.com and uh feel free to connect with Margot on LinkedIn. All all of those links will be in the in the the show notes as well in the podcast. That's it for for now. So yes, thank you, Margot, for your for your time. And uh we shall uh be back next time.

SPEAKER_00

Thank you very much for host for hosting me, Steve, and for some great questions.

SPEAKER_01

Thank you very much, Margot. Great answers too. Thank you.