That Tech Pod
Welcome to That Tech Pod, a podcast co-hosted by Laura Milstein, Gabi Schulte and Kevin Albert. Each Tuesday, That Tech Pod will feature in depth discussions about data privacy, cybersecurity, eDiscovery, and tech innovations with heavy hitters in the industry. Subscribe so you don't miss an episode! Visit thattechpod.com for more information.
That Tech Pod
AI Knows Everything. But Does It Know Who Can Get You in the Door? with The Swarm's Olivier Roth
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
This week on the pod, we talk with Olivier Roth, Co-Founder and Chief Growth Officer of The Swarm, about what happens when AI starts mapping the human relationships behind business.
Companies have mountains of relationship data sitting inside their employees, customers, investors, partners, advisors, CRMs, and inboxes. But knowing that two people have exchanged emails or worked at the same company doesn't necessarily mean one of them would actually pick up the phone. So how do you teach AI the difference between a connection and a real relationship?
Olivier explains how The Swarm uses AI and relationship intelligence to help companies find the people who can actually open doors, rather than relying on LinkedIn searches, cold emails, and endless asking around. We also get into the uncomfortable questions: When does relationship intelligence become creepy? Who actually owns the relationships inside a company? And can AI really understand trust, credibility, and willingness to make an introduction?
If AI can eventually understand not just what we know, but who we know and how those relationships actually work, cold outreach may start to look very different.
Olivier Roth is the Co-Founder and Chief Growth Officer of The Swarm, a technology company helping businesses find smarter ways to grow through the people and connections they already have. Olivier works at the intersection of AI, relationship intelligence, and business development, helping companies uncover warm introductions, move away from cold outreach, and get more value from tools they already use, including CRMs and Clay.
Hello and welcome to that tech pod where we discuss all things e-discovery, data privacy, cybersecurity, tech innovations, and anything else that we feel like we should be discussing. To help us, each week we bring on heavy hitters from the industry to help us break down these topics. And today, Kevin, who are we talking to?
SPEAKER_03Today we're talking to Olivier Roth. He is the co-founder and chief growth officer of The Swarm, a technology company helping businesses find smarter ways to grow through the people and connections they already have. Olivier works at the intersection of AI, relationship intelligence, and business development, helping companies uncover warm introductions, move away from cold outreach, and get more value from the tools they already use, including CRMs and Clay. Olivier, welcome to the show.
SPEAKER_00Thanks for having me. Hey Kevin.
SPEAKER_01Oh, the same as France.
SPEAKER_00Uh yeah, it's the closest thing to France that I found there. So yeah.
SPEAKER_01Yeah, that's what everyone's like. France, San Francisco, basically the same. So that makes sense. Well, for people that don't know who you are, can you tell all of our listeners, you know, who are you? What are we doing here at the Swarm? Like, what are you building today? What's going on?
SPEAKER_00Yeah, so I'm co-founder and chief growth officer at the Swarm. Uh we've been building this for the last five years now. And it is a technology that lets you essentially know who knows who. And there's a bunch of use cases, but like the main one is sales. It's like I have a list of people and companies I'm trying to get in touch with to sell to. And I want to know if there is a warm introduction to them. So that's what the swarm at super high level is doing, combining the relationships of your investors, your advisors, your customers, your team. So you can see all of the potential intros that are around your company that sort of were always there, but you just didn't know. So that's where we've been uh busy building for five years.
SPEAKER_03So I mean, uh again, I guess the concept is pretty simple, right? You have all these relationships inside your employees, your customers, partners, and everything, but most of that network is invisible. How did you come to see relationships as a technology problem that could actually be solved?
SPEAKER_00Yeah, we we see that the best salespeople are very good at leveraging their networks, right? They are they they they they know a lot of people typically, right? They they're very good at like nurturing their networks and so on. So at the individual level, that's something that that we've seen happen since like the beginning of the business. I mean, the oldest thing in in business is relationship. It's like it's what it is. And the best people are the best salespeople are extremely good at it. But a lot of times uh referrals, uh intros, they sort of happen like opportunistically. I mean, it just like happens that you figure out that like your colleague knows someone that you're trying to get a hold of, and there is no real way to do it at scale. There is no real way to sort of like systematize that, uh mapping the network, looking through the network at the company level. And so we took something that's happening already like organically in business for the longest time, and we we built technology around it. That's basically what a technology does, right? It's like taking something manual and complicated and making it like easier and scalable.
SPEAKER_01I feel like easier is the is the real word that for me, am I like, is it hard? That's what I want to know with anything, because everything sounds so complicated these days. So I'm always just kind of like, is it though? Like, tell me, tell me more. So imagine I gave an AI one simple instruction. Like, I want to meet the CIO at ABC Company. Traditionally, I'd search LinkedIn, I would search maybe my CRM, I would ask around, maybe send a bunch of cold emails. What can AI do differently if it actually understands my relationship network?
SPEAKER_00Right. So LinkedIn, I guess, is the the elephant in the room. Um, everybody thinks LinkedIn because that's the closest thing to the network of your network that you can see. I mean, you can see on LinkedIn you're up to your third-degree connections, right? And so most people are trying to use LinkedIn to see, okay, let's find a mutual connection to the person and ask for an intro. That doesn't really work. That's what people like try it, but it doesn't really scale up because LinkedIn is not a network, it's an audience. Like LinkedIn, on LinkedIn, you're getting, of course, connections, but like over the years, you're you're getting you know, some people that you've met at events and conferences, some people that you were on on a call with but never spoke to them again, some people you know that you you you may have reached out to for sales, but you like you don't know them, and so on. And so LinkedIn is that this this strange accumulation of uh connections, right, with no context. And so if you try to look through mutual connections, you might find sometimes you have like 50 mutual connections with a person. So, first of all, you have to manually sift through who you might ask, but then you don't know who you don't know if the person that you're gonna ask actually knows the the person on the other side, right? And that's because of that like looseness of LinkedIn connections by nature, and because it's a social media, uh that LinkedIn wants you to have more followers and and more connections just to to sort of like grow your audience, but it's not really where your your network lives. So the swarm is uh going to uh map the former colleagues of your the people you add in your swarm. So you you add your investors, your advisors, your customers, your team members, people who know of your business and may be interested in in its success, right? So that's a pretty big list. Okay. And then we map people who work together, the same company, similar functions, similar seniority, similar geographies. We map people who went to the same school the same year, we map board overlaps, like so people who are investing in you know in uh in different companies, we we can see that too from our data. And on top of that, if you invite people to connect their LinkedIn, they can. And if you invite people to connect their email, they can. So you've got former colleagues, school alumni, investor overlaps, and then optionally LinkedIn connections and email and counter context. So in that setup, LinkedIn connections are there, but they're just like a piece of the puzzle, right? They're just one layer of the network. And we can score relationships. We can see, oh, there's a linked connection, but there's also an email, or there is a LinkedIn connection, and there's also like a work overlap for five years. So the difference is it's gonna give you a much better sense of who knows who. And it's a controllable set of people, you know, we call them the reconnectors that you add initially. So it's not like this this loose, you know, set of connections that you have, and then there are connections that are also loose. So it's a much like it's a much different data set than than LinkedIn. But we incorporate LinkedIn as one layer of that data set, if that makes sense.
SPEAKER_01I mean, literally, I'm pausing for a second because I was trying to get this thing to download. For those of you who are listening and are like, what is she talking about? He has this microphone that legit looks like a robot to me. And I can't stop thinking about it. And I just took a snippet of it, just so you're aware, and I put it in and was like, what robot is this? And I was like waiting for it to download and was like, I need to have the answer. And then it said wally or a vintage tin robot vibe. And then I Googled that, and it doesn't look like WALL-E or a vintage tin vibe, but this other one comes in, and I'm gonna send you a snippet of this so that you can see. And everyone who's listening, you're probably like, What in the world is she talking about? But you it's like a very cool looking microphone, but to me, I'm seeing a robot. How do I how do I get this in a chat? Hold on, you guys have to see this. And I know this is not Kevin. How do how do I get it to let me to chat? Oh, here it is. Chat chat. I'm new to technology, it seems. Okay, that's your microphone to me.
SPEAKER_03Isn't that your first zoom?
SPEAKER_01That's what I'm seeing, and it's not even like what I was originally thinking, because that's kind of like a demon version of it, because it has like the black you guys listening. Here's the moral of what I'm saying: technology is so ingredded ingrained in our brains that I'm seeing a microphone and I'm like, that's a robot. And he's like, No, it's actually just a microphone. So anyway, back on track of what we're talking to. Okay, I want to further this. So if a connection isn't a connection, you know, like if we take that approach, how do you teach a machine the difference between like this person follows me on LinkedIn and this person would actually take my call? You know what I mean? Like, how can you make it see I know you versus I wish I knew you, but I don't know you?
SPEAKER_00Yeah, so on the swarm, you sort of define your your connectors. We call connectors people you're gonna ask for intros. So that's a little bit of homework up front of like, okay, you're gonna you're who are who are your connectors? And the the definition is people who know about your business, about you, and are uh interested in the success of the company. So it's like to some capacity. So so it's like employees, I mean maybe like executives, if it's a bigger company, maybe you just do execs, you know, that typically have a strong network because they've typically been in their career for longer and so on. Investors, uh advisors, people who are on your cap table, if you have investors, or happy customers, because they didn't they give the best referrals because they've experienced your product and your solution. So you define your connectors, you add them, and then we map the network of your connectors. So you see, so you define like who you're gonna ask up front. So it's different than LinkedIn in the sense that you're gonna see you're not gonna see through mutual connections that that you may not know because you know you have a lot of connections. You're you're just looking at like the network of the people that you're gonna feel comfortable, you know, asking at the company. So that's the connectors, and then we map the network of the connectors. And then like to your question about like AI, I mean, we use AI. It's kind of funny because we use AI for this technology that's actually about connecting people. So I love it. It's like you know, a lot of AI is kind of like faking relationships. We're sort of like using relationships, we're sort of using AI to like try to surface relationships that are there that just do you know is humanly impossible to know. But yeah, the AI is gonna help with the mapping and the scoring. But then the intro request and the the handshake and the meeting and everything else, like that's people. Where we point AI to the swarm is really just uh well, it's all of the hard stuff. It's like figuring out who knows who, scoring, you know, sort of like prioritizing the relationships that are gonna that are gonna most likely yield like a warm intro and and so on. That's the hard stuff that frankly took us years to to build.
SPEAKER_03Well, so that's the piece of it that I think is really confusing, right? Because you're um my worry is that you're basically confusing relationship data with actual relationships. AI can tell that I know somebody, but does it know that they actually trust me and that they're gonna to make a warm introduction? How do you know that the swarm is actually improving business development rather than just sort of making the guessing process more sophisticated?
SPEAKER_00Well, we're you you define your connectors, so you're gonna only ask people that you know know you and you know, right? So you you're adding, I mean, if you're a company of like, I don't know, a hundred employees, you add all your hundred employees to this, you add your investors and so on, you're gonna have a few hundreds of people that you're gonna go and and and and ask. Now, how do we know that they know the person on the other side? That's like based on those data points that I described. So it's based on work overlaps, you know, and then function geography, seniority levels, education overlaps, and and all of those layers of a person's network. And it it is it is a sophisticated sophisticated guessing. I like how you put it. Like, I mean, it is just a much, much better sophisticated guessing than than what exists, you know. And does it mean like we know that the person we know that we know who everybody knows now? It sounds like uh, you know, it's it's not quite quite true, but it's it's just much better at at guessing, and that impacts business because you can then you know spend the like salespeople can spend time on asking for intros that are actually going to yield something, like a meeting, obviously, or I have a deal that's like in progress, but it's stuck, I'm gonna try to get an intro to someone like more high up and stuff like it's it's not just top of the funnel, it's also like okay, how how do I move a deal forward? And then even when I close the customer, if it's like a big company, how do I expand the account? How do I expand to other other places and so on? So, yeah, that there are there are many ways how it you know top up and down the funnel, how it impacts the the business for good. And we see it in the conversion rate. I mean, you see the con you see the conversion rates after warm intros are five times better than without the warm intro. You see the customers tend to stay longer. There's also like people who give warm intros are happier here. And so if you ask a customer for warm intro and it turns into a business, they tend to stay longer as well. So the the person who gives the intro, I don't know, I mean, I think it's like because then they have more skin in the game in a way, and they they just kind of like they tend to stay longer as a customer too. So it's like, yeah, it's the magic of you know, intros and relationships, I guess. That we we see that in tech business in in in many great ways.
SPEAKER_01I think everything you're saying makes sense, and like I'm thinking out of such like a sales thing, but I'm also just like very fascinated by you know, like when people say, and and I'll just ask this like at what point does quote I'm I'm constantly air quoting for y'all who can't see, and I feel like I have to say quote so you know what I'm doing. But at what point does quote relationship intelligence cross the line into being creepy? Like, if if an AI knows that I worked with someone 12 years ago, that we exchanged a few emails and that we both went to the same school, should it really be suggesting that person as my route into a deal? You know, like where's the line but between like this is super helpful and whoa, that was really weird.
SPEAKER_00Yeah, yeah, I I hear you there. I mean, the way we set it up is we and and what we're really trying to do right now is like optimize for the connector experience, what I called. So the the person you're asking the intro to, they should they should be in control. So, first of all, when you set up an account, you you can add your connectors and then you can have you can send them a link, and it is optional. And I mean, part of it is also culture. You don't like send to your employees like an email saying everyone, everyone, like we're gonna we're gonna use your network. I mean, okay, if you're if you're talking to a sales team, like they they're sort of incentivized to like do it, but but the like the CFO or you know, they're they're not necessarily gonna participate, and that's fine. I think like if we make it easy for the connector to say, yep, I'm in, or you know, no, I I respectfully decline, that's that's that's the best way we we protect you know their their experience. And frankly, yeah, obviously, like if someone asks you an intro and you don't want to give that intro, you just don't, you know, it's like it's where it stops, and then you you opt out. But like controls of like are you opting in or not? And then do you want to import your LinkedIn connections or not? Do you want to connect your email or not? Do you want to just do LinkedIn connections but not email? Do you want to get an intro request every month at the most, every quarter at the most? Hey, guess what? If you're like a if you're an external you know stakeholder, you're not an employee or whatnot. Or maybe like if you're if you're even if you're an employee, like do you want to be compensated if the intro turns into a million dollar deal? Do you want is there a ref share program and so on? So I mean, at this point it's just business, basically. It's just like, okay, I opt in, I upload my LinkedIn, I upload my email, I put my calendar. Yes, I want to participate in the RevShare program, and if there's a million dollar deal, I get 10% of it, great. Or, you know, or or no, I'm not gonna participate in this, that's fine. So yeah, I mean, that's how we we're we're we're building it with you know our customer, I mean, just working with our customers, getting that feedback, building the features that make it like anything but creepy, you know.
SPEAKER_03So who owns the relationship data at that point, right? If my company combines the networks of all the employees, right, we're 500 people, and those do those relationships become a corporate asset, or do you still think that they are fundamentally the personal relationship belonging to the individual who entered that relationship in the first place?
SPEAKER_00They belong to the individual 100%. I mean, and and by the way, if you have a as an individual, if you have a swarm account, you can get re we can remove yourself from an account, and then if you move company and you still have your swarm account, you can now you take your network with you, you know. So yeah, it's a hundred percent people ownership, and that's the thing that makes them that's the thing that actually makes them so powerful, is because it like my relationships are my relationships, like the swarm or anything is never gonna LinkedIn or whatever is never gonna never gonna steal that. And that's what makes it like the most valuable asset for every company, because your competitors they can't steal your relationships, even if they have the data. Let's just say like someone like gets the whole data, the swarm is secure and so on and whatnot. We could talk about privacy and security and stuff, but let's just say like a competitor steals the data set. They can't do anything with it because if they go and ask your investors for intro, they're like, okay, well, we're not gonna so the relationships of your connectors, you know, how is as we call them, is both the most valuable thing, the most like underutilized thing, and the uh and it's uh it's it's a it's a moat in a way, in you know, like your competitors can't steal it, can't have it at the person level and also at the company level. But yeah, we we we we need to put as a software provider and you know, tech builders that we need to put the right like we just need to architect in a way that's like gives people controls, opt in, opt out. We have a do not sell like link on our website, people can like fully opt out. We have like we're data brokers in like many states, we've done the SOC2 compliance, we can talk about all of that, but that's one of the the most painful part of like building the business over the last five years and make it make it like you know rock solid. So so it's so so it's all you know not only secure but also like compliant with the latest latest regulation. Um and I know that's you guys are obviously like experts in that.
SPEAKER_01So compliance is always key. I always gotta start with how does this sound and can we make it work and can we make it work going into that compliance tie? If AI gets like really good at finding warm introductions, could we then at that point say that cold outreach eventually just becomes the weird old-fashioned way of doing business? Like, can we just say we don't need that anymore? Because AI is basically handling it. And if you're doing the cold outreach, like now you're the weird one, whereas like before you were great when you were doing it and it was like, oh, this is clearly a bot. But if it gets so good at it, could we reverse that?
SPEAKER_00Yeah, I love it. I think like over the last few years, just like with AI, especially, right? Cold outreach has has become weirder. And I'm gonna quote like one of our customers, Christy Lux, who is who owns an agency, like a RevOps agency, and she said like the other day, you can tell me what I had for breakfast, you know, you can tell me who my neighbor is, you can tell me like the name of my cat in a cold email. It doesn't matter if I if I don't if I don't know and trust you and I'm not interested in like what you're selling, I'm that's not gonna move the needle, you know. That's just not yeah, cold email has gone weird. I can think I'm getting some I'm getting some like some and some obviously like there's some some borderline like illegal stuff, I guess, you know, some scammy, scammy stuff, impersonating and whatnot. Like is is is weird. I like that reframe.
SPEAKER_03As a last question, where do you see the swarm going next? What becomes possible once AI doesn't just understand the information inside a company, but also understands basically the human network that surrounds it.
SPEAKER_00Like the next thing we really want to crack is like this like air traffic control of intros. Because if we go like we we work with a lot of startups, and then we're starting to go up market, bigger companies are like, okay, we want to do this and so on, and it's like a they can see that it's a company of advantage and whatnot. Like, okay, now you have a hundred sellers, two hundred sellers, 300 sellers. Can a salesperson ask an intro to an investor? Probably not. Maybe there is a check control and so on. So we are like we're working on like that, like orchestration. So we we I think like big picture, the arc is that we got the data, which we're still working on, but like the orchestration of intros, this air traffic control of intros inside a bigger company with the right checks and balances in place, that obviously foreign enterprise are gonna be like uh much bigger than like a startup. That's what's next. And that we have a lot to build there, and we're excited to dive that with uh with our customers. So lots to build. We're not done. Just getting started.
SPEAKER_01We're not done. I love that. I think I think that should always be the approach. Like, we're just getting started, there's lots to build. Anybody who's like, we're done, this is where we're at. I'm like, well, you better run quickly because that won't in this kind of world, that's then you're gonna fall behind in the next two hours. But you've been excellent. Thank you so much for coming on. I can't wait to visit you in San Francisco or France because they're one and the same and hang in real time. But this was lovely.
SPEAKER_00Thank you so much. Thank you so much for the chat. Thanks for having me.
SPEAKER_01If you would like to learn more about that tech pod, please reach out to us at contact at thattechpod.com. We would love to hear from you, or head on over to our website at www.thattechpod.com. Enter your email to subscribe, check out our merch, and just keep updated on the latest things with the tech pod and what we are doing. If you want to follow us on social media, head on over to LinkedIn.com slash that techpod, Instagram, or anywhere else that we may be on social media. Kevin, is there anything else our listeners can do?
SPEAKER_03Yep. Go to wherever get your pods, Apple Podcasts, Spotify, or what have you, and leave us a review, leave us a rating, let us know how we're doing. All that stuff helps us with the algorithms and is much appreciated. Thanks everybody.
SPEAKER_01Thanks, everyone.