Investing Billions - E388: The Future of Investing: AI, Expert Networks, and Information Alpha

Episode Date: June 11, 2026

What if the biggest edge in investing today isn't having more information—but knowing how to turn information into conviction? In this episode, I sit down with Matt Wells to discuss how AI is reshaping the investment process, why investors are drowning in data but starving for conviction, and where information alpha still exists in increasingly efficient markets. Matt explains the evolution of expert networks, how the best investors use expert calls and channel checks to build differentiated insights, and why qualitative information often drives quantitative outcomes. We also explore decision-grade AI, conviction building, private market diligence, and how the role of the analyst is changing in an AI-driven world.

Last time we chatted, you said that investors are drowning in data and starving for conviction. What did you mean by that? It goes beyond investors. I think all of us, right? If you're making a decision, buying a new car, anything, you have 20 tabs open, right? And even then you don't feel like you still have all of the information, even though there's more and more and more layers to dig through. It's the same thing from an investor's perspective, whether it's filings, expert calls, transcripts, news, market. market data, we've never had more data than we do right now to sift through. And it's not the data. That's the problem. It used to be, oh, well, if I can find this one piece of data, this one person to talk to, maybe I can have conviction. Now it's, I have too much data to sift through. How am I possibly going to get through all of it to find that conviction? And I think therein lies the problem.

Data isn't the thing. I've had this recurring, I guess, toxic relationship with AI where I ask a very important question. I upload a lot of data, and it gives me an answer. And I always double-check. What is your confidence on this? And it's always, I'm extremely confident. Here are these two edge cases where maybe it might be wrong. And maybe 50% of time, it's just completely wrong for anything that requires any kind of prognostication. Exactly right.

And it's pulling from all of that data. And I think what the troublesome thing is Gartner just released a study, I think it was back in Q1, that like the floor, minimum, between three and 25% sort of general AI error rates. That means the base case is that you're 3% wrong. If you're 3% wrong, hard to trust anything that coming out of it. And I think also you're not getting any real differentiated insights when you're talking about sort of generalized AI. And it's one thing if it's to me I'm researching a lot about sleep hacking, diets, all these things. I guess it gives me the wrong supplement. When I'm making an investment, that's pretty problematic. Yeah. And I think we're all using it for that.

recipes or whatever the case may be. And even in a work setting, I'm using it all the time for drafting or testing messaging, things like that. But when you're using it for something that I would consider high stakes, really you need the term that I use is like decision grade AI. And that's AI that's grounded in reliable source data, not just general training data. And I think that's where you can really draw conviction because that's when you can start to cite your sources. You can trace your sources and knowing where that actual information is coming from and having the confidence that the LM just didn't make it up per se is obviously a necessity in investing. That's the problem. The problem being that AI is, I would say, overconfident and also AI isn't

fully self-aware on when it's overconfident. What's the solution to this problem? So the solution to the problem, I think, starts with differentiated and trusted data, right? If we have If everybody has by and large access to the same information and information is no longer the edge, I think there's a bit of a nuance to that in that there is some information that does give you an edge, but it's that not generic information. It's not what everybody has access to, the K's and Q's, the filings, the transcript, the earnings calls. It's the real sort of ground level intelligence.

And whether that's coming to you through niche broker and sell side research, or things like expert calls where you can really get a flavor for like the messy reality, I call it. I think it starts there. So it starts with that differentiated data. And then it's layered on by how you apply AI to that. And can the AI that you're applying to that source and cite exactly what it's actually telling you so that you can be confident in where it's coming from so that you can use it in a

decision or in your investment committee or whatever the case made? I want to double-click specifically on expert calls. Before we go into why they're a source of competitive advantage, which I believe to be true, give me a slight history on expert calls and how it's evolved over the last 10, 20 years. Yeah, so expert calls have always been around in some way, shape, or form. I would argue that back in the day, maybe more than 20 years ago, investors' advantage could be, oh, hey, I knew a guy that knew a guy on the factory floor in Ohio that told me some piece of information that drove potentially how I acted or invested.

That's an expert call, right? Before that industry even existed. Then that was a bit more formalized where you can actually have somebody help you identify somebody to speak to you based on a topic that you wanted to diligence or go deeper on. That's evolved now to these pools of expert calls. We have this expert transcript library, right? 250,000 plus transcripts that, you know, even if you, do not conduct yourself a single expert call, you can tap into that expert knowledge that's

differentiated, again, that ground level intelligence, that signal. Or if you're not finding what you're looking for, you can identify the expert that's right for the topic that you want to discuss and go and interview them to try to get that intelligence that you may be lacking. I had, I believe it was episode 50, Mark Gerson from GLG, the kind of the OGs in the space. They did this product for hedge funds where they would connect hedge funds with experts and hedge funds would get information alpha from these conversations. You guys have made a pretty radical decision, which is you bring down the cost of the expert calls,

but in return, you give it out to the entire network. Talk to me about that decision. It's really to help populate that transcript library. I think our view is that an expert call is valuable, but a collection of expert calls then mapped against other market research and market data is really, where you can drive interesting insights and signals that you can act on. So for us, we allow our users to do those expert calls at costs.

And the trade-off is that they become part of the library. Obviously, there's an embargo period where that information is still exclusive and proprietary to you to act on. But after that period, it goes into the library where we believe it benefits everybody. And it's not just that you have access to that individual expert call or transcript. it's how you ask questions against it. It's how you interrogate it. So it's not the same for everybody, depending on your sophistication of what you actually do with it. Let's say I'm looking at the SpaceX IPO, and I want to make a decision on whether I should invest at the IPO. What's the best use case for

expert calls and how does investor go about utilizing it? It's a great illustrative example. Obviously not investment advice or anything like that. But in the case of, SpaceX and investors thinking about what assumptions they need to underwrite. And SpaceX being a late stage growth, but private company, there's not as much information as there would be on a public equity, obviously. So what an investor might do when trying to figure out at what level to invest in the SpaceX IPO is engage in a series of expert calls. And what does that actually do for them, right?

They're not trying to find out necessarily what margins are, what revenue is. or anything like that, they're trying to find out, again, that ground-level intelligence. So you might spin up an extra call with the CTO at a commercial satellite company. And maybe you're trying to understand, well, if SpaceX raises prices by 30% on their launch product, what are your competitive alternatives? Are you locked in with SpaceX? I understand the sort of pricing dynamics and supply demand. You might want to talk to a supplier of SpaceX, someone that machines, titanium,

pieces for rockets. And if SpaceX is saying they're going to do 100 launches next year, that's going to give you signal whether the actual supply chain can keep up with what SpaceX is publicly stating so that you ultimately understand whether what you're hearing as market narrative is actually feasible based on sort of the ecosystem that SpaceX. Said another way, SpaceX, now that's going to be a public company, publishes projections. We're going to do X amount of business in this business line. If you knew that there's space business or Starlink that customers were more bullish than maybe market sentiment or less bullish, you can now predict in theory, both in the short term and long term, how the stock might

trade. That's exactly right. Or at least that's the crystal ball that investors are, you know, trying to ascertain. And I think the composition of what management is stating publicly through earnings calls, et cetera, and what you're hearing on the ground and understanding if, you know, those two things are congruent or some things off is exactly where you start to get that signal. I like to think about it as this gathering a mosaic of information. It's kind of like you see those murder mysteries where they have all those pictures on the wall and everybody's like trying to connect all the dots. It's kind of what the best investors do.

And the most elite investors, especially today, they're not focused on purely quantitative data. Why? To the beginning of our conversation, everybody has. There's no more alpha in the quantitative. You need to have both quantitative and qualitative data. order to really generate alpha system. I love that analogy of the murder mystery board. I sometimes say it's like getting the 3D picture, right?

If you're talking to suppliers, customers, competitors, former employees, you can really start to see what that shape actually looks like before the light might shine on it. Perhaps this is extremely dumb question, but talk to me about the game theory on an expert call. Why is somebody incentivized to tell the truth? Is it just natural human behavior? I think it's natural human behavior, but I think it's fair that anybody that is taking a call or requesting a call, approaches it with some level of skepticism, which is why I don't think that anybody should ever make a trade or decide

if you're in private equity and you're doing diligence, you're trying to understand a space. A single expert call is not something that you should just trust wholeheartedly. I think the signal really comes, and I think that's where the expert transcript library is so valuable, when you can try to extract insights across a number of expert calls or go deep with a handful of experts to see where things start to line up. Because, of course, there's always the case that something maybe giving you,

maybe not even maliciously, but maybe giving you information that they weren't privy to or they think they know, but they don't actually. This is my pet peeve. It goes back to the AI. Maybe I have some trauma around it. But when somebody tells me something with extreme confidence and they're wrong, it's a big pet peeve of my. I think the beauty of it, especially with Alpha Sense,

is you can always map that what's said in that expert call trans, transcript with what's being said across other pieces of market research or news or filings or transcripts, sell-side research, things like that. So you can get a sense for, is this just an isolated data point that you're only hearing here? Is there some other signal in the market that's across all these different formats? Let's double-click on that, the meta-analysis of the expert calls. Stay on this example of a SpaceX IPO. How does one go about getting a meta-analysis? And what does that even mean in terms of expert calls? That's something that AI has actually been a real benefit around, right? Our system actually understands like tone and tenor and positive and negative sentiment.

So especially as you look at these calls over time, we're able to track how sentiment is tracking over time and almost give it a sentiment score, so to speak. So you'll see in those expert calls, some things are highlighted in green, some things are highlighted in red. That's actually tracking sentiment behind that call. And so valuable in just a one-off call to see what positive, what's negative, even more valuable over time, especially as you go into things like channel checks, which is like a derivative of expert calls. Double click on what a channel check is and what's the best practice. You could think of them as one and the same or a lot of people do, especially a lot of junior analysts, right? Like, what's the difference between an expert call and a channel check? And I think

of it as really depth versus breath. So one single expert call if I'm doing diligence on a specific company or even a specific customer and I want to go really deep with someone that might be one of the only few people in the world that know everything about that one subject. That's really an expert call. A channel check is mainly used by investors ahead of earnings to understand how everything is sort of lining up in that company or industry's ecosystem. So speaking to suppliers, distributors, understanding where orders may be tracking, inventory is growing or shrinking. And getting that picture, not just in one call, but in a higher volume of much shorter calls

to get that signal ahead of earnings is really how you should think about channel. I've been talking to a lot of growth investors. And I had individuals like Lucas from who runs the CO2 growth fund, Matt Wiltire from Wellington, as well as Paris Heyman from JP Morgan. These are massive crossover funds. if I was one of those three people or somebody else with a massive amount of capital, what's the best practice in terms of how much do you deploy against an IPO? And what's the framework to think about how much you should deploy into expert network calls relative to your check size?

I don't know that there's any sort of rule of thumb, right, of like, hey, don't ever spend more than X percent. What I would do is I would start there, right? Because I think the numbers that drive the models and the quantitative aspects, they all start as qualitative signals. So you may have your base case. You may have your model. But then you need to go and test those assumptions. Expert calls, channel checks.

Those are bar none the best way to actually go and test them and not test them in a way that I think the best investors don't go out there and look for confirmation bias. Hey, tell me why I'm right. I think that's like natural human nature. I think the best investors go out there and say, tell me what I'm missing. Tell me why I'm wrong. Disprove this thesis that I have. In that way, you're really able to either drive conviction even higher or save yourself from something that could be a really huge mistake.

That's where I would start. And I would spend a lot of time, effort, dollars there. Historically, that's taken a lot of human capital and even a lot of capital capital. I think with AI and the way that these have been democracy. a bit and the way that you can now run concurrent expert calls without even actually speaking to the expert allows you to really amass signal a lot quicker. That's the exact reason why I'm asking, where before, if you're Lucas at Koot or Matt at Wellington, you might have time to do 10 expert calls.

So just pick a number because you're limited in terms of how much time. Now you could have AI do 10 expert calls for you and specifically AI ask the questions. but you could also have AI do 100 calls. I want to ask you specifically, like, what's some extreme cases where you've seen people spend? Is anyone spending $10,000, $100,000 on a single investment? The biggest funds in the world, the household names, historically, especially around channel checks, they spend millions of dollars per year on that process. And that used to be something that they would potentially outsource.

Now, with AI and the way that we're approaching channel checks and democratizing that a bit, A smaller fund can have the same access to what the larger funds historically had from a channel checks perspective, which was a huge advantage for them. And again, I spent a lot of money. And you said something really interesting. You said that all quantitative data starts qualitatively. Double click on that. What do you mean by that?

If you think about the drivers in a model, whether it is margins or assumptions on regulatory risk, let's say, all that starts as what they're finding out for. from individual stakeholders in that market, right? Beyond just what the company wants you to know, beyond the highly curated earnings calls, when you are actually speaking to a supplier, a former employee, a competitor, and finding out these nuggets of,

in the enterprise software space, implementation is moving from three months to six months. And that's gonna delay how revenue is recognized potentially. That changes an assumption in a model and ultimately changes what the outcome of that investment going to look like. So while the model itself hugely important, usually all of that starts off as some sort of qualitative signal that eventually shows up in an earnings report or something like that as a quantitative metric. That's such a good point. Even their earnings projections, the revenue

projections, where do those come from? Those come from analysts. Where does that come from? That comes from qualitative information. Now, it looks like quantitative because it ends up with a number and then you do an average of the numbers. But ultimately, if it's not coming from the company and it's not a number from the company is by definition starts qualitative. That's absolutely right. And that's the same process that all those analysts are going through. You just don't really think about all the work behind what just shows up as a number in a report somewhere. There's deep, deep, deep analysis generally through expert calls and other ground level intelligence that is baked into that. I know you think a lot about the future and this post-AI future. Where does value accrue

when there's so much quantitative information out there. Is it all in qualitative information? Is it around process? Is it around relationship? Support for today's episode comes from Square. The all-in-one way for business owners to take payments, book appointments, manage staff, and keep everything running in one place. Whether you're selling lattes, cutting hair, running boutique, or managing a service business,

Square helps you run your business without running yourself into the ground. It's actually thinking about this the other day when I stopped by a local cafe here. they use Square and everything just works. Check out is fast, receipts are instant, sometimes I even get loyalty rewards automatically. There's something about businesses that use Square. They just feel more put together. The experience is smoother for them and it's smoother for me as a customer.

Square makes it easy to sell wherever your customers are, in store, online, on your phone, or even at pop-ups, and everything stays synced in real time. You could track sales, manage inventory, book appointments, and see reports instantly whether you're in the shop or on the go. and when you make a sale, you don't have to wait days to get paid. Square gives you fast access to your earnings through Square checking. They also have built-in tools like loyalty and marketing.

To your best customers, keep coming back. And right now, you can get up to $200 off Square hardware when you sign up at Square.com slash go slash how I invest. That's SQU-A-R-E.com slash go slash how I invest. With Square, you get all the tools to run your business with none of the contracts or complexity. Run your business smarter to Square. Get started today. Managing risk for your business may be complicated,

but your relationship with your insurance broker doesn't have to be. NFP and Aon company can help you navigate insurance markets and negotiate with carriers to build the right coverage for your business, helping you turn your risks into leverage. NFPs advisors are total business partners who help you protect your business and connect you with solutions to your toughest financial and workforce challenges. Whether your goal is to manage risk more effectively, attract top talent, empower your workforce,

or grow your legacy, NFP is ready to help you succeed. Visit nfp.com slash how I invest today to unlock your full potential. Managing risk for your business may be complicated, but your relationship with your insurance broker doesn't have to be. NFP and Aon company can help you navigate insurance markets and negotiate with carriers to build the right coverage for your business, helping you turn your risks into leverage. NFPs advisors are total business partners who help you protect your business

and connect you with solutions to your toughest financial and workforce challenges. Whether your goal is to manage risk more effectively, attract top talent, empower your workforce, or grow your legacy, NFP is ready to help you succeed. Visit nfp.com slash how I invest today to unlock your full potential. What do you think is going to happen in the future? I think qualitative information is a huge driver of that. And obviously, we've talked a lot about expert calls and channel checks and things like that. But I don't think it's all, like all the Alpha is just trapped in that data set alone. I think obviously AI has unlocked huge potential to work through that massive amount of data that we talked about at the beginning. But I think it's

really how all of those pieces come together in a way that one analyst or a huge team of analysts could never possibly synthesize that actually gives us a clearer picture or helps us find those needles in a haystack that is compounding in size. That's where everything is going. And building tools that allow you to tap into that, that map to your workflows and that run almost autonomously, that's where I think everything's going. Going back to the SpaceX IPO example, how should investors be using tools and be using both qualitative and quantitative data to come to ultimately a decision of do I buy or not at this price? There's so many great tools out there from an AI perspective, the big sort of like watch out is using the right tool for the right

task. And we talked about before recipes, et cetera. I think that's really true. And I think obviously general purpose LLMs have an amazing place in all of this. That's what powers everything underlying Alpha Sense. We're using the best models for the best task. But I think you need to make sure that that that model is grounded in, you know, trusted sources. And oftentimes that's proprietary sources, whether it's the expert calls, a real premium sources, broker research, et cetera. And then across filings and earnings call transcripts and qualitative data like that. But ensuring that LLM, that may be running a specific workflow for you,

is grounded in those sources. And then also able to specifically cite those sources. So you can see exactly what sentence of what report that can. came from already is massively important to actually making any of these decisions. If you don't have that, you're really just kind of playing with fire. It's one of the big things I think is massively underpriced and undervalue in the market, which is the rootedness of your conviction for your friends in 2010. I told you to buy Bitcoin at $10 a Bitcoin.

And now it goes up 16 times. You're extremely happy. Now next year it goes down 20%. Now it's at 120. What are you very likely to do if you have no real conviction on Bitcoin? Yeah. If you have no real conviction on Bitcoin, you're flying blind. You're going to sell it because you're going to lock in that 12x.

You're now fear-based because it's gone down that 20%. So you've mentally lost 20% off the 16x versus somebody else who has a very sharp. And what's the opportunity costs there? Today, Bitcoin's at $80,000. So somebody that has very high conviction, getting ready for my interview with Anthony Pompliano, and he's famously, and this is publicly verified, he doesn't sell his Bitcoin ever. He doesn't care about the day-to-day. He has this long-term thesis on it.

And whether he's right or not, that rootedness allows him to be so much more anti-fragile than every other investor. A lot of retail investors, more than institutional investors, but they want to know when to buy. Where they should be really asking about is the thesis. Why should I hold? What's your thesis? When should I sell? What should I be looking for when I sell?

These are kind of the unsexy, unloved things that differentiate the good from the grade investors. Conviction allows you to whether that's storm, right? Conviction allows you to know when to double down on your thesis when the market is reacting differently. And constantly testing that conviction and testing that thesis in a way that's not just looking for confirmation, but that's looking for the holes in it is what a lot of. allows you to drive great returns ultimately, whether that's in the public markets or even in the private markets where there's even less information to do it. It even goes down to the LP layer, some of the most elite LPs, and there's not many of them

that do this, but when the market's down, they will be ready for that time, period, because obviously it always happens. So you should never be surprised by that. And ahead of that, they would have built relationships with their top managers. They would have built conviction in their managers. Usually the conviction is not on the individual level. On the individual investment level, it's on the manager level. And when there's blood on the streets, they're able to capture some real gains. It goes back to that saying there's decades where nothing happens in weeks when decades happens. Sometimes some of the best investment opportunities come when everybody is running around like a chicken with their head cut off and there's this once in a generation opportunity to

deploy capital. Being prepared for that and playing out all of those scenarios so that when to act and you can act with conviction, that's what everybody's looking for. Going back to how we started the conversation, there's so much data and people are starving for conviction, not for information. At the same time, AI could be your friend. It could consolidate data. It could synthesize data. What's the best practices of utilizing AI in terms of all the data that's out there? And how does that help make better investments? What we talked about in the beginning, it's not lack of information. In fact, there's information overload and everything that we just talked about with expert calls and channel checks has just now added even more information.

So what's the most natural thing to do? Well, go have the machines read it, right? Apply AI. That is obviously what everybody's doing, but ensuring that you're doing it in the right way, again, through AI that's plugged into those actual sources of intelligence that are grounded and you're able to cite from them directly is much different than uploading some of that to a horizontal LLM, which if it's not actually getting the exact fact that it needs, is subject to going out and potentially making it up

or grabbing it from Reddit or grabbing the wrong time period into an output that sounds great, looks even better. And especially if you don't have a ton of experience in that market or if you're a junior analyst and you're reading through that output, very easily to get conviction in something that can be flatly wrong. Making sure that you're using grounded AI that's rooted in that domain context is vital if you're going to use AI against any of those sources. I was speaking to one of your colleagues and he mentioned that the future analyst is going to go

from an analyst to an architect or from an analyst to a prompt engineer. How do you? How do you? do you see that playing out? And are you seeing the early stages of that already? Absolutely seeing the early stages of that already. I think you're seeing a lot of the more senior deal makers or senior portfolio managers kind of do themselves now what they used to ask an associate or an analyst for because it's so easy to access that information. The real future value in an analyst is somebody that can understand how to automate a lot of these tasks, how to figure out that agentic workflow that would take them working multiple nights overnight to actually get that data or drive that, get those insights. Now they can actually do absolutely on their own

and let those run for a period of time to get that versus like them actually sitting at their desk. And they can do multiple strings of these at a time across multiple names or multiple potential investments. So I think the role of the analyst is absolutely fundamentally changed. It's going to look different in a month from now than it does. I have a bit of a contrarian thesis. Obviously, AI is very difficult to predict five years from now. But in the midterm, I think this goes back to this truthfulness and believability of models. I think the role of the analyst is going to be about veracity, meaning how do you confirm the data?

How do you make sure that all the data, all the data sources are good? How do you double check? How do you, I guess, channel check the data itself and make sure that what you're feeding up to your VP, or your managing director, that is not only the right data, but that's also packaged. How do you make that person more efficient? And I think you might be taking away the data crunching layer of it, creating the spreadsheets, creating the kind of the source code. But I think analysis will still be extremely useful.

And I think it's going to be an extremely useful skill set at least for another three to five years. I agree with that. But I also think that's why AlphaSense users love. of Alpha Sense because while you still always want to operate with that sort of trust but verify mentality, because we provide in our Gen AI answers or our outputs like slides and models, etc., every single fact is cited right back to that source document and highlighted the sentence where it came from, that's not how it works with general purpose LLMs. So I think analysts will naturally spend a lot of their time still verifying things.

But we basically do 99% of that for them. And ultimately, our theory is where you used to spend 80% of your time doing the data analysis, building the ICEMMO or building the pitch deck if you're a banker. And then 20% of the time actually making the decision, do we invest? Do we move capital here or there? We're really going to flip that, right? And you're going to spend 20% of your time doing the research, getting everything in order, in order to make that decision where you can now spend more time testing that thesis, sparring with your

partners, understanding, okay, do we actually do this and why? Despite devil's advocate, you referenced earlier, LMs have 3 to 25% error rate. And obviously, that could be catastrophic in big decisions and big investments. What is Alpha Senses error rate? Because our LLM, well, we use all of the best LLMs out there. We don't have an LLM, right? We're using Claude, where it makes sense. We're using chat chitbd where it makes sense.

We're using Gemini where it makes sense across different functions in our platform, whether it's building slides or asking a typical Gen AI question or deep research. Because that is grounded in our content library, as well as any data that the user connects to it in terms of like your intellectual property, your I see memos, your slides, your market research. That drastically reduces the potential. for things like hallucinations. So I think you're really, it is a bit of apples and oranges

when that 3 to 25% is really because the model could be just filling in the blank making things up. It doesn't want to say nothing. It doesn't want to sound wrong. It's very insecure. It's very insecure. It always not to sound right. But then when you say like, ooh, I don't think that's right,

then it likes to tell you that. Which is interesting because it's not game theory optimal. So if you're an analyst and your MD comes to you with a question, and you say, yes, this is the, answer and I'm extremely confident on it. Sure, that might make the MD happy for the next 24 hours, but if you're wrong, that could hurt you for years. Right. So it's interesting that the models have adapted this strategy. Yeah. And I think that's exactly why you can't just show up into investment committee and say, well, why is this so? Why are you recommending this? Oh, because AI said so.

And I really confirmed it because it gave me a conviction after asking a few questions. Like, with Alpha Sense, absolutely sure, because you have that cited context from those premium sources. With the general LLM, I think that's a little more dicey. Strategically, Alpha Sense, you guys have really focused on the public markets. In many ways, public investors have been the first adopters for technologies like expert networks because that edge is just so much smaller in the public markets. You really need every last piece of information alpha. Today, you guys are focused on the private markets.

How does that differ from the public markets when it comes to research? I wouldn't say we're heavily focused on private markets versus public markets necessarily. I do think you're absolutely right in sort of public markets historically is who adopted platforms like ours first because that edge is so small. But I would argue that in private markets, that information asymmetry is even more of an issue, right? with public markets, at least you have sort of this consistent reporting. There's this data out there that you can rely on. In the private markets, there's such an absence of data or data is so scattered that you really need to fill in much bigger blanks.

And that's where a platform like ours has been highly, highly useful. If you think of the case where you're a private equity investor and you just get access to a data room, well, now with a platform like AlphSense, you can integrate that data. data room directly and start to ask these sort of market intelligence, market context questions against what's actually in that data room to really fill in a lot of the gaps that you might have that otherwise you were either flying blind or you had to go do a lot of manual hours of research to figure out. Matt, a lot of people might know you were a serial entrepreneur prior to joining Alpha Sense. If you could go back to 2006 when you had just graduated college, what is one

piece of timeless advice you'd give a younger version of yourself? Yeah, it's funny. And this isn't even from my entrepreneurship days. This is sort of right before that when I was a banker. And we had this senior managing director, chairman, almost a figurehead who was like a legendary banker back in his day. He was part of the breakup of the bells. He would tell us stories about flying on the Concord and meeting the queen of England. And we would go into pitches. And before we'd go to the pitches, he'd always say, don't forget to ask for the cookie. Like, what are you talking about? It's like, I don't know what this guy's saying. And I'd say, if you don't ask for the cookie, you're never going to get the cookie. And what he was really saying was, you put all this work into this pitch. When you're in there with

the client or potential client, ask for the mandate. Go that last step that a lot of people don't take and actually close. And I think that's true whether you're in investment banking. You're trying to win a mandate, whether you are negotiating a new job offer or whether you're running for local political office and you're trying to ask people for endorsements, a lot of people don't always remember to ask for that close. And ultimately, that's what he meant by remember to ask for the cookie. Said slightly differently, in-person meetings are incredibly valuable. Great for sales, also great for resolving conflict. And the best time to ever ask for that sale is in person. That is the opportunity. The buying window is open, quite literally because you've made the pitch, but also because

you're physically there. And the highest likelihood that he or she will ever say yes will be at that window. 100%. Never underestimate the value of just genuine human connection. I mean, it even works at the most extreme level. You have these political delegations. So you'll have a king or a president visit another country and you think, why is that necessary? Why can't they just do kind of jump on a zoom and come up with a deal? As technologically advances we become, that in-person meeting, that one to one in-person meeting, whether it's with a GP and an LP, whether it's between two presidents, is just so powerful. It's so true.

I don't think that will ever go away, or at least I don't want to live in a world where that does go away because there's absolutely no replacement for that connection and that rapport building. That's one of the most positive futures post-AI is if we're not spending all the time data crunching, what are we going to do? And one of the versions of that world is we're spending more time in person. I really hope that's how it's. realize this. Well, Matt, it's been great to have you here in person and this absolute masterclass

on expert interviews on how to get the most out of information alpha. Thanks so much for jumping on. Thanks, David.