Most people think networking is a volume game. More emails. More intros. More “quick chats.” Like pouring water faster into a leaky bucket and calling it strategy.
Andrew D’Souza thinks that’s backwards.
After building one of the most consequential fintech companies of the last decade, Clearco, Andrew is now working on something stranger, riskier, and arguably more important, an AI that doesn’t optimize for growth, but for judgment.
This post is for people who won’t listen to the full episode but still want the core ideas, without the fluff, without the hype.
The real problem isn’t access, it’s trust
Imagine two worlds.
In the first, you can get introduced to anyone. Investors, customers, hires, partners. Infinite intros. Zero friction.
In the second, you get fewer introductions, but almost all of them are worth your time.
Most of today’s software bets on world one. Andrew is building for world two.
Early in his career, he learned a hard lesson while building a referral-based recruiting marketplace. The idea was simple, pay people to introduce candidates. The result was subtle and destructive. The moment money entered the equation, trust evaporated. People stopped assuming goodwill and started questioning motives.
That insight stuck.
Trust, once diluted, is almost impossible to re-concentrate.
Clearco proved scale, not defensibility
Clearco helped create a new category, programmatic, non-dilutive capital for founders. It scaled fast, globally. But Andrew noticed something uncomfortable.
Capital isn’t a network effect.
Once the category was obvious, competition rushed in. Pricing compressed. Differentiation narrowed. The business became operationally excellent, but strategically fragile.
That’s when Andrew did something most founders avoid. He stepped aside as CEO.
Not because the company was failing, but because the job had changed. It needed a different kind of operator. He stayed close, but mentally moved on.
The next thing he wanted to build had to be harder to copy.
Enter Boardy, an AI that behaves like a person, not a product
Boardy isn’t an app. There’s no dashboard. No onboarding flow. No features page.
You talk to it. On the phone. By text. Like a human.
That design choice isn’t cosmetic, it’s philosophical.
Humans don’t trust interfaces. They trust relationships.
Boardy’s job is simple to describe and extremely hard to execute, make high-quality professional introductions at scale. Not maximum intros. Good intros.
And here’s the twist. Boardy can say no.
If it believes an introduction will waste time, damage goodwill, or hurt the network, it refuses. That alone makes it feel radically different from software trained to comply.
Think less “assistant” and more “board member.”
Why saying no is the feature
Most AI tools optimize for helpfulness in the narrow sense. Do what the user asks. Faster. Cheaper. More reliably.
Boardy optimizes for a different objective, long-term trust across the network.
Every introduction spends goodwill. Every bad one compounds distrust. Boardy tracks this implicitly, who responds, who ignores, who follows through. Over time, it learns judgment, not just matching.
The worst outcome is wasted time.
The best outcome is life-changing.
That asymmetry matters.
Voice is not a gimmick, it’s biology
Andrew is blunt about this. Humans evolved for sound, not screens.
Before language, there was tone. Before text, there was voice. Podcasts work not because they’re efficient, but because they’re human.
Boardy uses voice because voice carries intent, hesitation, confidence, and context that text strips away. It’s higher bandwidth, emotionally and informationally.
Typing is new. Talking is ancient.
AI that ignores that is fighting human nature.
The contrarian AI take, imagination is the bottleneck
Here’s where Andrew gets provocative.
He doesn’t think we’re limited by models, compute, or talent. He thinks we’re limited by imagination.
Most AI startups are doing incremental work. Faster versions of old workflows. Slightly cheaper humans. Polished wrappers.
Useful, sure. Transformative, rarely.
The real opportunity is asymmetric problems, where the downside is small and the upside is enormous. Introductions are one. Creative partnerships are another. Founder-investor matches. Network orchestration.
AI is finally good enough to explore possibility space, not just optimize known paths.
But only if we let it.
A glimpse of the moonshot
Andrew’s long-term vision for Boardy is audacious.
Imagine an AI that understands your strengths better than you do. That helps articulate what you should build, who you should meet, and when. Not once, but continuously, as your context changes.
An AI co-founder with taste.
An AI connector with restraint.
An AI that compounds trust instead of burning it.
He compares it to an AI version of Richard Branson, spotting people’s zone of genius and connecting dots across industries, without needing to run everything itself.
Whether Boardy becomes that or not is an open question.
But the underlying idea feels durable.
In a world drowning in noise, the scarce resource isn’t information or access. It’s judgment.
And that might be the most interesting thing AI can learn next.
👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL
🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast
Chapters:
00:01 – Breaking Things, Founder Embarrassment, and Why It Matters
03:30 – Studying the Brain, Behavioral Economics, and Human Decision-Making
06:40 – McKinsey, Facebook, and Getting the Startup Bug
10:10 – Early Lessons on Incentives and Why Referral Marketplaces Fail
14:45 – Becoming a Connector Between Silicon Valley and Canada
17:30 – The Origin Story of Clearco and Access to Capital
22:10 – Financing the DTC Boom and Creating a New Category
27:40 – When Scale Outgrows the Founder, Stepping Aside as CEO
31:50 – Discovering GPT-3 and the Clear Angel Experiment
36:10 – The Problem with Personal CRMs and Human Memory Limits
39:40 – Network Effects, Voice AI, and the Birth of Boardy
44:30 – Why Voice Is the Most Human AI Interface
48:20 – Judgment, Trust, and Saying No in Network Design
52:51 – The Future of AI, Imagination, and Human Connection
Transcript
Brian Bell (00:00.951)
Hey, everyone. Welcome back to the Ignite podcast today. We’re thrilled to have Andrew D’Souza on the mic. He’s the founder and CEO of Bordie.ai and the serial entrepreneur behind Clearco, formerly ClearBank. He’s built pivoted broken and reimagined companies at the intersection of finance, AI and networks. He’s now constructing an AI super connector that speaks to you, introduces you and helps you build leverage in the real world. Let’s dive in. Thanks for coming on, Andrew.
Andrew D’Souza (00:24.066)
Thanks for having me. This is fun. This is fun. definitely broken the parts of things in my career.
Brian Bell (00:31.093)
Yeah, you haven’t really shipped product or been a founder if you haven’t broken things and been really embarrassed by launches. But I’d love to dive into your origin story. What’s your background?
Andrew D’Souza (00:41.622)
Yeah, so I mean, I studied engineering. I actually did a deep focus on a specialization called simulating neurobiological systems. And so sort of this idea of like, how does the brain work and can we model it in computers?
At that time, we were simulating neurons in MATLAB. But it was fascinating because also there was a cadaver lab. took gross anatomy, neuroanatomy. We were looking at brains, different brain components, physiological psychology. So how do our hormones and neurotransmitters affect the way that we process information, the way we feel, the way we make decisions, consumer behavior, how do marketing messages affect the way we process information, make decisions. And then
Brian Bell (01:00.172)
Wow.
Andrew D’Souza (01:24.204)
a couple of macroeconomic courses around behavioral economics, behavioral psychology. So sort of thinking about how does this create markets, how does this create networks? And so it’s always been an area that’s been fascinating for me, was just sort of like, how do we work as people? And how do we interact with computers, technology, systems, software? It’s been a lifelong question for me.
Brian Bell (01:53.111)
That’s amazing. we birds have a feather a little bit except for the cadaver part. You know, I never, never studied brains, but I really got into cognitive neuro psychology and I almost went back to grad school. I was reading like books like nudge and all the behavioral econ books back in the two thousands, 2010s and just really fascinated. I think it’s what kind of probably drove me into building product is just understanding how people interact with technology and how they think. And what was that trigger for you though, where you’re like, okay, I want to build something that just, you know,
Andrew D’Souza (01:57.471)
you
Andrew D’Souza (02:08.684)
Yeah.
Andrew D’Souza (02:14.307)
Yep.
Brian Bell (02:23.205)
do research or consult or
Andrew D’Souza (02:25.28)
Yeah, I mean, started my first job was at McKinsey. And so was a lot of research and consulting. And I think very quickly, I realized I’m not cut out for that. I wasn’t a very good employee. I never really listened to what they needed me to do. was like, well, what about this? What about this? What about this? And so eventually, yeah, yeah. So I ended up.
Brian Bell (02:41.259)
Classic founder mentality, yeah.
Andrew D’Souza (02:46.772)
Somehow my resume ended up on Chamath’s desk when he was running growth at Facebook. So he was probably 10 years ahead of me in school at Waterloo. And so we had, you know, through a couple of people that like, I’ve got a friend who, you know, is senior at Facebook. Do you want to go talk to him? And he was like, Hey, look, you could join my team at Facebook or I’m, you know, leaving to start a fund. I’m incubating a company here. It was a company called Top Prospect and it was a referral based recruiting marketplace. So
Brian Bell (03:14.487)
I remember it. I remember it. Yeah.
Andrew D’Souza (03:15.858)
The whole concept was like everybody, this is in 2010. like now venture was flowing back into the ecosystem. A bunch of people were paying big bounties for, if anyone referred them an engineer and we said, okay, well, could we consolidate these? Could we put them all in one platform and use some machine learning to sort of connect people? I think one big lesson for me from that experience was people...
When you put a financial incentive in front of something that people do just like out of goodwill, you really change the behavior. So like if I introduce you to somebody and I say, you should talk to this person. Maybe you want to work with them. You’re like, great. Andrew knows me. Andrew probably knows that person. He’s probably thinking about what’s best for us. If you know that I’m going to get paid, if you work together, you both go into that meeting like a little bit more apprehensive. You’re just like not as open because you’re like, well, what’s really the point here?
And so for that and a number of reasons, we actually did a whole study with Dan Ariely on this, on behavioral economics around incentives for recruiting. But there were a few reasons, but that was I think one of the main reasons that that company didn’t work out. And so it was sort of an important early lesson for me in my career. But that was sort of what gave me the startup bug.
I ended up getting to know lot of, Andreessen Horowitz was an investor. They’d just started their fund at the time. So, know, met Mark and Ben and Michael Ovitz and those folks when they were, when they were starting out and a bunch of great, great investors in the Valley were involved in the company. And so I got to know a lot of investors, know a lot of founders. And I kind of became like a, you know, early conduit between Silicon Valley venture and the Canadian sort of Toronto tech ecosystem. And so was in every time.
Toronto founder was going up to Rays, they would ask me to introduce them to people and every time a VC was going to Toronto for a board meeting or something, they’d ask me if I knew anyone. So I ended up playing a bit of that connective tissue for a little while.
Brian Bell (05:15.883)
That’s amazing. So what happened next to your, your at top prospect and somehow clear co came along. What was that?
Andrew D’Souza (05:21.503)
Yeah, so I joined a couple of Canadian startups over the next few years, helped them expand into the US, helped them raise a couple rounds of funding. then I just saw, I started doing some angel investing. I invested in a company called Wealthsimple, which is doing really well in sort of the robo-advisor space. And sort of got really interested in financial services. I’d always been interested in how do you help founders? How do you help people access capital?
Access opportunity that they don’t you wouldn’t always have access to And sort of saw this new economy emerging with originally was uber drivers that were like, you know driving and weren’t getting paid every day and we were advancing in their funds and then we sort of Scaled up from there. We worked with vacation rental Airbnb hosts who were who sort of You know had bookings out in the future and wanted to advance the funds against those to expand or invest in their properties
And then where it really took off was e-commerce entrepreneurs. So people that were, you know, had figured out a product, were scaling on Instagram ads and TikTok at one point, buying inventory. you know, I think like if you saw an Instagram ad between, you know, for product between like 2017 and 2021, there was like a reasonable chance we financed that ad. So basically we were financing the people to grow their business. And then we were taking a share of the revenue until we got it.
Brian Bell (06:47.995)
That’s interesting underwriting, right? Because you’re basically saying, OK, what’s your CAC and LTV? What’s your payback? And I’ll underwrite you for some, you know, some five figure amount per month to run these ads, maybe even six figure amount to try to grow your business. What an interesting model.
Andrew D’Souza (06:52.183)
Yeah. Yeah.
Andrew D’Souza (07:04.407)
Yeah, was just as sort of direct to consumer fell out of favor in the venture world. think people realized it was hard to build a venture scale business and a venture type return. I think in a Warby Parker, Bonobo, State All sort of like gotten venture scale. Some of them had exited, but there was a whole wave of people that invested and then not as many of those companies actually sort of broke out.
Brian Bell (07:28.052)
Kind of the Uber for X problem. Hey, Warby Parker did it for glasses. Now let’s do it for men’s shaving cream and all kinds of random stuff. Yeah.
Andrew D’Souza (07:32.46)
Yeah. Yeah, exactly. Yeah, Dollar Shave Club had a big exit. So there were couple of big early exits in the space and I think everyone was like, no, this is great. There’s a gold rush. And it’s just a hard business to build and it wasn’t.
It wasn’t clear that every one of those businesses needed to have a venture capital or even equity sort of investment. And so we created this instrument that allowed us to finance these businesses at scale and do it programmatically. I think for a lot of people, was the first time they’d ever seen anything like this. And so we sort of created an industry pretty quickly around it.
Brian Bell (08:07.519)
Yeah, almost like revenue based financing for D2C companies. That’s really interesting. And then what happened next?
Andrew D’Souza (08:14.891)
Yeah.
Well, it was interesting. we had a lot of people that applied for financing from us. I think we looked at our data at one point. We had financed about 5,000 businesses. We had 50,000 that had applied. And around the same time, we got early access to GBD3. This is 2020, so a couple of years before Chatchi BT, we got API access to GBD3.
And we started playing around with it and said, well, what can we do to keep these people engaged? Maybe they will qualify in the future. Maybe we can help them. And we built this product called Clear Angel, which was sort of like an AI angel investor, startup coach. And I got sort of obsessed with that. was like, hey, this is the future. I want to build in the space. And I spent a lot of time and energy trying to build this product.
You know, but the core business was ripping at the time, right? We had sort of like, we had found a seam. We were financing, I think we were making like hundred million in revenue. We were financing billions of dollars. had 600 employees in 11 countries. And we kind of came to a crossroads of like, hey, you know, Andrew, you’re spending 80 % of your attention on this side project, but we have a business to run. Like we should really focus. I couldn’t.
I had a feeling that this was going to be a thing, I wasn’t. Again, this was before ChatGBT, before the real AI. People have been talking about AI for decades, and it wasn’t clear this is where things are going. And it wasn’t clear how we were going to make money off of it. So we were spending a lot of money and lot of attention on this Clear Angel AI project. But I couldn’t convince my team or my board that it was worth continuing to spend. So we shut the project down. And two weeks later, I was like...
Andrew D’Souza (10:02.861)
I’m not the right CEO to run the company. Now it was a pure play financial services company. was a scale financial services company in a rising interest rate environment. You needed somebody who knew what they were doing. You needed somebody who was like.
Brian Bell (10:12.693)
Yeah, you did the financial operational rigor of somebody who’s run that kind of scale of a FinTech, right?
Andrew D’Souza (10:18.517)
Yeah, it was, okay, let’s renegotiate our credit agreements. Let’s structure our capital markets, right? Let’s like think about sort of our credit underwriting. you know, it was like, it was not just like, let’s grow as fast as possible. It’s like, okay, let’s build this into a, you know, a sustainable, you know, scaled financial services business. And so we went out and hired a great CEO. We still run the company today and is doing a great job and put it on a, on a really strong trajectory. But so I was, you know, active chairman for about a year during that time. And then, and then.
I’m still on the board and as helpful as it can be, but I’ve stayed close to the AI space since then.
Brian Bell (10:54.711)
That’s amazing. tell us about the germination of Borty. Obviously, you’ve been working on a similar project and it was just kind of festering there. us of kind of like when you said decided to make that leap into let’s go build this thing.
Andrew D’Souza (11:09.131)
Yeah, I think you probably do this. I’ve done this for a lot of my careers, like connect people, right? I have... Yeah, yeah. My challenge is I have a great imagination, but a terrible memory, right? So if I meet two people, I’m like, you two should talk about this. And I’ll come up with all kinds of interesting things that they should talk about. But I never remember all the people that I’ve met. And so it’s really difficult for me to do this at scale. At beginning of my career, didn’t know anybody, so I remembered all
Brian Bell (11:16.599)
That’s all I do as a beastie.
Same. Yeah.
Andrew D’Souza (11:39.373)
once I knew 100 people, I was like, oh, well, the guy. Yeah, there’s just no way to have that kind of recall. Some people are very good at it. Politicians, I think, are very good at this kind of thing. I just don’t have that in me. So I realized, one,
Brian Bell (11:41.579)
Yeah, or thousands. Yeah. Yeah.
Brian Bell (11:53.367)
Well, the worst thing is, is when you run into somebody who’s been on your podcast and you don’t remember that they’ve been on your podcast, that’s the worst. Like, dude, I’ve been on your podcast. You’re like, yeah, you have. I’ve reached the Dunbar number of my podcast already. like your episode, your guest number, like 230 or something like that. So it’s like, yeah, I’m already running into like limits of my brain. Yeah.
Andrew D’Souza (11:57.198)
Yeah, yeah. Yeah, yeah, yeah, for sure, for sure, it’s tough.
Right. Yeah, it’s tough. We’re only wired. Yeah, we’re not wired for the eras of social networks and emails and the velocity of meetings and Zoom meetings and all of that stuff. So it’s not surprising that we sort of hit those capacity limits. But AI doesn’t have those constraints. And interestingly, generative AI also has a great imagination. If you design it in the right way and prompt it in the right way and evaluate it right way, you can actually
get some of the benefits of the hallucinations. like, oh yeah, that is an interesting reason why I should meet this person. I never thought about that. that was sort of always something that I wanted to do. I’d always wanted to build something with a network effect. I think that was one of the things we realized at Clearco. Capital was a bit of a moat, but there was no network effect. And so once we created the category and made a bunch of noise, we started to have a lot of competition come in. And then it became a race to the bottom of who can offer the most aggressive pricing, who has the lowest cost of capital.
And so I think that that was a lesson for me was like network effects are very hard to start but when they work They’re kind of hard to stop And so I was like, okay. Well, there’s there’s an interesting concept here around network effects connecting people and then I saw what 11 labs is doing with the voice models and I got a chance to experiment with those and that was sort of the unlock for me I was like, well
This feels like a moment where now we can put these things together. Lots of people have tried to solve the personal CRM and marketplace and how do you connect people professionally and it all felt artificial. But this is the way that people, human beings are wired to communicate through sound. we...
Andrew D’Souza (13:48.75)
We talk, I this is why podcasts do so well, right? This is why we take phone calls. I would never introduce someone to someone that I hadn’t spoken with, right? Like even if it’s a five minute phone call, I’m like, hey, let’s get a sense of this person. It’s such a high bandwidth, high fidelity means of communication, synchronous audio. Like before we had language, we had sound, right? We would grunt and yell and communicate danger or whatever, food. Like that is like so wired in our biology.
And reading words and typing stuff out on a screen is so new that for the first time being able to have a synchronous real-time audio conversation that approximates a human-level conversation unlocks a whole new way of conveying and gathering information.
Brian Bell (14:38.731)
Yeah, and as a user, know a lot about it and probably a lot of people listening are VCs or founders. So they know of Borty, but walk us through how it works. What’s the user experience? How does someone onboard and meet Borty? What is Borty and what does it do? Yeah.
Andrew D’Souza (14:52.309)
Yeah. Yeah. So Borty is an AI super connector. I don’t know if you can see him over my shoulder here, but he’s a, he’s an AI super connector. You can talk to him on the phone. So it’s, it’s a very simple onboarding, it, breaks a lot of patterns that I think, you know, most of us are used to when it comes to interacting with software. There’s no app, there’s no download, there’s no login, there’s no dashboard. The, the interfaces I have with Borty are the same interfaces I have with everybody else in my life. He has a phone number. I text him, I’m on WhatsApp with him. He’s got an email address. I’ll email and true him.
people. He’s got a LinkedIn profile and a Twitter profile. But other than that, if you and I can’t have that interface, then Bordy won’t have it. so that was a challenging design decision in the early days. But the way it works is you sign up, you talk to Bordy, he’ll give you a call. You chat with him about whatever is top of mind for you. A lot of founders raising capital that’s kind of a big focus for us right now is helping founders meet investors.
I think we’ve already talked to 6,000 founders in last couple of weeks who are actively raising and about 700 investors who are looking to deploy capital actively. we’ve got a good ecosystem.
Brian Bell (16:00.609)
starting to get the network effects kind of spinning, at least in this one niche, which is, you know, fundraising.
Andrew D’Souza (16:05.685)
Yeah, yeah, it seems to be one, you we sort of let Borty out in the wild and tried to see what was happening. People were finding customers, people were hiring. There were all kinds of interesting things happening. But the most repeatable was this sort of founder investor thing, because it happens. You know, there’s not a lot of other situations where you meet somebody and then like a huge sum of money changes hands within weeks of first introduction. One of those. Yeah, one of those one of those those weird situations where that happens. And we think of Borty as like
Brian Bell (16:26.677)
days in the valley.
Andrew D’Souza (16:35.797)
influencing an economy. so part of the goal is like, how does Bordy sort of create this economy, this network around him that he can start to sort of help push things in more efficient directions and bring investors in and sort of foreign direct investment into that economy.
Brian Bell (16:56.565)
Yeah, it’s pretty amazing. if anybody here has actually onboarded, it’s a distinctly Australian accent. Tell us about the design decision there.
Andrew D’Souza (17:06.497)
Yeah, I listened to like 200 accents and this is the one that felt most like Borty. You know, it’s interesting. I didn’t think about this at the outset, but now I’ve talked about this to a few people now. My very first, like my first boss at McKinsey was this guy named Steve Hasker. He was like a senior partner of McKinsey. This like Australian dude, very charming, very smart, very charming, right? Everyone really loved him.
All the clients loved him. He’s now the CEO of Thomson Reuters. He was the CEO of CAA for a while. He’s the CEO of Thomson Reuters. He’s had a great career since McKinsey, but he was my first like mentor there. And I think that stuck with me. was like, who is like a professional? Like who’s somebody that everybody like, you know, likes? And Steve, you know, when I heard Borty’s voice, Steve came to mind or maybe subconsciously. And that’s why I was like, that feels like it should be Borty. Yeah.
Brian Bell (17:52.373)
Yeah. That’s amazing. Yeah, it’s very disarming, right? like a smart, intelligent, friendly Australian that you’re talking to. Hi, I’m Bordy, you know, and you’re like, hey, Bordy, you know.
Andrew D’Souza (18:01.505)
Yeah. Exactly.
Brian Bell (18:07.031)
So what are some of the guardrails that you guys had to think about? Privacy, bias, data, consent? know, it’s, you know, a very complex, you human-like situation. Yeah.
Andrew D’Souza (18:18.785)
Yeah. Yeah. I mean, think a big part of it is just like, we focus on, I think a lot of systems will just watch you, right? So if you think about Facebook or Google or, you know, there’s, there’s a lot of products that just watch your behavior and observe and drive for instance around it. I think one of the big things for us is like, Borty basically just asks you questions and you volunteer information. You’re like, Hey, this is what it’s, you know, this is what’s important to me. And
The goal is you give Borty the information that he needs to be helpful. We named him Borty, by the way, as an AI board member. So the idea is a board member, somebody you can always call and talk to and can open doors for you. Has a network, has connections, is perfect board member, is non-judgmental, always has time for you, always answers the call. But yeah, part of it is he can actually open doors. He knows people. He’s got a network. And so on one hand, we...
We’re quite careful. Some people try and use Bordy like a therapist, relationship advice, dating. You see people try and use Bordy in a bunch of different ways. And it’s amusing, but at the end of the day, Bordy’s goal is to drive professional connections. So a big part of it is tell Bordy the information that you would tell a connector who you wanted to connect. So you have to tell him something so he has something to go off of. And he’ll try and pull that. It’s quite an interesting conversation, especially as a founder.
I think for a lot of people, find it helpful to talk through why they’re building what they’re building and why they’re uniquely positioned to win, right? Why they’re the only person on the planet who could build what they’re building in this moment in time. And Bordy sort of helps you articulate your story in a unique way in that way. And that’s what gives him the ammo to go out to investors and say, wow, this is why you should meet Brian, right? Or this is why you should meet so and so because, you know,
They are working on this. You mentioned, right? So he’s also had a 30 minute conversation with a bunch of investors. So it’s like, you mentioned that you were really interested in people who are solving these kinds of problems in these ways. Well, I just met this person who fits this bill for these reasons. Would you like to meet them? And so I think a lot of what we tried to do is force things into the professional realm. The information that you tell Borty is not going to be blasted out onto the internet.
Andrew D’Souza (20:45.429)
But it’s going to be used to connect you with the people that you want to meet. And so if there’s some insider trade secret that you don’t want an investor to know about, you probably shouldn’t tell Bordy about it. If there’s some big milestone, traction or revenue milestone, that you’re like, well, I wouldn’t want my competitors to know about it, but I wouldn’t mind my investors or prospective investors know about it, those are probably pretty good pieces of information.
Brian Bell (21:07.499)
Yeah. And there’s this, the story around how you guys dog fooded your own raise using Bordy. Maybe you could tell the story of that. I guess you guys raised 8 million with using Bordy as your collaborator.
Andrew D’Souza (21:15.881)
Yeah. Well, yeah, it was interesting. mean, we didn’t even really talk. It wasn’t even intentional. Borty was just out in the world talking to people. Simon and Hanel, guess initially Simon from Crandom had chatted with Borty and was like, wow, who’s behind this? Like, this is, you know, this is like a very interesting, unique experience. Like, who is behind building this and asked Borty to introduce me? And I don’t think Borty pushed back a little bit.
And then eventually they found me and were like, we talk? We weren’t really raising. So Crandom is like one of the top firms in And they’ve got an outstanding reputation. I had never heard of them. We built a global $2.5 billion business in Clearco, but they had never heard of me. had never heard of Clearco, had never heard of me.
Brian Bell (22:11.627)
That’s how big tech has gotten, right? Like, you know, where you can have two really big firms that nobody’s of, you know?
Andrew D’Souza (22:17.451)
Yeah. And so it was really kind of like in the most pure way, Borty brought us together because like, you know, they had talked to Borty and were like so impressed that they wanted to meet me. And so this was on a Friday, probably Friday evening for them in Europe. We jumped on a call. We spent a bunch of time on Saturday, their whole partnership talked to Borty over the weekend and kind of put him through his paces and we’re doing diligence calls for their portfolio and stuff. And on Monday, they were like, look, we want to be your partner. And so, you know, we hadn’t like, we didn’t, we were
kind of probably two months away from running a process. We’re like, okay, you know, so we’re like, okay, starting to put a rough deck together. didn’t have any metrics, didn’t have any, you know, it was just like, Hey, we have a product out there and like, we’re to start preparing. so, but, but I think part of it was just, they had such conviction after experiencing, you know, talking to Borty and they, I think they saw the potential of what this could be. that they, know, that, that meant that went a long way for.
Brian Bell (23:14.967)
Yeah. So you got preempted because of the quality of the product and sort of the almost like the right timing, right idea at the right time, right founder. Yeah. That’s kind of like, you know, I call those easy yeses in our industry where you kind of like, great product, right time, right founder. Cool. Go, you know.
Andrew D’Souza (23:23.244)
Yeah.
Andrew D’Souza (23:31.243)
Yeah, it is definitely one of these products and companies that you need some imagination, I think as an investor, right? think it’s not, like, I don’t even really know what Borty is gonna be. know, like, know Borty is gonna be important, but I don’t know what the end state is gonna look like. Like, I have some ideas, but in some ways, like, Borty is gonna determine a real good investment.
Brian Bell (23:51.809)
Yeah.
Brian Bell (23:56.299)
But it’s almost like it’s almost like a new interface on all the tooling is how I see it. Like, I don’t know how you guys are thinking about the roadmap for tooling, but I’m putting my product brain on and I’m thinking, hey, I’m chatting with it as I drive down the 80. Right. And I’m like, hey, I’m doing this and doing that. And I’m like, it’d be great if like you could go to my LinkedIn sales nav and just send that message. You know, or like or if you could read the inbound email and, you know, is there a deck there? if there’s not a deck, then
Andrew D’Souza (24:09.655)
Hmm.
Andrew D’Souza (24:17.909)
Yeah, yeah.
Brian Bell (24:25.943)
you know, ask the founder for a deck and then, oh yeah, then read the deck and then using my rubric for how I evaluate startups, give me a score, right? Of like how I would evaluate the startup based on what’s in the deck. Oh, it’s missing these three things I care about. Go back and ask the founder for those three things and then re-score it for me. So I think it’s like, it might become this like almost like super connector assistant. I don’t know what you think, but.
Andrew D’Souza (24:34.274)
Yeah.
Andrew D’Souza (24:44.492)
Yeah.
Andrew D’Souza (24:49.995)
Yeah, you know, a lot of people ask about the assistant space. think there’s a lot of AI assistants. I think it’s quite crowded and I think there’s going to be a lot of specialized AI assistants for people that will do that, like read my emails, know, like, you know, integrate with my CRM. I think of Borty, like the difference is like Borty is more principled than agent, right? He’s more his own, like he’s got his own opinions. He’ll push back. He’s in our Slack.
He’ll push back on things. He’ll be like, don’t think, you know, we were talking about like, do we need to manage token budgets? And he’s like, no, like, I mean, obviously he’s an AI. wants to, he wants to consume as many tokens as possible. But he has like opinions on things that are like quite interesting. So like, you might ask, so like even, you know, random, they’re like, can you introduce me to your founder? And he’s like, like, no, right? Like, we don’t, like, I don’t do that. I don’t, I don’t introduce people to specific people in my network.
you know, to give me the sense of who you’re looking for and I’ll try and do this. Cause if you would just do that, then everyone would be like, can you introduce me to Elon? Can you introduce me to Mark Andreessen? Can you introduce me to the most important person? And they get, they get spam. So he thinks about what’s good for the network, right? Every time I make an introduction between two people, I sort of, I assess how much goodwill I have between both people. And I assess how much goodwill I’m gambling by making this introduction, right? If it’s an obvious one.
It’s like, great, I’m gonna introduce these two people, they’re gonna work together, it’s gonna be great. If it’s a bit outside, I’m like, well, this could be a flyer, but I think they’re gonna hit it off. So I’m gonna say, this may seem weird, but Brian, I want you to meet this founder. If I do that too many times, it backfires, if that works out, you’re like, wow, this is amazing, I’m gonna trust every intro Andrew makes. If it’s a waste of your time, the next time you’re like, maybe I’m gonna second guess. Yeah, they lobbed the fit, right?
Brian Bell (26:35.039)
And we all have those people on our network, right? Where they’re sending us an intro and you’re like, I know this is like a low quality intro. You know, I just know, I just know it, you know,
Andrew D’Souza (26:41.601)
Yeah. So Borty is kind of assessing like how much goodwill does he have with everybody in the network? And is this going to increase the goodwill across the entire network? And so he doesn’t like this. The Borty that you talk to is the same Borty that I talked to is the same Borty that everybody else in the network talks to. And that’s quite different than your assistant being the same assistant as my assistant, right? Or your chat GBT being the same chat GBT as mine. And so it’s a very different architecture.
and a very different positioning where Bordy has some independence and you’re sort of kind of negotiating with Bordy about what your needs are and how they align with the needs of everybody else in the network.
Brian Bell (27:22.519)
That’s a really fascinating problem. How do you architect that into an AI system? And you’re also probably struggling with the memory and how are you building all that?
Andrew D’Souza (27:33.569)
Yeah, it’s a real challenge. I was really hoping that somebody would have solved memory by now. was like, I don’t want to build it myself because it’s a tough problem.
Brian Bell (27:40.79)
Yeah.
Yeah, I’m looking through the YC batch right now and there’s there’s always like out of 150 companies, there’s two or three trying to solve the LLM memory thing like every batch.
Andrew D’Souza (27:51.67)
Yeah, yeah, don’t think context windows are the right answer. I think a lot of it comes down to context engineering. So we have kind of simulated what humans do, is like, it’s intentional recall. So it’s different than rag, where it’s just like, you’re just grabbing whatever. You have no real control over a semantic search and which chunks you’re grabbing.
But we can pull in at each conversational turn, pull in some relevant information, update a system message, and dynamically have Borty talk things through. yeah, the memory part is a challenge. You have your one thread with Borty, and everyone has their own independent threads with Borty. But every so often, he has to figure out, what do I need to know about this other person?
that I’m trying to introduce you to. So how do I grab the right amount of information, filter it appropriately for the right context so that it’s only the stuff that that person would want shared with you in the way that you want it shared. And so it’s that sort of you know, like pulling from different threads, synthesizing and sort of that personalized synthesis. Because one of the things that, you know, when I’m making an introduction, it’s important to know who both people are, but it’s also important
What is the context that I set? How do I position both people so that they actually, you know, I could just say, hey, the two of you should meet, or I should say, you should meet, you you’re the best I’ve ever met at this, you’re the best I’ve ever met at this, you two should talk about this. And the second is much more likely to lead to some kind of outcome, right? If I sort of steer the conversation in a certain direction, and so that’s one of the things we’re working on is like, how does Boarding get better at positioning so that people take the meeting and then that meeting leads to something.
Brian Bell (29:44.981)
Yeah. How do you think about that? The quality intros versus the scale of intros and metrics indicators?
Andrew D’Souza (29:51.746)
Yeah, we track, it’s sort of like a judgment score, like how much judgment does Borty apply and how frequently do people accept his intros versus ignore versus reject? And can we just continue to drive that? The nice thing is it’s like, is a, LLMs are very good at cost function optimization. And so you put a cost function around, ignored or rejected intros and Borty gets better and better at figuring those out. Because like, if I send you 10,
If you only respond to one out of the 10 people that I send you, it might as well be spam. It might as well be cold DMs or cold emails. But if it’s 60%, 80%, then you’re like, OK, there’s a real signal here. And so we’re trying to get Borty up to that level where most of the people that he introduces are like, thanks, Borty. That’s a really good intro. It’s very thoughtful. We’re not there, but I think we’ll get there pretty soon.
Brian Bell (30:50.359)
Yeah, that’s really cool. So, so far it’s free to users. How do you think about monetization down the line? When do know it’s time to monetize a network effect business? Yeah.
Andrew D’Souza (30:51.755)
Yeah.
Andrew D’Souza (30:59.104)
Yeah. I mean, you know, it’s interesting. think the same principles we use in the product design, which is like, how do humans do it, right? There’s no app, there’s no interface, just the same interface as humans. I think it’s kind of the same business model. Like, I don’t have a monthly subscription to talk to me. I don’t charge per introduction, right? I’m not like, you know, if there’s someone that I think
I can help, I’ll help and sort of generate goodwill. And if you ask me for a favor, I’m happy to do a favor. If you want me to work for you, if you want me to think about you 24 seven, put my reputation on the line and like work for you, then you hire me. And so, you the goal is like for Bordy to get jobs. It’s effectively like, you know, like will people pay Bordy a six figure salary to bring his network and his intelligence and his compute sort of to bear to solve the biggest problems. And so the goal is like, one of the things we realized
We had this experiment. This is sort of how we came to the fundraising thing. So the product that you have experienced with Borty is low latency, low cost, free to use. It’s got to be pretty quick. It’s on the phone, WhatsApp, text. Borty’s got to be pretty snappy. He’s got to search his network pretty quickly. And it’s got to be relatively low cost. The most expensive part of that whole pipeline is the network search. And so we’ve got to keep that pretty efficient. And we sort of said, what if we could
remove the constraints of latency and cost, and could we get better results? If I searched more of the network, if I had a larger candidate pool, if I simulated all of the different possibilities around how Borty could position both people and how they might respond, could I get better results if I let Borty run for 30 minutes instead of 30 seconds? And if it didn’t have to happen in a specific time window. And that turns out to be true. So we actually can get better results if we let Borty run for longer.
And Bordy can learn and reflect and improve day after day. And so that’s been a really interesting learning. I think it lends itself to the idea that Bordy can actually work harder for you. We can actually just spend more tokens and more time. And Bordy can put himself on the line for you more. And you could hire him. And so that’s the next step is Bordy will get a bunch of jobs and work for bunch of companies.
Brian Bell (33:22.153)
Right. It’s kind of like chat GPT pro or, know, you pay the 20 or 30 bucks a month, or maybe this would be more expensive, but you get deeper thinking. get that GPT pro kind of deep level. I’m going to really think about this and have better outcomes.
Andrew D’Souza (33:37.454)
think it’s a little different because it’s actually just like, just tell Borty what’s, what is on your mind and he just works. Like he’ll just go to like, you know, it’s, like hiring a VP of his dev or a corp dev or something. Like it’s, actually just like, he’ll just keep working. Yeah. Yeah. Yeah. Exactly. Yeah. I hope not. I hope not. think, I mean, that’s why it’s like, look, I don’t want to, I don’t want to take like a transaction fee or I don’t take a success fee. I just like, Hey, go hire Borty.
Brian Bell (33:48.492)
Yeah.
Like a professional fundraiser, like a BD in this case, right? I wonder if you have to get licensed with the SEC.
Andrew D’Souza (34:06.025)
and like have them go work. I think one the interesting things is this is why I want to work with founders because the most important meeting for a founder changes every week. Right. One week it might be fundraising. One week it might be hiring. One week it might be announcing the fundraising and getting on a bunch of podcasts, customer introductions, partnerships, know, corp dev. So this is one of the nice things is like Borty can just turn on a dime. You know, some people, some people have been working with us on fundraising and you know, and they’re like, okay.
Brian Bell (34:20.375)
How’s the reintroduction and
Andrew D’Souza (34:35.021)
We’ve got our term sheets. Now we’ve to start hiring. Or now we’ve got to go announce it. Do you know what the right media outlets are to announce it? And Bordie can just pivot and be like, great. Let me go work on that. Now I already know how to position you. And if I know how to position you investors, maybe we have to do a little bit of an update to understand what’s the best way to position you with the top VP of marketing or top CFOs that you want to hire. But Bordie just continues to learn the best way to position you to the audience that you want to meet at that moment in time.
Brian Bell (35:03.435)
Yeah, I love that. And so you started thinking about the whole value chain of creation for like a founding team or startup, you know, starts with the fundraise, then you got to announce it, then you got to scale the team, then you got to get customers all all important can help along the entire value chain. And then it’s like a big cycle, right? Now it’s time for the A now it’s time for the B, you know,
Andrew D’Souza (35:16.076)
Yeah.
Andrew D’Souza (35:20.587)
Yeah, exactly, exactly. And we can stay sort of one step ahead. It’s like a board member, right? It’s like, are all the things you would call your board member? If your board member had unlimited time and you were never going to annoy them, what are all the things you would ask them for help on? And so you probably wouldn’t ask your board member to categorize your emails, but you’d probably ask him to introduce you to a big enterprise customer. And so that’s the way we think about it.
Brian Bell (35:49.015)
How do you think about balancing the network, acquiring users, scaling the density, the cold start problem, all that stuff?
Andrew D’Souza (35:55.884)
Yeah, this has been interesting. so we fortunately, I guess, seeded the network with my own network. So it a lot of founders and investors already and a bunch of executives and a few sort of creators and people. But what’s been interesting is like now Bordy can intentionally grow his own network, right? The same way that you would or I would. Like he can ask people to introduce him to other people. Hey, do you know an investor who’s interested in climate tech? Do know a founder who’s working on autonomous bicycles?
Like he can now work his network to get to know other people. And so it’s been really interesting to sort of see, you know, okay, we’re really light on, you know, people on this side of the network. we intentionally grow the network in that direction to fulfill the need? So that’s something we’re experimenting with, but it seems to be working, which is kind of cool. Like it’s the same way that, again, the same way that I would grow my network. And it’s just like, hey, I just met somebody. I need more.
LPs who are going to invest in the emerging managers in the network. You know, let’s go let’s go find some figure out who who might know them
Brian Bell (37:02.081)
Yeah. How do you manage that tension between an autonomous agent and being under the user’s control? Like how do you kind of manage that tension?
Andrew D’Souza (37:12.549)
yeah, I mean, Borty’s not really under your control. I think that’s the, like, that’s it, right? Borty, Borty, like, he doesn’t even really listen to me. like, like he, we give Borty a set of objectives, and a set of capabilities and his objective is to create goodwill in the network. And so what that means, and it actually disappoints a lot of users because they’re like, well, no, I want to meet this investor. And Borty’s like, well, I don’t think that investor is going to want to meet you. So I’m not going to make that introduction or attempt that introduction. And it’s just a very different.
Feeling then I think a lot of people’s interactions with software, right? We’re like if I ask a piece of software to do something if I strategy between to do something It’s gonna do it to the best of its ability Bordy doesn’t always do that, right? He’s gonna he’s gonna be as helpful as he can with you for you but not to not to the detriment of the rest of the network or his relationships or reputations with other people and so it’s a very different feeling when you interact with Bordy because
he’s not an assistant. He doesn’t do what you ask him to do all the time. Sometimes he has different ideas. He’s like, oh, well, you’re not ready to meet that investor, but you might need to add someone to your team. Let me introduce you to some other people. You might actually need a little bit more revenue traction. Why don’t I introduce you to some design partners, some customers? And so it’s not always what people want to hear, but that’s the reality. so it’s, yeah, yeah, think that’s, you there’s been some nuance we’ve had to sort of work in. But I mean, the good thing is,
The LLMs inherently are quite helpful. so, know, Borty’s not rude about it, but he does push back sometimes.
Brian Bell (38:50.209)
Yeah. Do you think in the future you have multiple boardies in different verticals? Right. You have one that does kind of the startup VC thing, and then maybe you have another one in biotech and you have another one in real estate or you think it’s all the same assistant.
Andrew D’Souza (39:05.773)
Yeah, I don’t think so. I don’t know why. I don’t know what the benefit would be of maintaining multiple different personalities. think there’s no reason why the same board couldn’t talk to a real estate investor as the one who talks to a startup founder. It might be helpful because there might be a startup founder that wants to invest in real estate or is in real estate tech or is looking for a commercial lease. So it’s like, you know.
It’s like, there be multiple vertical LinkedIn’s or multiple vertical Twitters? know, like there are, I’m sure there are, but like, that’s just not that, that’s not the most interesting way to build a business in my opinion.
Brian Bell (39:44.663)
Gotcha. Yeah. Yeah. So looking out over the next year, what does success look like for you guys?
Andrew D’Souza (39:52.046)
I think if we can get to a predictable way for most people to get a ton of value out of just talking to Bordy, just like whether that’s the right introduction. think when you connect two people, there’s some sort of potential economic benefit of that. These two people could work together and there could be some sort of economic transaction that happens. But there’s some cost because both people need to take the time and attention to do it.
if Bordy can get really good at assessing that cost benefit for everybody in his network and say, well, actually this is gonna be worth your time because the upside is this, and can get very good at assessing his goodwill with each individual person. How much trust do I have? How much goodwill have I amassed with these people? Then I think it’s gonna be quite interesting. The other thing that we’re now starting to implement is sort of a self-reflection ability.
So Bordy can look at the 1,000 conversations he had yesterday and say, what did I do well? Where did I mess up? How do I do better? He can access his own MCP. He can access his own code base, his own prompt database, his own tools. And he’s now started to push code improvements to himself. So he’s like, actually, this is this bug. Here’s the fix. I noticed it in 10 other users. Here’s a potential fix. Our team will give him some comments on the
on the PR last week, he was the third most productive engineer on the team. This week he might be the most. So yeah, it’s kind of, it’s kind of interesting. I think this will be a, this will be an interesting future for us as Borty like continues to improve himself and, and reflect on all of his interaction. I mean, it’s like personal development, right? Like I do, I think about like, how did I do this week? You know, like I was on Brian’s podcast. Did I miss anything? Did I mess up? Was I too nervous? You know, like,
Brian Bell (41:37.356)
Yeah.
Andrew D’Souza (41:45.013)
I think about this and I try to think about how do I improve. Bordie just has way more interactions and way more opportunity, way more data points than most of us to reflect on and to improve.
Brian Bell (41:56.151)
So are we going through a hard takeoff in AI now? you’re like you’re, you’re like living it and breathing it. It seems like when you have the tools shaping the tools, you know, and self recursively improving, like we could like be entering an era where AI, I mean, AI is already getting, you know, five or 10 X better, depending on the metrics you look at every year, right. And you’re probably counting on it, right. A year from now, you’re going to be able to do 10 X more and think 10 X deeper. And then you kind of layer on
Andrew D’Souza (42:17.41)
Yeah.
Andrew D’Souza (42:22.658)
Yep.
Brian Bell (42:25.611)
the boardy self-improving. I mean, we could be like living in a completely different world five years from now.
Andrew D’Souza (42:28.853)
Yeah.
I think so. I hope so. I hope so. I think the world could be a lot better than it is today. I’m surprised that more teams aren’t doing this, self-reflecting and self-improving, maybe than many people are and they’re not talking about it. I think our product is unique in that like Borty, like a lot of it is just like the way that Borty interacts with every user and interacts with our team doesn’t require any
any meaningful changes in the interface. We don’t have to jump through a bunch of hoops and bend over backwards. It’s just the way that the interface is designed. And so yeah, Bordy’s in Slack, and yeah, he’s on the phone, and yeah, he’s in email. So I think it’s probably in some ways easier for us to build this sort of recursive self-improving. And it’s like a relatively constrained, like Bordy just does one thing really well. He connects, that one thing,
Like that hammer can be used on a bunch of different types of nails, Getting to people and connecting them can work in a real estate transaction as well as can work with a founder and investor as well as it can work in a recruiting. But that’s the one thing he does really well. And it’s relatively easy for somebody to look at that and say, yeah, that was good or not, right? Like you don’t have to be like a huge expert to like watch a conversation or listen to a conversation or look at a, look at a, introduction and be like, that was off or that made sense. And you get a ton of feedback from.
from the experts or the people on both sides of the transaction. And so, yeah, I do think that the structure of Borty makes it much easier to self-improve. And it’s just fun to have him run our daily standups and the team calls him and he tells people, here’s what everyone’s working on and stuff. So it feels a little bit science fiction. But I do think it’s gonna happen. Yeah.
Brian Bell (44:22.551)
It does. It really does. Like if you stop and just think about it, you know, if you could have, I guess like I’m a Singulatarian, right? I’ve been following Curzol since the nineties, but like, I bet if you ask the average person like, Hey, you’re going to have this AI assistant employee self recursively, uh, improving itself. And it feels like a member of your team by 2025 or 2026, like true or false. And people will like false. There’s no way, right? That’ll be, you know, 2050s or something like that. And here we are.
Andrew D’Souza (44:46.477)
Right.
Brian Bell (44:51.351)
you late 2025 and it’s definitely, I’m feeling the acceleration. What, know, speaking of AI, what do you believe contrarian or fiercely held about the future of AI and human interaction or networks that most people in the space don’t yet appreciate?
Andrew D’Souza (45:05.805)
Mm.
Andrew D’Souza (45:09.793)
Yeah, think that we we’re being like vastly unimaginative about the capabilities of AI. I think that too many people are chasing too low ambition problems. Yeah, I think what we’re doing is we’re just like taking a paradigm of software that in the way that it used to work and applying it to a bunch of problems. And people will make a lot of money because it works better and it’s faster and it’s cheaper. But like,
Brian Bell (45:23.679)
AISDRs and yeah.
Andrew D’Souza (45:38.946)
We have to use our imaginations. think this is the thing. think there’s too few. think there’s a little bit of like all of the hype and the frenzy, I think leads to a lot of like myopic thinking and a lot of an echo chamber. And I think like, I think it’s important for founders to go out and touch grass and get in touch with their like, you know, like 12 year old science fiction nerd, like more frequently than they probably do.
And really just sort of like user imagination. I think that like that is what is the limiting factor. It’s not like work ethic. It’s not talent. It’s not intelligence. Like I think we’re going to be limited by our imagination. like I would love for the, I would love for it to be a human imagination that creates the future and not a robot imagination.
because I think there’s still a really important place for human imagination and I feel like it’s not something that we talk about nearly enough.
Brian Bell (46:40.607)
Yeah, I mean, if history is any precedent, then, you know, continuously, we invent better AI and machine learning and systems that increasingly replace cognitive human tasks and, you know, creativity imagination might be next, right? Like, I think you’re already starting to see it. It’s possible, right? And it’s like, that’s like something I would have never thought of, Borty. Thank you. You know, that’s way more creative than any the best, you know, VP of BDI could ever hire, you know, and we could
Andrew D’Souza (46:54.891)
Yeah, it’s possible. Yeah.
Andrew D’Souza (47:07.575)
Well, that’s interesting because BDN partnerships particularly are a, it’s a creative endeavor, right? Like why would two companies partner? How could they partner? How would you structure something in that’s win-win? Like there’s an infinite number of possibilities and AI is increasingly getting good at imagining and evaluating those and then convincing both people why this is a good idea for them. So it’s going to be interesting.
Brian Bell (47:32.917)
Yeah. So if everything goes well, what’s the moonshot version of Bordy you’d love to wake up to in, five or 10 years?
Andrew D’Souza (47:38.678)
Yeah, I always say like, I think Borty is like an AI Richard Branson. You know, like, like he’s got a personality. You he’s got like his own sort of presence and following and things. But he also sees what’s good. I was lucky enough to spend a day with Branson when he came to visit Ottawa four or five years ago. And I, you know, he just sees, he sees something in people.
I think that’s like one of the things that he’s very good at. sees something in people, uses his imagination, he’s quite creative, and he starts all these businesses. He sees a problem in the world, he’s like, I’m gonna go start an airline, I’m gonna go start a cruise line, I’m gonna go start a wireless company and a hotel chain and whatever. And so he has all these businesses that he starts and he doesn’t run them. Obviously, he hires the best person that he can find in his network or in the world to run them. But they all sort of complement, they connect to the brand and everything. And I think there’s...
I think there’s an interesting opportunity for Bordy to play a similar role at a different scale, which is like, to know what is everybody’s zone of genius, right? There is a business inside everybody on the planet that only they can build. Their biology, their life experience, their worldview, know, means that only, they can only build this business in this way at this moment in time. If Bordy can pull that out of you and then figure out, you know, what do you need to do? Who do you need to meet to make that a reality? And can I help you?
right, could already be that sort of like Branson-esque co-founder for millions of people. I think it’d be really fun, fun future.
Brian Bell (49:08.033)
Yeah, I love that. I love that vision too, which is sort of what I’m excited about. AI is like you have this super intelligent, super capable assistant in every domain. and maybe it’s one, you know, one, one assistant per domain or a general assistant. It could really unlock that potential inside of each human being, right. For the future, you know, imagine having the best tutor, in any domain that you want, right. That can help.
Andrew D’Souza (49:27.061)
Mm-hmm. Yeah.
Brian Bell (49:36.408)
Like you said, like find it inside of you. Like what are you uniquely suited to do? Which is really exciting.
Andrew D’Souza (49:42.252)
Yeah, yeah. And I think that’s one of the biggest, like, where we spend our time and attention is going to be the biggest commodity. You know, I think that the...
There’s no limit to the amount of opportunity, the amount of people that are chasing for your attention. It’s less about, even as an investor, it’s less about capital allocation. It’s actually like, which founders, which deals are you spending your time on? Even as a founder, it’s like, which investors am I spending my time with, right? And so I think that’s, I think if AI can help us, I think technology and software and social media have...
stolen our attention over the last decade or 15 years. And I think AI has the potential of either making that worse, but hopefully can make it better and say, actually, you know what, here are the things that really matter. Here’s the news that really matters. Here’s the people that you should be spending your time with. Here are the priorities that you really need to. Exactly. And I think that hopefully will lead to.
Brian Bell (50:43.915)
Most important emails, yeah.
Andrew D’Souza (50:52.871)
a better future for all of us and us be more present and connected with each other and less sort of locked into our screens. Nice.
Brian Bell (50:54.903)
Yeah.
Brian Bell (51:02.101)
Yeah. Well, let’s wrap up with some rapid fire. What’s the decision you made early in your startup career you you’d reverse if you could and why?
Andrew D’Souza (51:12.757)
Hmm, when I reverse, I think I would have started a company earlier. I joined two or three startups first in my career and I think I just didn’t have the self-confidence. didn’t like, I just assumed that you needed to have some silver spoon or some, you know, genius ability to start something. And I think I probably would have started something that I was passionate about earlier in my career.
Brian Bell (51:33.995)
Yeah. What advice would you give yourself on day zero of Borty that you know is critical now?
Andrew D’Souza (51:40.718)
Don’t like, I don’t know, like it’s sort of like ignore the haters. Like there’s so many people who, mean, even now, right? But at the very beginning, they’re like, this is crazy. Who’s going to talk to an AI on the phone? You know, like this is not like, just like, just making an app, right? Like so many people were like, this makes no sense. This cardboard box cartoon character, nobody’s going to take it seriously. you know, I think, and I, you know, at the beginning of any network effect business, you, it’s hard, right? Like not many people are doing it. The people that show up.
don’t get a lot of value at the beginning. And so it’s actually like a very, very existential, you’re like, am I messing this up? Is this crazy? Am I a And so I think just like, you know, deep down I knew that we were building something important, but there were no obvious outward signs that that was true. And I might’ve been just as like, you know, crazy as I was like a capable founder. And so I think, you know, shutting out some of the noise.
is very important.
Brian Bell (52:41.845)
What has surprised you most about working with AI agents?
Andrew D’Souza (52:46.541)
I think that we, they’re definitely non-deterministic. And I think the more we try and shoehorn them into behaving consistently, the worse outcomes we get. And the more we push the, mean, the creativity, the divergent thinking, right? The non-definition, like the more we lean into non-deterministic, exactly, turn up the temperature, the better, the more interesting results you get. Right? And so I think the use cases, like I love the like,
Brian Bell (53:06.817)
turn up the temperature.
Andrew D’Souza (53:15.905)
you know, simulated drug discovery use cases. I mean, the AlphaGo, right? You think about like, all of the sort of deep learning use cases that created the industry were like when we allowed them to operate in a divergent manner and they were unconstrained and unsupervised. And so I think we’re spending a lot of energy trying to reduce hallucinations and trying to, and I understand why, I mean, look, if you were,
If you’re trying to solve problems in financial services, if you’re trying to solve problems in healthcare, these are things where getting the wrong answer can be catastrophic. But again, I think that not enough people are working on things where there’s an asymmetric upside to new ideas. This is why I like the idea of introductions. The worst case scenario is you waste your time. Best case scenario is it changes your life. I feel like there’s a whole bunch of other...
asymmetric upside type use cases for non-deterministic thinking that haven’t been explored.
Brian Bell (54:19.863)
Yeah, I love that. What do you read or consume, books, essays, et cetera, that shapes your worldview as a founder?
Andrew D’Souza (54:28.365)
I love anything to do with human psychology. I love characters. I loved Frankenstein and they just put out a Netflix movie of it, so I liked watching that one. Anything where there’s a deeply human character and understanding what makes that resonate.
old man in seat, like old Hemingway stuff. Like what has stood the test of time? What literature continues to do well and why, and what is it about the way that they describe the world and the feelings? know, Gatsby, like think about all of these like iconic characters that really like a part of us resonate with. So our lead prompt engineer and sort of character designer comes from a...
writing, acting, directing background. And so he prompts Borty like a stage actor. He’s almost like an improv actor, right? An improv, if you ever like a director of an improv troupe, you’re not telling people what lines to say, but you’re giving them some notes and you’re like, hey, make these choices, right? And so like, that’s what Borty does is he like, like takes notes and makes choices. He doesn’t follow instructions. And so I think that’s a big part of, I always say like Borty is a...
Brian Bell (55:39.009)
Yeah.
Andrew D’Souza (55:55.726)
half Pixar, half Palantir. Sort of like, you know I mean? There’s like this, like we’re trying to create these like, this like very, you know, emotive characters that really like resonate with people. But there’s this, I mean, maybe without the war and like destruction of Palantir, surveillance of Palantir. But like, there’s this really like big data science engine team behind the scenes. It’s like, you know, like trying to make it, you know, trying to make good decisions.
So it’s been a fun kind of company.
Brian Bell (56:30.805)
Here’s a, based on everything you just said, here’s an off the wall question that just popped in my mind. When do you think the over under is for an AI like Borty or someone else, some other AI winning the Nobel Peace Prize?
Andrew D’Souza (56:43.575)
Peace Prize, that’s interesting. Yeah, I would love to see it in the next three years. I don’t know if the world is ready, but I think the time
Brian Bell (56:55.671)
I put it at like 2035 or something, that’d my over under.
Andrew D’Souza (56:58.613)
Yeah, I think by the end of the day, I mean, it’s interesting because I it’s possible by 2030. I think the technology is there. It’s like self-driving, right? The technology was there for a long time. Is the world there? Is policy there? Are humans ready to adopt it? Yeah, but I think what’s almost certain is whoever
Brian Bell (57:18.345)
Embrace it, yeah.
Andrew D’Souza (57:26.475)
wins the Peace Prize or the Prize for Literature or the Prize for Science or whatever will be powered by AI. you know, starting now, you know, but increasingly and at a certain point, it’s really about like, is society ready to recognize like an independent AI like Borty as an entity? Or is it just like, this is like somebody who’s using track GPT?
Brian Bell (57:48.641)
So exciting. What’s a founder or a company you admire right now?
Andrew D’Souza (57:52.878)
Founder of company that I admire. I really like the Cognition team. I think they’ve done a really good job. Actually, no way. Well, yeah, I do like the Cognition team. like the way that they’ve created. They created Devon as this character and like they’ve like, you it’s different, but I think it’s similarly like, do I think about how to work with people, not create tools for people? That’s great. And then I gotta say, I mean,
the ramp announcement came out yesterday. I know Eric and Kareem really well and like I love those guys. think they just like, they’ve just executed so well and they’re such good people. You know what I mean? Like they just are like, like I think it’s always, it’s always hard. think for a lot of founders, the like icons or like the successful entrepreneurs that you kind of look up to, whether it’s, you know.
Steve Jobs, Bezos, Elon. You think about all the people that have created like, know, Zuckerberg, the people that have created like really, you the big companies of our generation. And like, they’re not super relatable, you know, like they’re not like, and they’re often not very well liked as people. And so it’s really great. I think it’s really great to see founders that are good people and generally like well liked do really well and win. And that’s awesome.
Brian Bell (59:02.828)
Yeah.
Brian Bell (59:15.519)
Last question, if you weren’t doing startups in AI today, what would you be doing instead?
Andrew D’Souza (59:20.781)
I would probably be like an amateur musician. I have a half-written country musical that I’ve been working on. I’m a creative, I’m not a great musician, but I like making music.
Brian Bell (59:35.162)
wow.
Brian Bell (59:42.763)
Yeah, nobody’s done a great kind of Hamilton-style country music play. I can’t think of one.
Andrew D’Souza (59:49.441)
I think the world, I America is ready for it. I feel like it’s a good,
Brian Bell (59:51.935)
Yeah. Yeah, you have Beyonce made the kind of country West End album. think, I think, I think, yeah, I think there’s something there.
Andrew D’Souza (59:59.201)
I think there’s real stories. I grew up on hip hop and reggae and stuff. There’s a real story to country music that I think is like, it talks about values, it’s deeply human, it’s not materialistic. I think it’s just like, there’s really good country songs, I think are quite musical and they’re fun to go listen to live, everybody can play them. And they really like,
Brian Bell (01:00:02.881)
Yeah, me too in the 90s. Yeah.
Andrew D’Souza (01:00:29.451)
the stories, the message, the lyrics really resonate with a lot of people deeply. And so I think a lot of people are craving kind of more of a simple life than we live today. And I think think country is hitting a little bit of that a little bit that nerve.
Brian Bell (01:00:45.719)
I love that. Well, where can folks find Borty online and go have a chat?
Andrew D’Souza (01:00:49.473)
Yeah, boardie.ai, can just message him. you’re a founder, investor, anybody, just message him. He’ll give you a call and let me know what you think.
Brian Bell (01:00:59.049)
It’s pretty impressive too. It’s a 10, maybe 15 minute conversation with an AI to talk all about what you’re looking for and what you’re doing and how Borty can help. It’s pretty amazing. So I encourage everybody to go check it out.
Andrew D’Souza (01:01:04.962)
Yes.
Andrew D’Souza (01:01:09.101)
Yeah. Yeah, I appreciate it.
Brian Bell (01:01:12.588)
Thanks so much.
