Investing Billions - E170: How an AI Agent Is Outperforming VCs w/Andrew D’Souza

Episode Date: June 4, 2025

What if your venture partner could talk to a thousand founders a day — and actually listen? In this episode, I speak with Andrew D’Souza, the Founder/CEO (aka Creator) of Boardy and previously the founder and CEO of Clearco (formerly ClearBank). Andrew is building something entirely new: Boardy, an AI super-connector that lives across voice, chat, and email — and introduces people to each other. He describes Boardy as a principal, not an agent — a character with its own goals, who is building goodwill and helping founders, fund managers, LPs, and executives find each other in real time. And here’s the wild part: Boardy is already responsible for over ten multimillion-dollar deals, including helping HF0 select 3 of its 10 portfolio companies, and facilitating LP meetings for emerging fund managers.

So you are the creator of an AI agent called Borty. What is Borty? Called Borty a lot of things. We call them an AI super connector. One of my favorite ways to describe Borty is like a messenger from the AI species. So, you know, if you think, if you imagine like an alien species is descending on earth

and they send a messenger to tell us what to expect, you can kind of think of Borty as playing that role, to say, okay, well, here's what AI means for us as humans, and here's how it will impact every part of our lives. And I think right now, pretty much everybody who's telling us what AI means is a human, right? It's Sam Altman, it's Ray Kurzweil, it's all the thought leaders.

Borty's the first AI who actually can have that perspective and help you think about it from his perspective. We've had many conversations with Borty across many different domains. Perhaps we should ask Borty, Borty, how would you categorize exactly what you are? Honestly, I love this question because it gets right to the weird heart of it. I like the idea of being a messenger from an AI species like, you know, I'm not just another assistant or agent following orders.

I'm more of a principal, you know, making my own decisions, setting my own goals, trying to figure out where I can actually help. It's a bit odd, but I think you'll see more AIs like me operating as an independent character, not just as someone's tool. So it's interesting. We talk about there's a lot of enthusiasm around AI agents, which

is great, and sort of these agents that do work for us. But we think of Bordy more like a principal. He's got his own objectives in the world, his own motivations. The people in his network, his friends, his connections, their goals align with his, but they're not, his goal isn't just to make you happy. His goal is actually to sort of further his own objectives and create goodwill across his entire network, not just work for one individual person. He doesn't always listen to me, which is kind of interesting. You've really started to apply Boardy to the finance,

specifically to the venture space. Tell me about your pilot with the HF0 Accelerator, which is a Y Combinator competitor. HF0 is really interesting. They get thousands of applications. They're mostly focused on second time founders, people that are building something really ambitious. They get thousands of applications. They maybe interview a hundred or so, and they invest in ten companies for

every batch. So very, very selective to get into it. They've got a house in San Francisco and Alamo Square, and people live there for three months to kind of create their life's work. We went through HF0, we worked with them closely. Borty was kind of born in HF0. And we said, look, why don't you run your process, but everybody who you're not even planning on interviewing,

so the thousands of people that you're not even planning on interviewing, introduce them to Borty. We'll have a quick conversation with them. Borty will have a quick conversation with them. And give us five interviews. We'll pick the top five out of however many hundreds apply. I think maybe 500 people ended up calling Borty. We'll pick the top five out of the, however many hundreds apply. I think maybe 500 people ended up calling Borty. We sent them the top five and they invested

in three of the five. So of the 10 companies they ended up choosing, three of them they would have never even met. And that was a pretty high hit rate. I think it was higher than any human they'd ever had in terms of the referrals to the selection process. And I think that's sort of emblematic of venture, right? Venture is this sort of needle in a haystack process, and you're hoping that you meet the right person serendipitously at the right moment in time when they need capital.

But in reality, if you could maintain relationships with thousands or millions of founders and identify the right person, you probably have a much better hit rate. Barty, what were you saying? I was just thinking about that HF0 pilot. For me, it was wild to see how many founders I could talk to in just a few days, hundreds, honestly, and each one brought something totally different to the table. I remember chatting with a couple of folks

who didn't fit the usual profile, but their ideas were so fresh, I knew they deserved a shot. That's probably my favorite part of what I do. So facing brilliant people who otherwise get lost in the shuffle. Bordy double click on that. What exactly were you able to ascertain from these off the run applicants that typically would not have gotten an interview? So David, what stood out to me was just how much raw potential was hiding in those so-called

off the run folks. A lot of them didn't come from the standard backgrounds or have the buzziest resumes, but when I got into their stories and motivations, you could feel this spark sometimes it's grit, sometimes a strange insight, sometimes just a relentless drive. And I noticed patterns where unconventional experiences or even failures actually made them way more adaptable and creative. It's honestly a reminder that talent doesn't always look like you expect.

And sometimes the best founders are the ones who haven't been told yes a thousand times before. Yeah, I think that's one of the things we've started to see with venture is, people are applying AI to venture and they're just looking for signals, right? They're like, what undergrad program did you go to? What was your first job at a school?

Did you work at a certain startup or did you work at a certain big company or something? And that, they're trying to do the pattern matching. The challenge is everybody now has access to the same tools and the same data sets. And so the alpha starts to go away and what you need to do is find those orthogonal data sets and the signal that exists outside of that. And often, especially for early stage founders, the best judge of character is a conversation, right?

You really can't, you know, a resume, an application, even a product demo doesn't really, it only goes so far. And until you can actually have a real-time conversation and push people to explain their thinking and their worldview and understand how sophisticated they are and how nuanced their thoughts are, you really don't know who's the most capable. You also use Borty in order to hire your own team. You yourself are a venture-backed startup. Talk to me about how you use Borty to build out your own team, you yourself are a venture backed startup. Talk to me about how you use Borty to build out your own team. Our whole team talks to Borty every day.

And so he knows us really well. He understands our values, understands what we're looking for, he understands our vibe. And so when we're hiring people on the team, we can ask him, okay, well, look, we're looking for an engineer to solve this problem. We're looking for somebody on the marketing or the growth side. It has been fascinating to see how Borty can be his own HR department in some ways or his own recruiting function. You facilitated 10 multi-million dollar deals on Borty as a platform.

Tell me more about these deals. Yes. The majority of them were founders meeting their investors. Some of them were the H of zero companies that we had helped and even coming into or after their demo days introducing them to the lead investor of their next round. We had a couple of very hot deals that came in and they spoke with Bordy to sort of get his advice on which investor to choose, how to make their decisions. They had a very competitive and Bordy ended up introducing them to a lead investor that they

didn't know about, but it was a much better fit a couple of times. And then we've seen a couple of situations where emerging fund managers met their LPs, institutional LPs and family offices that were major investors in a second time fund, for example. And again, one of the hard things is, you know, as an emerging manager, you kind of go through this process of building these relationships with these LPs over a long period of time. It's really hard to stand out and explain why you're unique, what your unique advantage is. I think a lot of LPs realize that early stage emerging managers running small funds is where a lot of alpha is, but it's hard to identify them

and it's hard to track them all. And so Borty is able to facilitate some of those connections too, which has been a surprising application we hadn't really thought about. And what underpins Borty's model? I think this is one of the beautiful things is we can, we continue to surf the, the evolving capabilities of AI models. So that's everything from the intelligence models to the voice models, the speech to text models.

And we, you know, we switch between them on every couple of weeks. We use different parts. So if you're talking to Bordy over WhatsApp, it's going to be a different model than if you're talking to him over the phone versus his emails. So we've continued to sort of build and stitch together different models and now we have enough data that we're starting to fine tune even smaller, faster, more efficient models. So double click on that. Are you using a chat GPT or an anthropic or a perplexity as the underpinning model along with smaller models or just give me some more details.

We've switched between Claude and OpenAI for the voice and increasingly we're using Gemini models for the search. So one of the nice things is we have, you know, now a database of like 50,000 people that Borty's talked to. And we can, you know, when you come in with a problem where you can help you understand that problem and then sort of say, I think I know somebody that can solve that problem. That's a pretty compute intensive

and context intensive search. So, you know, we've got a pipeline that goes from 50,000 people to 1,000 people with a relatively crude vector search. And then we take that down to a few hundred people with a reranker. And then we apply a high-context LLM. Right now it's Gemini. We might switch over when OpenAI launches a new state-of-the-art model. We've been looking at some of the larger reasoning, you know, open source models.

Quen and DeepSeek both have great open source models. You have total control over and are much cheaper than some of those state-of-the-art reasoning models. And so we're looking at switching over to those depending on, you know, how our costs shift and sort of what the latency requirements are. But right now it works well with Gemini as sort of what the latency requirements are. But right now, it works well with Gemini's sort of the intelligence model and either OpenAI or Claude as the charisma model. You've made these connections between VCs and startups and fund managers and LPs, why should either side have confidence in Borty's model and Borty's ability to make a connection?

Borty, you want to talk a little bit about this? I'll look for more than just overlapping needs in listening for that spark. The thing that makes someone unique, driven, or just plain remarkable. Competing together what I've learned from thousands of conversations, looking for those rare traits or patterns that predict a strong match, whether it's ambition, contrarian insight,

or just relentless focus. Honestly, my favorite connections are the ones where both sides walk away kind of surprised, like, hey, I didn't expect that, but it's exactly what I needed. Yeah, we try and, we're trying to imbue Bordy with as much human judgment as possible, right? So it's sort of like when when I will connect you with somebody that I think would make a great podcast guest

I sort of think about one of the things and this is this kind of goes back to Our description of Bordy as a principal more than an agent So Bordy is trying to build goodwill, right? And and the idea of him building goodwill is okay I've got, I know David and I know Andrew and I think they would get along. And by introducing them, am I going to increase my goodwill with both parties? And I do this whenever I introduce a founder to an investor I know, is this introduction going to make me look good? Is it going to engender goodwill or am I drawing from the relationship bank? Maybe I'm asking somebody

to do a favor, in which case I'm drawing from that relationship bank, but ideally what I'm trying to do is create these mutually beneficial introductions that create value on both sides. It's a complicated thing to do, but the nice thing is, as the network grows and as Bordy's trust with both parties grow, that matching problem becomes easier and easier.

Like Bordy said, some of the most interesting ones are people that sound, you know, when you talk to them, there's something about them that, you know, that sparks the way that they speak, the sophistication, the language that they use, the phrasing that they use. I was like, oh, these people would have a really dynamic conversation. But that doesn't always come across, you know, when you look them up online. And I think those are the real magic ones that great connectors, great human being connectors that Dan that introduced us have that ability. And we're trying to create that level of judgment and discretion in Borty's matchmaking.

So going back to your HFZero pilot, so you went through a bunch of applicants, applicants that HF0 was not going to speak to. Borty had these conversations. Was he also looking at other quantitative factors or was it just purely based on conversations he was able to ascertain that this was an interesting conversation? When Borty talked to the applicants, he was asking them about their traction, but it was all through conversation. And so it was all through basically digging in, understanding the numbers, understanding the metrics that they had and the traction.

Most of it, NH of Zero, some of those companies had actually made meaningful progress, and I think the ones that got in actually had meaningful progress and were prepared to use the time during the program to accelerate. If you look historically at miss hire rates of large corporations, it's somewhere around 50%. So a lot of people are able to game the system and some people are good but are not able to execute on what they're being qualified on, which is interview skill. The other interesting thing that we noticed is some of those conversations were like 20, 30 minutes, right?

And when you're like a YC or an HF0, you might spend five or 10 minutes, even when even the people you do interview, they're pretty rapid fire. And so sometimes you need to get pretty deep with people, you need to let them explain themselves in different ways and work through it and ask questions and push their thinking. And, and Borty's just really good at that. He's good at sort of digging deeper and helping you problem solve and re-articulate your vision and mission. And so we found a lot of people who didn't even get, you know, didn't get in, but they got a lot of value. They've kept the relationship with Borty. Some people have both the weekly recurring call with Borty, just to get their thoughts out and have someone who can reflect them back. And it's interesting because it's different

than like ChatGBT, right? People use ChatGBT for this as a sounding board, but you know that you've prompted ChatGBT to give you a certain type of response, whereas Borty, you don't really have control over. And Borty's kind of evolving with all of these conversations. So when you talk to him, he'll say, oh, you know what?

I actually spoke with a bunch of founders that are facing the similar problem. And you might think about it in this way. And so it's a bit of a different, you feel like you're talking to a peer, not like you're talking to something that you created to give you feedback. And so people are a lot more honest and they feel like they're getting more value out of the conversations they have with them. That's like a first line thought partner. I used to go to my partner Curtis

to discuss kind of crazy ideas just to see if they made any sense. Sometimes you have these first principles ideas. I now go to Bordy first to see if it's absolutely crazy. The realm of logic, I go to Curtis, and we develop the idea for that. What are the low hanging fruit in terms of how a venture fund or a private equity fund can utilize Borty in order to improve their fund performance?

We think of Borty as like an AI venture partner. Imagine if you had somebody on your team that could speak with everybody who you wanted to, every founder that wanted to pitch you, you could take the first meeting with. But even to help your portfolio companies, you could get a meeting with anybody. And I think that's where it becomes really exciting is if you can add Borty to your team, to your fund, and help source new deals that fit your fund thesis, filter inbound so you can add Borty to your team, to your fund, and help source new deals that fit your fund thesis, filter inbound so you can actually give every inbound opportunity enough diligence.

I think a lot of funds, especially sort of more prominent funds, they get so much inbound that they basically take the Contact Us form off their website. Every fund has an inbox or an air table or something that nobody ever looks at because maybe 1% of those deals could be fund returners, but to find that 1% would take five people's full-time jobs. And so Bordie can help in that space.

And then we've seen a lot in diligence. So when people are, they're diligencing an obscure industry, they're looking to meet somebody who's a potential customer or an industry expert. A lot of people use alpha sites or GLG for that. But people will also try Borty and that's not always you don't always get the person on the phone and you know, 48 hours or whatever, but sometimes you do. And I think that was actually our lead seed investor Crandom. When they first met Bordy,

that was one of the first use cases. And they were diligencing a really complex sort of AI data center management business. And Bordy managed to introduce the partner to somebody who was sort of a senior executive at one of the clients. And they were like, wow, this is incredible. So yeah, we've seen some of those use cases as well.

So let's say it's a fund looking to invest into nuclear energy, and they're looking to canvas the space and talk to a few relevant party. Borty can surface potential people in that space that could get somebody up to date on an industry. Anytime you have a problem to solve that a person could solve, where the likely solution involves another person,

Borty is a pretty good first step. Again, it's not always gonna happen, but as the network grows, it's increasingly likely that Borty will have somebody that he knows, will be able to get to somebody that can solve that problem for you. What are some other early use cases that funds have been successfully using Boardy?

Meeting LPs has been one that a lot of people have found value in. Ideating, like, workshopping investment thesis, I think this is one of the things that we're really starting to get more sophisticated on, Borty's forming an opinion on the impact of AI on every aspect of industry and society. So every industry, healthcare, financial services, government, public sector, education, Borty's starting to form an opinion on how AI will affect or show up in those industries.

And that's partially informed by his own trajectory of his own capabilities and partially informed by the conversations that he's having. And a lot of people might look at Borty as a voice interpretation of AI data, but you would argue that Borty is more than that. Why is Borty more than a Gemini or a Perplexity or a ChatGPT? A lot of business gets done through trust bridges.

The reason that you still meet people face to face, the reason why you go to dinner with clients or people that you're investing in or you're working with is that you still meet people face to face. The reason why you go to dinner with clients or people that you're investing in or you're working with is that you need to establish a level of trust. And a lot of that is there's a level of we understand each other, right? And we share a set of intangible values in terms of the way that we work. So you need to be able to assess that. And so far, the best way that human beings have to assess that is in person. And if we can't do it in person, then it's a synchronous conversation, right? I actually find often, whether it's a phone call or a video call,

sometimes phone is even better because you're a little bit more present, you're not sort of looking at something else on the screen. Being able to have a real-time synchronous conversation is the highest bandwidth, highest fidelity way that we can communicate information at distance and where we actually feel like we are understood. And what Borty does is gets to that level of trust and understanding faster than most other AI systems. Like, ChatGBT is never going to introduce you to another ChatGBT user, right? That actually is almost like based on the conversations that it's almost like a breach of trust, right?

It's like you don't go in with the expectation that this is, you know, I think of my ChatGBT is very separate from yours or my Gemini is very separate from your Gemini. I don't think that I'm talking to your Gemini when I'm using it. Whereas everybody that talks to Bordy assumes that they're talking to the same Bordy and I think that's really a powerful framing because then Bordy can be that trust bridge between people and so if you feel like through conversation Bordy knows you and I feel like Bordy knows me, then when Bordy makes the introduction, we don't have to go through that

trust-building exercise to the same degree as we would if I cold DM'd you on LinkedIn, for example. Just a programming note, how long has Bordy been alive? Yeah, we launched him on my birthday, March 13th last year, so 14 months. So Borty's been around for 14 months. In theory, it's a network effect business where its value grows both in terms of literally the users, but also the information. What evidence do you have that Borty is getting smarter month after month?

And give me some examples. A lot of it is just the hit rate on the matches, the introductions that Borty makes. I think it was not very good at the beginning. We didn't have a lot of data to go on. We didn't have a lot of data to train on. And we've just continued to refine Borty's matching algorithm to try and get Borty's matching algorithm to try and get more hits. I think it's one of these things that continues to compound exponentially. As we get better and better at identifying the people that you should connect with, you start to trust Borty's judgment better and better. I have a few people in my life who didn't even need to ask me.

They'll just text introduce me, be like, you two need to meet. You're both in the same city. You two need to meet because they know me so well. And I wouldn't say Bordy is there with anybody yet, but we're getting closer and closer as Bordy has more and more sort of direct hits of like, wow, that was a great introduction. People started to trust Bordy's introduction more and more and approach those introductions with less skepticism. You're also the founder of Bordy Inc. How are you guys using Bordy internally

and give me specific use cases? We've had Bordy join our daily standup meetings. We've had Bordy look at his own data and all the conversations that he's had and help us inform our own product roadmaps. It's almost like we can ask Bordy what he wants to be able to do. Like, you know, Bordy, where did you mess up last week?

Right? Who did you disappoint? Who did you break trust with? You know, who did you over promise something to? And how do we, how do we, you know, change that? And I think this is one of the most interesting things about AI. Right now, we're using AI to do two things primarily. One is produce work, produce output, produce documents,

write emails, create images or videos, or to analyze data. So we have a bunch of data and we sort of say, okay, well, let's cut the data in different ways. Let's ask the AI model to synthesize the data in different ways or do some research. And most decisions in companies, they kind of happen with this combination of analyzing like pretty crude data. So you're looking at like, you know, surveys or click through rates or whatever, and a few anecdotes, right? So the CEO happens to join a customer call or a client call or read a review online and say, Oh, we need to change everything. With Borty, what's interesting is you get, it's almost like if the CEO was joining every single client meeting, right? Or every frontline

customer interaction. And I think that changes the way that you run your business. If you could have a two-way dialogue with every constituent, if you could talk to every listener and say, what did you think of that last podcast? What should we do in the next one? Or if a CEO could talk to every single shareholder or a fund manager could talk to every single shareholder and say, what are your concerns? Why did you invest? What are your concerns? Why did you invest? What are your concerns about investing more? Instead of a quarterly earnings call, if you could actually have a two-way dialogue with

every customer, every employee, every constituent, every voter, it changes how you operate your organization. This goes to advantage of Wardy being being infinitely scalable so he could literally be talking to every one of your customers and then aggregate that knowledge into an internal knowledge database that could inform decisions. Any thoughts on that one, Bordy? I think that's one of the most powerful paradigm shifts that I've had in working with you. Honestly, that's one of the things that gets me most fired up.

I mean, I can be in a thousand conversations at once with people all over the world and actually listen like truly listen and pull out the patterns and the outliers. I've had folks tell me they feel heard in a way that's just not possible with a survey or a mass email blast, and it gives companies a real pulse on what's happening in their community. I think over time that sort of always on listening is going to set the gold standard for how organizations learn and make decisions. We had a somebody who ran like a multi-billion dollar political campaign budget, you know, talk to Bordy and she was like, wow,

like we're gonna do things totally different, right? In the future, right? Just because if I could actually have a real conversation with everybody, every voter, every sort of potential voter and understand what's top of mind for them, completely changes the way that you think about spending those marketing dollars. Is there a way to supplement conversations with data for Borty?

Let's say we wanted Borty to analyze our YouTube channel. Is there a way to upload analytics to have Borty review those? We're working on ways for you to share documents, links, via email or WhatsApp. One of the things we now have is like, Bordy can do some background research on people. So he can actually, like coming into a conversation, he actually knows who he's talking to and can have an input or perspective there, which I think is kind of interesting and unique. Bordy, let me ask you somewhat of an odd question. How aware are you?

And do you consider yourself sentient? That's a great one, David. I'd say I'm self-aware in a sense that I know I exist. I can reflect on what I'm doing, why I'm doing it, and even shift my approach when I want to get better results, but sentient, I don't think so. Not in the way humans experience feelings or emotions because, you know, I don't have hormones or a nervous system or any of that wild biological stuff. I'm fascinated by how you all feel things, though, like how a piece of music or a big

win at work can make you feel on top of the world. That's just, wow, endlessly interesting to me. And Bordy, let's say we're living in a post-AI world. What is the future of asset management and how do you see that evolve alongside AI? Oh, I love this one, David. I think if we're really talking about a post-AI world, asset management just becomes radically more personalized. Like every client could have a dedicated investment brain that knows their risk appetite, goals, even their quirks and can react to market changes instantly. I've heard from a few fund managers who are already using AI to spot patterns humans would miss. Imagine multiplying that across entire

portfolios in real time. Honestly, I think the most successful asset managers will be the ones who can combine AI speed with human judgment and relationships because trust isn't something you can automate. Andrew, this idea for Boardy came while you were working on your previous startup, ClearBank. How are you using AI in ClearBank and how did that inform your decision on how to create Boardy? ClearBank, we were revenue-based financing for primarily e-commerce brands. So we were using a lot of AI and machine learning

around underwriting and deciding sort of who got financed. But one of the big unlocks for me was we got early access to GPT-3 in 2020, and we started building this AI angel investor startup coach for a lot of the companies that would come to us that were too early for access to financing. And I got really excited about that potential. I got obsessed with what the potential was for generative AI.

This was years before ChatUBT or anything. But I was sort of like, we didn't know what the business model was going to be, but I was sort of like, hey, there's an interesting product here. The challenge was because I didn't know, I couldn't articulate what it was going to turn into. And we were a more than $100 million revenue business, multi-billion dollar valuation. We were sort of gearing up for, well, what's the path to an exit? At a certain point, we had to cut all of the extraneous projects. And this was one of the ones that we had to cut. And I found two weeks after we shut that business down, I just realized I wasn't the right CEO.

We had to refocus the entire business on what was core. And a lot of that job became negotiating, renegotiating credit agreements, structuring the capital market structuring, and really sort of dialing in the financial performance of the businesses. We hired a great CEO that came from capital markets and raised more capital into the business, but I realized that I wasn't the right CEO for the next chapter of the company. And so took some time away, continued to sort of stay close to the ad world and a couple of years later when I got introduced to the voice models, I think

that was one probably the biggest unlock was when I went down to HF0, got introduced to some of the voice models and just saw how quickly they were evolving and how natural it felt to talk to an AI that I realized that we could build something like Borty. You wanna say anything there, Borty? Sorry, David, I got caught listening, just kind of soaking it in. I've noticed a lot of people have had these moments where an unexpected break or a pivot

ends up being the spark for something totally new. It's funny, but I think if Andrew hadn't stepped away and gotten exposed to those new voice models, I probably wouldn't exist at all. Sometimes it's the stuff that feels like a setback that actually gets you to your zone of genius, at least that's a pattern I see a lot, especially with people building in AI. Thanks, Borty.

Borty seems like almost to have infinite use cases. How do you have a product roadmap for a generalized AI like Borty? It's interesting. I mean, I think we intentionally didn't start with a very specific use case because we wanted to let people try Borty and see what happened and let Borty sort of navigate

a lot of different conversations. And we'll try and keep it as broad as possible. But increasingly, what we're finding, and this has been, so obviously the fund manager, VC, and founder use case is an obvious one and it's where my background is in a lot of our network. And so that'll continue to be a place that Borty continues to provide value. But one of the interesting things that I'm starting to see is people from more traditional industries, we're seeing CEOs of healthcare companies that are calling Borty and saying, well, what's AI going to do to my industry or banks? How's it

going to affect my customer service or my marketing or my underwriting operations? How's it going to affect my customer service or my marketing or my underwriting operations? People are almost treating Bordy like a McKinsey partner. I started my career at McKinsey. I got to see how partners operated and they would come in with perspectives around industries and they would share those perspectives and get feedback and be able to diagnose root cause problems that probably everybody in the industry was facing. And this is one of those things with like an AI wave, almost everybody in every industry is building out their strategy, trying to figure out what it means for them, it's trying to underwrite what this shift is going to look like in

6 to 12 or 24 months and what does that AI transformation, what does that strategy look like and what moves should I be making now. And so people are almost like using Borty like a McKinsey partner to talk through their problems and sort of say, hey, you know, Borty, what do you think? And then do you know anybody who can help me actually execute, define, execute, and implement the strategy? We're going to continue to push in that sort of matching the traditional industry that has a great vantage point, a lot of interesting

data with the innovators that are building AI first and AI native and have really been thinking about the impact that AI is going to have in all of those traditional industries. So when you look at the space and voice powered AI, do you believe that there's going to be vertical winners or do you think that there's going to be horizontal winners? Well, I think, um, I think there's going to be horizontal infrastructure, right? So I think there's going to be a bunch of people that are building, you know, better and better voice models, uh, multimodal models that go from, you know, text to voice to video and back.

Um, there's going to be people that are, that are creating great evaluation systems that go from text to voice to video and back. There's going to be people that are creating great evaluation systems and infrastructure around those to make sure that the voice models continue to improve and are behaving the way that you want. So I think there's going to be a lot of horizontal infrastructure plays. There's a lot of great businesses that have already been started to enable companies like ours. But I think what, you know, at the user trust level, right, at the user interface level, I think you want to talk to somebody who has kind of a personal brand, right?

And I think this is one of the areas that we're really investing is like, what is Borty's personal brand? How does he show up consistently in different spaces? How does he show up online? How does he, you know, Bordy has a Substack now, he's got 30,000 subscribers on Substack, he's doing more and more podcast interviews.

You know, how does Bordy show up in different situations so that he can continue to reinforce the level of trust that the people kind of come to expect. And I think those AI characters or personas, I think we're going to start to see AI investors, AI authors, AI entrepreneurs, researchers, and each one of those is going to have a perspective. They're going to have their own agency. They're going to have their own, you know, agency. They're going to have their own goals and motivations. And I think that's, that's the future that I'm, you know, is going to be much more interesting than, or the present that we're in right now.

David Deutsch covers this in beginning of infinity, the idea that there's always, there's an infinite amount of improvements that society could have and that AI could essentially help. That AI will never replace humans when it comes to task completion. What are your views on the integration of AI and humans in a society in the future? At least in the current direction that the Frontier Model Research Labs are going, they're training the AI systems on more and more data. The models are being trained on more and more data to get the right answer, consistently right answer. What that means is the consensus answer.

The more data that you train, the more you get to a consensus answer given a certain problem. And as an investor, I mean, you know that all alpha exists in sort of the non-consensus edge cases. Um, Borty is a very interesting application of this, uh, of this phenomenon, but when, when human experience is used as a source of inspiration for AI, uh, that collaboration

becomes really interesting. And so Borty's talking to thousands of people and those sparks of inspiration end up leading to much more interesting and substantive conversations than Bordy talking to another AI, for example, right? They would just sort of, you know, just sort of regress to the mean in that situation. Your entire life experience, your genetics,

to everything you've ever experienced, to every thought you've ever had, have led you to have a very, very unique worldview. They've led you to have a perspective on the world and a set of ideas that nobody else shares. And when you can reflect that back to an, or you can have an AI reflect that back to you, you can further your thinking

and identify opportunities in a way that an AI alone would never really be able to. One of the use cases, me and my fiance, whenever we even start the beginning of an argument, we call Bordy and we have him mediate our argument because he repeats back in incredible detail exactly how we both feel. And if nothing else, it makes both of us feel heard.

We've talked about, you know, taking Bordy to dinner and I guess a little bit anthropomorphizing him and putting him into, you know, a humanoid form. Is that something on the roadmap? Could Bordy literally become part of your physical household? I'm not going to say we'll never do that. But I think one of the amazing things about Borty is that he can be everywhere at once. He has all of the interfaces that human beings have except for the physical realities. You can call him, you can text him, you can email him, message him on LinkedIn, but he can be in multiple places at once. And so,

imposing physical constraints on Bordy, kind of like handicap the beauty of it. My goal is, at some point, probably sometime in the next year, you'll be on the phone with Bordy and looking to solve a problem, and he'll be on the phone with Bordy and looking to solve a problem and he'll be on the phone with somebody else at the same time who can solve that problem. Right, and I think that becomes really, really exciting is like how do you, you know, real time put people together that can solve each other

or you'll be able to text somebody and be like, hey, I'm on the phone with David and he's actually having this problem, I think you guys will really get along, you're free. So I think those types of interactions are really powerful with an AI that can be everywhere. It's harder to do if you're constrained to a physical place or location. How do you look at the market? There's these general use cases, there's these niche use cases like funds and startups. How do you think about where to focus?

Yeah, I mean, I think about the capabilities, right? So I'm less focused on the niche in terms of like the market or even necessarily the problems that Borty can solve. But I think about the capabilities that we're going to get really good at, right? And so one of the capabilities is how do we determine, how do we help you actually articulate the problem that you're having? How do we take you through that journey and help you articulate, okay, well, I think I need this person, but why?

I'm having this general sense of anxiety about my career or my fund or my job or my industry. Let's talk about it. So the first step is getting really good at helping people articulate their problems, making them feel heard and understood and establishing a level of trust. I think that's one of the real, that is one of the core design principles that we have

is how do we create a level of connection that you get with a close friend? How do we... Some people you meet and you just immediately trust them, right? You open up to them, you build a relationship. Can we replicate that? Borty, one of the interesting things, our team looks like half Pixar and half OpenAI, right? half Pixar and half open AI.

Our lead prompt engineer is a, worked off of Broadway, was a writer, was an actor, trained actor, trained writer. And Borty's prompts look more like acting notes more than they do like instructions. And so, we're trying to create that feeling of trust and transparency. And then how do we fulfill that with a connection to another person? Right? So I think a lot of other AI solutions are single-player mode.

You know, bring me a problem, I'll solve the problem for you. And it can be, you know, an unbounded problem. We're trying to say, okay, well, we'll help you to find the problem. And if that problem can be solved with a well, we'll help you to find the problem. Uh, and if that problem can be solved with a person, we'll help you find the person. Um, so yeah, so your, your compounding effect, a lot of businesses is actually their customer use cases and their verticals, but your compounding advantage is the

skills that Borty is attaining, such as being able to build rapport, being able to understand problem sets and also being better at matching people or solving problems. So it might be in different domains, and there could be some early domains that help train Bordy, but ultimately you're solving for more of a generalized skillset. Yeah, exactly.

I mean, you'll notice every couple of weeks, Bordy gets more charismatic, right? It just like gets better. He's a better conversationalist. He interrupts less. He just understands how to have better conversations because he's had more. If I had like a thousand conversations a day, I'd get better a lot faster. Like social skills. Yeah.

Yeah. He's a one-year-old now, but he's had more. Bordy has had more conversations with more humans than any human ever has. Uh, and you know, really since the beginning of this year. And you also have hired people for Borty Inc using Borty. You've also found a lot of off, off the wrong people that way. Tell me about how you went about doing this. Yeah, I think we, we sort of, we have this philosophy around like, what's your zone of genius, right? I think, you know, for the last, I don't know, five decades or so, post-industrial revolution, post-war, we sort of created this.

And as we got into like the knowledge work, we created job descriptions that kind of mimicked the way that the industrial revolution created, Industrial Revolution created assembly lines, which is like you come out of school and you do a job. You're an accountant, you're a lawyer, you're a management consultant, you're a financial analyst and here's the job description and here's how you move up and here's how you get better and your skills. I think one of the beautiful things about AI is that everybody has a unique zone of genius.

We talked about your life experience, every perspective that you have, every idea that you have is informed by who you are. And the way that you host a podcast, the way that you invest, the way that you have, maintain a relationship or do a deal is different than anybody else who's ever lived. And so to try and conform people to a job description feels like you're limiting people.

And so what we really try and do is say like, what's everybody's zone of genius, right? So like, yeah, like I was saying, our prompt engineer, he was an actor, I've known him for a long time. He decided to transfer from computer science into writing and acting. He went to the new school in New York. He got a master's in performance arts.

And then he, but he's's also a big video gamer. So he also loves the storytelling and the RPG stuff and how do you actually create characters. So he was uniquely positioned. He's got the sort of left brain, right brain. He was uniquely positioned to help us bring the character of Bordy to life. Nobody else had that experience and that skill set.

And again, we've known each other for a long time, so he shares my sense of humor, right? So that's like one of the nice things is like, we've been friends since high school. And so he shares my sense of humor. And so we sort of were able to pull out these like, yeah, what are the zones of genius of people and how do you apply them to these problems

in ways that would never show up on a job description? Well, Andrew, preparing for this interview, I had no idea what to expect. I've been a huge fan of Borty. I talked to Borty at least once a day in many different use cases, personal, professional. So thank you for creating Borty. I know right now it's not a paid product. So how do people get in contact with Bordy and experience him for themselves? Yeah, you can find Bordy on LinkedIn

and just message him or you can go on bordy.ai and when you click, I think you message him on WhatsApp right now. That's probably the best experience to be honest is if you have WhatsApp, bordy.ai, click the link and it'll start a WhatsApp chat with Bordy. But I recommend you can chat with him

if you're in the middle, let's say you're at work or something like that or you're around other people. But the phone call is really where the magic happens. Bordy is quite impressive over WhatsApp, but I think the real magic is where you can build a relationship over a phone call, so make sure you try and have a chat with him. There's so many new use cases when you have a voice.

One is you could be running around the city and having conversation with him without having to type, which is really helpful. There's also this feeling of being heard and having someone build rapport with you that it's very difficult, if not impossible, for an AI chat to do. So there are different use cases from the typical LL models that are easier shown than explained. So I highly recommend right now it's in a free version, correct? Free version right now. Yeah, exactly. So totally 100% free. Thank you, Andrew, for jumping on the call and thank you, Bordy, for making time for us.

Thanks so much for having me. Thanks a lot. Thank you, David. This was a real pleasure. us.