Dimitri_Enterprise_Ready
===

[00:00:00] 

Grant: All right, Dmitri, thank you so much for joining me.

recording-1_2025-07-29_07-40-35-1: Thank you for having me.

Grant: right. So, we normally do a list, do this a little bit differently where I have you kind of go through the background. We'll get, we'll get into that later.

I'd love to just start with a bit about big id, kind of where you are today, what's going on with the business, gimme some context, you know. You've been in business for, what, 10 years? I can't re you know, just

recording-1_2025-07-29_07-40-35-1: Eight years since 2000.

Grant: You've raised, you know, hundreds of millions of dollars. Your teams, how big, and then what are you really going after today?

Like, what's the core business? And then we'll talk about some of the history in a little bit.

recording-1_2025-07-29_07-40-35-1: Sure. So, today we're kind of focused on what we described as connecting the dots in data and ai, but set another way. We basically focus on helping organizations get visibility control around data and ai. So [00:01:00] we help them uncover what data they have, what, what kind of data, the context of the data who's using the data, who has access to the data.

And now we do the same thing for ai. So we help 'em find models, we help 'em understand, if the models are sanctioned or unsanctioned. We help them understand the risk around the models. We help them understand what data is going into the models. And then we kind of build the bridges. But we started off very much data centric and now like every other company under the sun we've expanded more and more functionality in the AI arena, especially gen AI specifically.

Company size, a little bit under 600 people raised about 400 million. Started formally in 2017. Basically started selling in 2018. That was year one for us. We did about 5 million in that year. And I'm in Miami. I moved here during COVID. But the company's fairly distributed. We had kind of global headquarters in New York, but we've kind of changed that and frankly just gone with a whole decentralized virtual thing.

It, it works for me and I think it works for most of the people [00:02:00] in the company. And so today we're like on every continent except Antarctica. And we have customers everywhere or Middle East Asia Europe north America, Latin America as well. And the customers typically care about data security and compliance.

That's usually the, the driver. And now it's increasingly data and ai, security and compliance and governance. And that's kind of the, the big shift over the last year or so.

Grant: That's great. Thank you for that, for that context. And, and this is your, you've started one other company, two other companies. I know you, this is

recording-1_2025-07-29_07-40-35-1: Two, two other companies. So I had an API security company, I had a network security company. Security seems to be a, a theme. Next time, who knows, what would I do? Maybe if there is a next time I'll, it'll probably be more AI centric. But yeah, multiple security companies. First two were in Canada where I lived prior to moving into New York in 2013.

And then obviously I've, I've kind of lived in many places. Seattle, Vancouver, Winnipeg, Montreal, New York. And now, now I'm kind of [00:03:00] probably settled down in Miami.

Grant: A real nomad. You've been everywhere, huh?

recording-1_2025-07-29_07-40-35-1: That's right. Well, that's, that's just in North America. Then there's all the places in Europe and the, and Middle East I've lived so that, that comes next.

Grant: Wow. All right. Wild. And then, I mean, te tell me, I mean, I know the business really was accelerated by sort of like everyone's concerns about GDPR when you first launched, right? That was kind of the initial like real catalyst for growth. Can you tell a little bit about that story, just in terms of like catching that market trend, timing, what that meant in terms of product and scaling?

recording-1_2025-07-29_07-40-35-1: Sure. So when we started so one thing I've learned is when you kind of look at beginning a company, there's kind of two options really. And, and you, most of your audience have probably heard this framing around blue Ocean, red ocean or greenfield brownfield which sounds terrible. But you basically have those two options and, and put another way, typically it's like, am I gonna do something [00:04:00] novel where I have to evangelize and basically invent something new, which is fantastic 'cause it creates a clear path.

Usually legacy vendors. Aren't addressing that issue 'cause it's new. Or I'm gonna basically tackle a legacy market with maybe like open source or taking it to the cloud or creating an ag agentic version of that. That's kinda like the two options usually companies have. So when we were kind of beginning big Id, I wanted to do something new.

I sold my last company, I, I was kind of, my term at the company that acquired us was expiring my vetting was expiring. And you know, we knew about me and the guy that I was gonna start a company with knew that there was something kind of boiling in Europe. I knew Regulat around personal data specifically, and he and I also know enough about the data and identity and other parts of the security landscape to know that there hadn't been much innovation.

Data security since 2003, 2004, and enough of a history buff, even in boring places [00:05:00] like security to know that the motivator for some of the innovation in, in data security like DLP, which was kind of big, you know, Scott got it started around 2003, 2004.

Grant: Data loss prevention, right? Yeah.

recording-1_2025-07-29_07-40-35-1: Data loss prevention. Right. Stopping data from, but, but part of the motivation was things like new regulations for our credit cards like PCI and, and you know, those were the first kind of like, Hey, we're gonna, we're gonna crystallize this kind of fear of certain types of sensitive data.

And so regulations have a way of providing a tailwind, right? They create a catalyzing event. There's a compelling. Instance, there's a penalty usually a fine, maybe jail time. there's a deadline that you have to be in compliance. And so as we were kind of thinking a little bit about where we wanted to do something, I thought, well, look, data security hasn't seen much innovation in over a decade.

You know, beyond DLP, there was only really one other piece of innovation, which is Varonis, when they were doing basically insider risk around 2008. So again, we were starting big [00:06:00] ID around 2017, so that's a, that's nine years. And so we, we knew about this event and we, and we knew that it was gonna make everybody subject to it.

It wasn't just covering, it was covering any organization that had European citizens, 28 countries at that time now, 27. And so, yeah, so that was kind of, we said, okay, you know, what, what can we do there to be innovative? And then we also realized that. The GDPR focus on personal information wasn't just random.

It's because when you think about criminality, especially digital criminality and, and people trying to steal data, personal data. Now it could be pass passwords, it could be, you know, your address, it could be your credit card still represents the lion's share where people steal. Clearly there's other things like intellectual property and so forth.

But personal data is still the stuff that gets the headlines in the Wall Street Journal. And so, you know, you put 1 0 1 together and we said, okay, there's a story here about creating a better mouse trap for protecting and privacy, a per protection privacy, a personal [00:07:00] information API I. And that was the kind of that was the, that was the headline, that was the cover page of the, of the PowerPoint, the deck.

And that's how we got started and started kind of socializing with a couple of investors.

Grant: And then, I mean, from, from what I know about the backstory initially maybe the market didn't really, it wasn't like a resounding, oh, this is clearly obvious. Like, we need this, this is the, you know, this is the next, you know, unicorn. It was a little bit of a slower response for the first year or so.

Right.

recording-1_2025-07-29_07-40-35-1: so yes or no. So let me kind of, so we had pro, so, so take a step back. So one, we misjudged who the audience was. We were thinking that the audience would be security, and security really didn't care. The audience was actually privacy. So there hadn't been a historical kind of privacy buyer.

But to some degree our eyes were opened up by another vendor, a company called OneTrust, that was really kind of [00:08:00] said, okay, you know what? We could create a new buying audience from Chief Privacy Officers. So Chief privacy officers were a new thing. 

Grant: But you like had to identify that or something in the GDPR thing,

recording-1_2025-07-29_07-40-35-1: Yeah, so look, I'm gonna say my background is security and the company was thinking about the context of privacy as a security function.

But we didn't really appreciate that there was all of these privacy professionals and they were gonna be the budget holders initially. And so we actually did pivot a little bit to kind of cater to them. We found organizations that speak to them like the International Association of Privacy Professionals.

I will, I will give some credit to OneTrust. 'cause you know, that was kind of the, the learning for us is that, hey, you know what, we've always catered to security, but one thing I've learned in the intervening years is there are multiple stakeholders and data. There's multiple budget holders, right?

There's chief data officers, chief privacy officers chief security officers, and, and there's more today. There's compliance, there's data lifecycle management. There's even CTOs from a development [00:09:00] standpoint. And so. Once we kind of, realized that, okay, this is, these are the people that are gonna be at the forefront.

They're gonna be the vanguard. They're gonna care about this first 'cause they're tasked with this mission to conform with these regulations, that that's gonna be it. Now, there is a second part to, to that, in terms of a learning, we, we did, we only raised 2 million, which in today's market seems almost embarrassing.

Right. You, you hear people re raising seeds, 10, 15, 20. It's,

Grant: Yeah.

recording-1_2025-07-29_07-40-35-1: how things have changed in a, in a relatively short amount of time. But the other kind of realization for us is because we were approaching a problem from a security mindset, security landscape, we were missing 95% of what a privacy professional wanted to do.

Right. So a whole bunch of stuff as I like to say, there's 99 articles in GDPR, so kind of 99 problems as JC would say. And we really only had a solution for one, right? So we were missing a lot of the pieces [00:10:00] in terms of kind of, governance risk and compliance that was required. But instead of kind of saying, you know, woes us how are we gonna sell, we kind of doubled down and said, okay, you know what?

Let's make this one thing that we can do, which was all about understanding and identifying personal data, what data belonged to what person, what identity that was, the idea and solving for that problem, which today people refer to as data subject rights. So we only had money to really find one swim lane.

We couldn't take over the whole pool. We really had to kind of, key off of one small swim lane, but we made a bigger deal about that swim lane. And about what we did around it on the backend, and even partner with companies like OneTrust and other GOC vendors like ServiceNow. So we doubled down and then we, we, we went, you know, I didn't know any chief privacy officers, so I went to the sis who were all building practices around catering to chief privacy officers.

And I said, look, we don't do [00:11:00] 98 of the 99 problems, but we do this one and this one is a big deal, and if you want to address this one, we're the only ones that could do that. And that's how we basically went from like, Hey, what do we do? Swimming around trying to find the buyer. In a matter of six, seven months.

I, I think I, I did in three months, just again, the team was six people. I was one of two people in North America. We did I think like 1,000,003. so with that little bit of a pivot where we said, okay, how do we get to CPOs? You know, I didn't know any, anybody, any CPOs, so I gotta find a partner network.

And then rather than kind of shying away from the fact that we only did one thing, we really kind of, you know, really kind of tripled down and quadruple down on that thing and kind of made that, hey, this is an important thing that nobody, that you can't get from anybody else in the data security landscape or in the privacy landscape.

And that was what changed things. And it changed their ra relatively rapid, rapid mode. Like I said, we had a kind of a prototype and we did over 5 million in year one with no, no help from investors. And we did over [00:12:00] like almost 16 million in year two. Just with that, those two kind of little insights.

Grant: And so how did you, I mean, did you have much experience with like channel partners before? You mentioned, you know, the system integrators, assuming this is like, are you going to Accenture? Are you going to like more niche players? Like how did you approach that?

recording-1_2025-07-29_07-40-35-1: I did. So look, I, I did. I'm not, like, I'm not, I, I'm kind of an expert in nothing. But I, I have a little bit of experience in everything including jewelry making for my dad. So I know how to make a snake, rain a snake eating itself with the, the Romans would've been impressed with. So I knew enough and I basically, we started going to this privacy conference I mentioned, or privacy organization, IPP, and there's, there's sis or system integrators that attend those.

And it's really about finding one person. I don't need to find the whole organization and bring them with me, but if you find a partner and kind of animate them, not even a partner, usually it's somebody a little bit below a partner who's a bit more eager, who wants to make partnership. And that was always the trick.

So [00:13:00] whether it was pwc, ey, Accenture. Deloitte. Those are the first four big ones, which is not bad for, you know, six people. And, but again, I didn't boil the ocean as, as, as the frogs would say. I just I just focused on one or two individuals who I knew were trying to kind of, you know, doing the more vanguard things.

Were both technical and ambitious. You know, so they were still trying to drive revenue. They weren't a partner. They weren't kind of like satisfied yet. They weren't taking six week vacation. And I laser focused on one or two of those and every one of those organizations. And, and that's what worked.

And again, I basically just said, let's find one account, two accounts. That's it. We don't need to find 20. Let's just find one or two. But one or two multiplied across four is enough to get you to kind of, five

Grant: and that's one or two,

recording-1_2025-07-29_07-40-35-1: opportunities.

Grant: opportunities. Okay. So let, let's, let me kind of rewind how you're mapping bsi. So you're, I assume, I mean, there's kind of these [00:14:00] like. Specialization. So there's probably at Accenture or Deloitte, somebody that's like in the privacy practice or you know this, you know, they're, because GDPR has become a big deal, all these big companies are gonna start to turn to these consultancies, you know, system integrators to

recording-1_2025-07-29_07-40-35-1: It's new, right? Whenever it something is new, nobody has anything. And if you also put yourself, I learned, you know, I'm a quick study. Privacy professionals are not technologists, they're lawyers. So they have a budget, they have a mandate, but they could barely turn on their computer. So who do they, you know, who do you call?

Who are you gonna call? As they said, as the Ghostbusters set you're gonna call the system integrators because they're the experts right? Now, you may call others, but I knew, I, I knew generally how system integrators work. And as I mentioned, there was one industry association and they started putting on events.

So you go to an event and there's like, Hey, PWC has a booth. Deloitte has a booth EY has a booth. Look, that's how you, you start there, [00:15:00] you start kind of with the basics. Hey, hello, my name is Dmitri. I'm with the company called Big id, and we're focused on this problem. What are you doing in privacy?

And so it begins fairly, basically innocuously, right? This is maybe not something you learn in, in MBA school, but it is something you learn in, in sales school, right? This is how you start selling. Hello, here. This is who I am. Who are you, what do you do in this space? And that's how it basically began very, very simply.

With, with just kind of, but again, also not trying to pretend that we did everything. We, and to this day, big Id probably different than a lot of startups and we're still, you know, we're over a hundred million now, but we're, we're still a startup in many ways. We've always been careful about saying, this is what we do do, this is what we don't do.

So we always kind of circumscribe in, in the scope and I think sometimes startups say, oh, we do everything. And then you get into, in, into a kind of a production, it's like, oh, we do none of none of it. 

Grant: You're, you're a mile wide and an inch deep, right?

recording-1_2025-07-29_07-40-35-1: yeah, and, and again, for us, and maybe it's just a [00:16:00] personality, Hey, this is what we do do, and then let's focus on this and not, and like make that the important piece.

Grant: Yeah. I, I think, I mean, this is, I think for a lot of people, the, the channel is a bit of a mystery, right? So like, even just this process you use, going to the events and then not focusing that necessarily on the. The end customer buyers, but saying like, Hey, I want to get to know these sis. And then you're doing the same thing you would do in an account.

If you're trying to sell someone, you're, you're identifying a real champion, right? You're looking for someone who understands the problem. You said like, they're hungry, they're not maybe a partner yet. They're like, they're actually out there trying to make a difference and, and you're trying to make their career to some degree.

You're trying to say, Hey, you know, you wanna be really successful at privacy. We can be this great partner. And then they have opportunities and they're gonna, you know, you kind of, I'm assuming you're educating them on everything you can do about the market, how the put the whole thing [00:17:00] together. And so they are getting, they're becoming more of an expert because of you.

And then, you know, ideally they're bringing your solution into some number of their, you know, large, large accounts that they're working with.

recording-1_2025-07-29_07-40-35-1: Yeah, look, I think one thing that, that is a tell, so I'm not, I'm American, I'm a naturalized American. You know, don't call ice on me or anything. I don't know if that's a good or bad, but I became a citizen two years ago, but I grew up in Canada. And, and, and I, as I kind of hinted at, I lived in Europe and elsewhere previously.

One thing that differentiates the US from all other places is in most countries, these kind of buyers, CPOs, CDOs, CISOs, they wanna protect their job, right? They want security. They're averse to risk. America's the converse. There's always a striving class in every company, whether it's an end user or a GSI, there's some individual that wants to get ahead and they get ahead by.

Looking for the new, new thing for by taking a gamble, right? America is the land of riverboat gambling, right? [00:18:00] It never developed on the Rhine. And so there's always gonna be a group in the states that sees their future attached to something new, doing something innovative, doing something pioneering, right?

Again, America is a pioneering culture, and that is very different even from Canada. I'm not saying that there's nobody in Canada that likes to, likes to kind of do stuff, but in most places around the world, people wanna save their job, preserve their job by avoiding risk. The US I think one, one thing as an observation as somebody who's, who's kind of been here for a decade is the people here like to take a risk, right?

Because it's a way to kind of get further in their career, and so to some degree we play to that, right? You see the same thing happening in ai, right? I would recommend, there's always gonna be some people maybe that go, oh, you know what? I don't wanna take a chance. But one, one thing about about here in America is there's a lot of people that do wanna take a chance.

There's a lot of people that feel their future is tied to doing something [00:19:00] different. 'cause there's so many people here doing so many things. How do you set yourself apart? And so we played that same, same angle. And again, it's not like we have to go find a thousand. We literally found four or five people.

But that was enough because you don't need a lot at the beginning just to start that flywheel, right? You just need a few customers to get you over the hump. And then you start creating momentum and, and awareness. And so that was, that was enough. And that's something we did. It worked, it worked well.

And I recommend to anybody that's kind of in a new, new thing, doing something innovative and pioneering. You don't have to start with now. It's great if you raise a hundred million and you just start flooding the markets with marketing and just, you know, basically like hope for the best, the spaghetti strategy.

But if you don't have that, if you have a more modest kind of beginning. You are gonna be able to find people that want to get on your bandwagon. And it's really just a question of finding those people.

Grant: And then any insight in terms of like how you would structure one of those early deals. Is this, you know, what, you know, what [00:20:00] percentages, you know, how do you know for how long? Anything

recording-1_2025-07-29_07-40-35-1: Yeah. So, you know, look great question. So my very first customer was Nike. And I remember they wanted to pay me, I probably shouldn't say this, but, but hopefully they're not listening. 50,000. And I said, look, you know what? I'm not gonna be in business, so what's the point? So I said it has to be, I think, you know, much, much more like, I think eight times higher.

And I, I like, literally there was me, the world headquarters for the company was One Pirates Cove, ma, New York. That was my house. So kind of a weird global headquarters. But that was the world headquarters for big ID at the

Grant: Yeah.

recording-1_2025-07-29_07-40-35-1: And our next customer was Intel. And so like I, again, you had to have, and look, the only leverage I had was to some degree guilt.

I said, look, you know, I need something meaty because if I don't get something meaty, I'm not gonna be able to raise another round. And so you telling me you want to give me like 2000 or 3000 or whatever that number is, doesn't help me. I'm gonna have to go through a lot of work but I'm not gonna be [00:21:00] able to get to where I need to get to where you need me to get to.

And you know what, we were able to make it work. I think our first that's how, ' cause I, I was not able to raise maybe 'cause I was my world headquarters one Pirates Cove. I was not, again, today's very different. I was not able to get like people throwing money at me. It was quite the. And the feedback was, Hey, you need to be like, we need evidence that anybody cares about privacy and protection of PII.

Which today sounds crazy, but you know, back in 2017, that was the state of affairs. And so I had to go and in basically three months figure out how I'm gonna basically get over a million in revenue. And that's what I did. but again, part of it was like, okay, you wanna pay me this? I'm not gonna make it.

You gotta pay me this. And so there was a little bit of this kind of a give and take, but also transparency, right? I was honest. I said, look, you need this solution. I'm the only provider or purveyor of that solution. I'm not gonna be in business. If you want to basically onesie twosie these.

Now again, [00:22:00] I do see today we were in a different, more of a predicament because we had, we only raised a couple million, right? So that only goes so far. I think there are companies today that raise so much and frankly. It's really about just blanketing the market and that's a valid strategy. There's no criticism of that.

I think if you're fortunate enough to be in that situation that's a very legitimate situation. And then you can basically go out and do much broader marketing trying to find those people that are responsive to your marketing. We just didn't have that luxury, so we have to be a little bit more precise and pinpoint in, in how we go about things.

Grant: But I think, I mean, for anyone, I mean, the interesting thing is what you're really doing there is testing the value, right? You're saying, Hey, look, you, I, I can't accept less than yeah, 400 K for this contract because I'm not gonna have a business. And so they're saying, okay, well we thought we could get it for 50, but like, is it worth 400?

And so if they agree that it is, then you've, you've, you know, you've just tested the [00:23:00] value a bit more. I think that's a really powerful, you know, rubric.

recording-1_2025-07-29_07-40-35-1: It is. And remember, we've, we doubled down on our difference and highlighted like only game in town, right? So we knew, and this is the Greenfield brownfield Red, ocean, blue Ocean. We knew we were doing something that even the more established vendors and privacy like OneTrust were not doing. And so we found something that was technically complicated.

It actually used ai which was then still kind of novel. It used a type of ai, not LLM, but graph. And so we we knew that it was gonna be a requirement, but you're right, it provided validation that this is in fact the requirement.

Grant: And then from a novice outsider perspective, from what I heard you describe like in just sort of some of my knowledge of GDPR was like the, there was this idea that you needed to be able, still do, need to be able to complete like a data deletion request. Right. And so from what you described to me, I'm like, without your, [00:24:00] without big ID early on, knowing how to map all of that data to an individual person, I feel like that data deletion request would never be truly complete.

And so was that part of the, like

recording-1_2025-07-29_07-40-35-1: It was, look, I'll be, I'll be candid and truthful. Look, I think if, if every single company had basically embraced this requirement through automation or technology we would be a public company today. They have a lot of companies still do it manually, right? The one there hasn't really been any fine.

Two, you have 30 days, three lawyers. There's enough, there's enough ambiguity in the law that you can say, well, we're only gonna look in these two places. And so every company has this, so. We do have companies that are using it for today, but, but again, as lawyers, do they sometimes kind of, create a particular definition or particular kind of interpretation of the law?

And so it obviously benefit. So, so again, I think we found a way to [00:25:00] get to scale. I think there's still great opportunities. I think today there's 20 states that now have privacy regulations in the US there's still 30 to go federal possibility. There's not harmonization across the, the states. We're still the only technology now the companies claim it, but there's a, there's a kind of an asterisk.

So we're still the only company that truly has automation for it. But like I said, it was never like our end goal to do this and nothing but this. This was part of our broader journey around data security and compliance and being the broadest platform. The other thing that kind of the privacy endeavor did bring to us is.

Prior to big id, every data security technology was very focused on specific data sources like email, right? Like a proof point or a file system, like a semantic. Everything was very precisely. Now, part of that is because in the data center there's like five or six data sources you cared about, [00:26:00] but in privacy, because it's done by legislators who write the laws, nobody knows what a snowflake is, right?

They think a snowflake is something that falls in Finland, right? Nobody knows what a data brick is. The da they think a data brick is something that somebody in Croatia using to build a, a new house. So they don't know what any of the, like they don't know what an NFS or an SMB or a or a Mongo is.

A Mongo is something that sounds like a fruit but, but not, not, it's not spelt the right way. So maybe it's so one thing that privacy did, and you see this pervasive now in the data security landscape. It forced us to build something that was agnostic to the data source. Well, we were the first company that says, look, we need to build a solution for data security and compliance.

And today data and AI security compliance, that's agnostic of where the data resides. That may seem obvious today, but at that time, that was kind of, wow. That's, that's crazy. Because everything here to, for and you still see some legacies of that in the DLP space and certainly even in the Microsoft arena with purview, where you look under the hood and it only [00:27:00] works on SharePoint in, in OneDrive, you know, like it's still only very specific to the, to the data source.

And so one thing, and it forced big ID besides, you know, how we sell and, and trying to find these channels, these routes to market it also forced us from an architectural standpoint to make something that was ubiquitous, something that was cross platform, cross data source. And today that is like every new company in DSPM or some variation, that's what they promise Now.

Very few have achieved it, but that's what they promised. But that at the time was, was still fairly novel. Because again, that's not the way data security was done. You, you focused on one data source, like in the case of Varonis network attached storage and that, and you became good at that and nothing

Grant: Mm. Got it. So you went from sort of like the solutions were almost like verticalized around each data source and then you went more horizontal, but very, for a very specific kind of PII. So you kind of flipped

recording-1_2025-07-29_07-40-35-1: started with pi I, [00:28:00] but we very quickly, so once we were able to raise more money and look, we went through our we went through a period, we went through this kind of period where we couldn't raise money to say, despite ourselves. Then we went through a period that was the exact opposite. We there's a competition in the security world called, so there's a conference called RSA Conference, R-S-A-R-S-A-C, I'm sure you've attended in the past.

And on the first day they have a competition for most innovative company called the RSA Innovation Sandbox. We applied more as a lark. Just, you know, a couple of investors said, oh, you should, you should do it. So we, we did it and we applied and we actually made it to the 10 finalists. Now, timing is everything, right?

It's a, it's a little bit like in real estate, location, location, location. So in, in tech business, timing, timing, timing. And obviously by 2008, by 2018 when we were, we made it to the finalist, GDPR was like, not just on the horizon, it was a little bit like in the movie Terminator, you know, in the conclusion of the first episode where the clouds are coming [00:29:00] in, you know, like something menacing is happening, right?

So not only did we, we get finalists, we won it,

Grant: Oh wow.

recording-1_2025-07-29_07-40-35-1: and there was a glorious period of probably three years where, you know, we were raising every few months. So again, we went through a period where nobody really understood why do you, why do I care about personal data protection or privacy?

Despite all the obvious reasons we went. We flipped and we used those proceeds to basically very quickly expand. So we went from, hey, we could identify personal data really well to finding any crown jewel, any high value, high risk data, any Glen Garry, Glen Ross data, right? The stuff, the good data.

And we went from like, just looking at it from a privacy lens to a security lens to a, a data governance lens. And then we added a framework to say, look, visibility is fantastic. But visibility without actionability just equals liability, right? Because if you know your problem and you can't do anything about it, then you basically have a, have [00:30:00] a, a, a, a heart attack waiting to happen.

And so we, so we kind of use the, the kind of follow on proceeds to very quickly expand the kind of datas that we can support in terms of crown jewel using various kinds of AI and machine learning. We expanded the universe in the cloud and SaaS on-prem where we could look. And then we said, okay, you know what the GRAS is gonna be, not just telling you you have a problem, IE visibility, which is really where a lot of the vendors in the space today are.

That started in 2020. But saying, okay, well how am I gonna basically fix the problem? Right? I'm not just gonna be a plumber that comes and tells you, you have a leak. That's not why you called me. You called me to tell you that, that, you know, how do I, how would I get that leak fixed? Right? And every other industry, that's, that's the point.

And so we added a whole set of capabilities around 2122 to start fixing the problem. And that remains a, a strong differentiator for us. A big ID today.

Grant: [00:31:00] Okay. And did that whole time, did you maintain the channel as your primary route to market, or did you start to, you know, do combination direct and channel, or how did you think about that go to market

recording-1_2025-07-29_07-40-35-1: So, so yes and no. So you have to build so the channel is great but it, it's hard to scale a channel to the level that you need. So you need some control over your destiny. So you have to build a direct motion, right? So getting a Deloitte or EY to bring you a handful of opportunities is one thing.

Getting them to bring you a steady stream of opportunities is a very different thing. And partly it's 'cause they just don't see them. So we did a couple things. So on the channel side, we expanded from just focusing on GSI to focusing on VARs as well. So the reseller network. So I think in 2022 or maybe 23, we became kind of a channel only company.

Everything has to be routed. But now we had the reseller network as well like the Guidepoint and so forth. So we're a hundred percent channel. Every, every deal touches a partner. And we, besides the VARs and the GSIs, we also looked at the ISVs. So we took [00:32:00] money from Splunk, hp hp, sorry. SAP ServiceNow.

And, and with the, with the intent that they become kind of partners. And so that was another, and that includes the cloud scalers, right? Who now is their marketplaces are pretty important. So we're in all three marketplaces. And so that became a much bigger approach. So today, the channel's a fairly large organization for us.

I think approaching 20 people that covers ISVs. They all have three letters, so it's almost like a physics exercise where you're kind of, categorizing everything, ISVs, GSIs, FARs. And we recently just this year added MSPs managed service providers for downmarket. So that still remains important, but in parallel we said, look, we need control over our own destiny, so we need to be, even if we fulfill through a partner, we need to be able to drive demand ourselves.

And so we did build out a marketing function and a, and a direct sales function. Now initially the direct sales function focused on the enterprise, so larger accounts. And we're slowly kind of expanded that into. Big, smaller accounts, right? Smaller accounts are a little bit [00:33:00] easier, you know, as you start preparing for an IPO, big accounts are great.

They could be large, but they're lumpy and hard to predict in terms of close rates. So you need, you know, the, the, the ho pallo you need, you know, you need a large a large mark mid-market to balance that out, especially as you get to IPO and you want that predictability. And it's hard to predict, predict if you have like 20 large accounts in the quarter and you know, maybe four of them get delayed.

So you need, if you have 400 accounts maybe smaller A A SV or a CV it's, it's much better 'cause it's much easier to forecast

Grant: Okay, so, but you are still going a hundred percent so if a, if a deal is sourced.

recording-1_2025-07-29_07-40-35-1: through the channel. But sourcing is now partly done by us.

Grant: Okay. But you will, you'll bring in a partner. How do you decide what partner to bring in? It's just like kind of figure out what's the best fit.

recording-1_2025-07-29_07-40-35-1: Yeah, like, look, usually there's a partner already in the account. So we bring in partners. So I'll, I'll talk about that, the two ways we bring in partners. So one, from a VAR standpoint, it's a determination based in the [00:34:00] field. So the field team in the particular region, whether it's Europe, Latin America, wherever, has the best kind of relate, you know, knows who the players are, knows who they want to kind of work with.

And it's a little bit like priming the pump. So they'll figure out who they wanna bring it. On the GSI side, it depends on the type of work and who has a relationship there. It's hard to bring just some random GSI into a customer if they have no history with them. But the other place where we now partner with GSIs is, as I mentioned, we, we.

We started building this application framework for the controls, right, for the actionability. And so now that gives us an opportunity to say, Hey, you know, GSI, we have, we have a module around data retention that's a little bit more involved for a customer, like more programmatic stuff or labeling or access governance.

And so we try and now bring in a partner and say, Hey, this is a great thing for you Mr. Or Mrs. Partner to work with the customer. So that's still a work in progress, but the apps do give us an opportunity to start you know, basically [00:35:00] bringing in, like once we sell the app, to introduce a partner to kind of help fulfill or deliver that that functionality.

Grant: And then do you still have a Prosser team or how, or do you do anything in-house?

recording-1_2025-07-29_07-40-35-1: We do. We do, but we position it very narrowly. So the ProServe team, we don't consider, we don't call Prosser. It's really an onboarding team. We position it as a quick time to value team. So they basically just get customers up and running,

Grant: Yeah.

recording-1_2025-07-29_07-40-35-1: in 30 to 90 days, depending on the complexity of the customer.

And that's it. So think of it as a Best Buy Geek Squad. And, and it's mostly comprised, like, it's not comprised of like, you know, 40-year-old pwc or Deloitte expats. It's comprised of kids, right? So it's basically about getting everything wired and connected. Our product is simple enough, like you, you can do it on your own, but usually companies are in a rush.

And so what we've done is we kind of said, okay, this is the scope of what we do, but we're not a general purpose consulting arm. And, and [00:36:00] that's by design. We don't wanna start conflicting with what the GSIs do or even some of the VARs do from a services standpoint. So, you know, can it be framed as a services maybe, but we position it more as an onboarding.

Organization and it's really getting customers live in a very basically connected and live in a very short amount of time.

Grant: sure. Okay, that makes sense. And then you mentioned the cloud marketplaces. Tell me about sort of, you know, what's the, what type of integration have you done there? What's the value you're getting? You know, how, how do you, how do you see those cloud marketplaces?

recording-1_2025-07-29_07-40-35-1: Yeah, so look, it's hard to get through. We get occasionally referrals from the big cloud scalers, but generally speaking, the marketplaces are just a transactional a place to transaction. And the deal for a lot of customers is they already have prepaid or pre-comm commits to Amazon, Azure, GCP, and this allows 'em to burn it down.

Where the marketplaces get tricky, and it's still more art than science is if [00:37:00] you want to do a three legged transaction where you don't wanna exclude the, the var. You wanna include you wanna let the customer have the flexibility of being able to burn down, their, their the dollars they've already committed to a particular marketplace.

Now the good thing is there's a whole bunch of VARs that already have these kind of bilateral relationships with the cloud scalers as well. So there is a way of doing it. It's still a little bit more complicated 'cause it varies from partner to partner. But generally speaking, the marketplaces are for the benefit of the, of the customer because it gives them essentially an ability to burn down commits, which, so basically it's dollars that they're gonna have to spend one way or the other.

And now they could kind of spend it on us. So it gives them a, we don't always use it, not every customer wants to use it, but it's, it's a powerful way to pay for a particular program.

Grant: My parents preloaded this, this debit card and, and I, I need to spend on that.[00:38:00] 

recording-1_2025-07-29_07-40-35-1: Basically, yeah. Like every, every college student knows knows what that means, right? Every month if I spend it, you know, like, look, I know it's gonna get reloaded at the end of the month, so I might as well go to the bar one more time.

Grant: And then do, I had one thesis around this, which is like, do you feel like when you are included in the cloud marketplace Bill, that like, it's sort of like, that's just this big bile of spend and like they're, they're a little bit less, they don't scrutinize the, you know, the line items within the, they're just like, oh yeah, we spent 5 million IWS and even if like 2 million of that was on ISVs, they're not, they're not like the CFO's not digging in as deeply there, you know, to like each of the invoices that's coming through.

Is that

recording-1_2025-07-29_07-40-35-1: I haven't. Look, I think that's a great question. I think it is probably a little bit easier, especially in you get these occasional kind of blips in the market when interest rates spiked in 22. When COVID came on more recently when the, the tariff tantrum, I think in April where you had, [00:39:00] wait a second, everyone's like, oh my God, you know, what's going on?

Is this gonna be a recession? And you have this kind of like, almost perturbation, right? In physics, we would call it a perturbation. And it's just kind of like, wait a second. The bottom falls out in the market. You

Grant: Look at all of our spend. Yeah,

recording-1_2025-07-29_07-40-35-1: Yeah. So there's a, everything gets more scrutiny in those instances.

And so I do think that there's a way to kind of. Camouflage some of that. Look, I think it's generally a great thing. Honestly, I'm not sure why we, we don't always do it. But for customers, you know, it's hard to find a customer that doesn't have some, some already committed dollars to the three clouds.

maybe there's, maybe there's certain restrictions around what different departments could do in terms of, 'cause you could also see a situation where every department go, great, I'll throw in my lunch into the, you know, the cloud. You know, everything will go into the cloud. Commit, right? Maybe not a lunch, but every, every piece of software.

So I'll just buy every piece of software under the sun. But look, it's, it's a great convenience for, for end [00:40:00] users and it's a great conveyance for, for vendors.

Grant: Yeah, I mean the one argument I would make against it is like, is basically the Amazon basics, like, you know, sort of comparison, which is, you know, because everyone's selling through Amazon, Amazon has visibility to where like all of the spend is and what products are really getting purchased a lot, and they have like all the analytics.

And so I think it just sort of maybe gives them a little more data around where they need to roll out, like the, you know, AWS competitor to something if it, if it really takes off. But that's only because they've been so, you know, Amazon particularly has been so, I don't know, like just to have done that pretty pre, in a pretty predatory way on, on the Amazon mark, you know, like.com side.

recording-1_2025-07-29_07-40-35-1: there's, look, with all of these things, there's discounts, there's gross margins. So look, as we become a bigger company, you know, when I was starting, I didn't think about gross margin a lot. I do think about gross margin a lot now, [00:41:00] right? I think about all the KPIs around sales efficiency, like constantly.

Probably my board reminds me to think about it. So, so look, all of these things have bad effects, right? So if you could imagine you pay X percent to Amazon, you pay another large percent to a var, maybe a GSI comes in and, and takes you know, also has a kind of a fines fee or something. So, you know, you could go bankrupt if you basically, you know, just sprinkle, sprinkle it around everywhere.

So I do think you have to be a little bit more thoughtful and this is where the, you get some complexity, as you know, later, later in life around routes to market, how you fulfill, how you sell. 'cause there's a desire I'm gonna sell through everybody. But you also have some ISVs that want to take a percentage as well.

So if you wanna be able to tap their Salesforce, they wanna also take a percentage. So again, you know, we're still navigating some of that. And again, we have a alliances team, and then we have a, a sales team. That helps us navigate, we actually combine them. So alliances now fits under sales, which it didn't before.

So now sales is more [00:42:00] just a customer team pre and post. But again, you know, like everybody else, we're kind of, making our way through some of these decisions.

Grant: Yeah. That's interesting. Okay. And then talk a little bit more about how AI is changing your business, both from how you work, what the market opportunity is you know, anything that you think is relevant because it is such a impactful, you know, shift in, in the business. So,

recording-1_2025-07-29_07-40-35-1: There is. So look, I think it's, it is impacting a couple of ways, right? So look, we've, we've been using, I think I mentioned we used AI from inception. The very first technology we built was now it used graph, not LLM, but AI predates large language models, large language models, our particular flavor of gen, ai, and transformer model and performance, a very specific task.

But we were kind of using AI in inception. So look, I think I'm gonna talk about a few impacts. I'm gonna try and enumerate them. So first I'm gonna talk about just just the kind of operation of the business. [00:43:00] Then there's the kind of making the technology easier. So this is where you do the MCP and Agentic and and all the kind of co-pilots in your product.

And then I'm gonna talk about. Products we build for ai, right? What are the use cases we wanna support that other people have as, as part of their AI prep? And the last thing I'll just talk about is kind of market perception. And I think frankly, the immaturity of where the market is. And I think where a lot of people, like, I don't think people have a clue, certainly outside of the data and AI space, I think if you talk about security professionals, oh yeah.

LLM will solve all my problems. LLM is not a one trick pony, right? It's not a hammer to every nail. It's it solves very specific problems. NLP solves some NER solves some, you know, customer not, you know, go, go down the list. But I think there's this kind of misperception about, you know, what can and is and isn't possible.

And some of the, so anyways, lemme kind of start with the way we kind of view ai. So one is we are looking at using AI [00:44:00] across the company in the tooling. So we have started replacing some of the, the tools that. We we, we used for sales, marketing hr with more AI forward tools because a, they come with a, typically a lower headline cost.

And there's a little bit more automation. So that's why, and we actually now have an internal it, a r that's kind of mandating that saying, okay, what else is possible? The other thing that we're doing is we are looking for opportunities at savings and efficiency, right? So we are trying to mandate across the various departments we haven't formalized it, but it will get formalized and crystallized over the coming months that look you as a, as a, as a department, need to be more efficient.

I wanna see what you've done either through software or through other types of efficiencies to leverage that. So that's one thing we are doing, and you see that with the big companies, and now it's kind of, percolating. Now in our product we've also kind of embraced all flavors of ai. So the way [00:45:00] we identify data as part of our general privacy security, we started with graph, I think, which is the type of ai we now have methods that involve N-E-R-L-L-M now it's not a one size fits all.

Now we don't tell the customer pick which AI you want. It depends on what they wanna do. You wanna do exact value matching. There's an AI for that. You wanna do look for combinations of attributes. There's an AI for that. If you wanna look for how data is connected to an individual, there's an AI for that if you want to be able to infer data like inference.

So I wanna look, I wanna be able to figure out who the data owner is. There's an AI for that. I wanna find all the derivative data. So if here's the parent, here's the children. There's an AI for that. So we have, I think, more in our space patents than anybody else around the application of ai. So identification of data is perfused even like LLM, right?

We'll be announcing a black hat in in less than a week the first prompt based LLM classification. So what does that [00:46:00] mean? While a lot of vendors saying we do LLM, we do LLM mysteriously, none of them offer a prompt. And if you've ever used open AI or any copilot, you know that, well, that's kind of weird because, you know, most ais are not black boxes.

They have a way for you to inter interface with them to converse with them. And so, and the, and the value of the prompt, which also begs the question as to what they're doing in the background. But the value of that is that allows you to basically give the power to a business user, not technical user, to go find whatever they want.

Even the ability to go say, here's a new law. I'm gonna copy paste into the prompt. Go tell me what, what's in my data? So that's an example. So we have a flavor of that. In addition, we've offered, we, what we were calling copilot, but now we're, we're now calling ag agentic assistance. A set of that to basically help you do various tasks inside of the product reporting be able to do certain types of inferences.

Like I wanna figure out who the data owner is. I [00:47:00] wanna know, here's the, the raw data, but I wanna know what the right business description is that for the data based on my taxonomies. So we now have a very comprehensive copilot or agentic assistant that basically provides you a way of working, in the, in the product even the way we do search.

So we've always provided core search so you could find findings that is now being, that now is natural language and vectorized. So, so again, all these things. So AI is completely all across the product from MCP to agent Agentic, assistance to kind of vectorization for search, all that kind of stuff, to various types of AI from LLM to NER for identifying data.

Thirdly, well, how do we help companies with ai, right? So we've introduced a whole bunch of things, right? So in compliance, we now do, risk assessments for ai, for vendor ai. I think we're the first company to introduce vendor ai. So I bring in a [00:48:00] vendor, they say, oh yeah, I have an amazing ai What's the implication to your company?

'cause remember, most ai, especially LLM, is actually pass through very like no company that you and I know, you know, say for Facebook and, and Google. No one's building a frontier model. It costs like $4 billion. Like nobody's building that. So all you're doing is you're either doing some type of rag, which means you're training it with some data.

From your customers or somewhere, or you're just doing complete pass through, you're just basically sending it, rerouting it to open ai. Well, what are the consequences to a company? So anyway, so risk assessments secure data pipeline. How am I? So I wanna do rag which for your audience is basically kinda like a type of data enrichment.

Well, I need to get data ready, but the problem is I can't just give it all my data because that data basically gets memorized by that ai. And so I want to be able to a, find the right kind of data, like maybe mortgages. I wanna find all the mortgages, but also make sure that I'm not including social security numbers, addresses, all this other kind of thing.

[00:49:00] So preparing that data is something we do. We call it secure data pipeline for ai. It's a use case. There's a whole bunch of things from redaction to curation to compliance that we do. Shadow ai, you familiar with it? That's another use case where customers are going, you know what? It's so easy to either download like a deep.

Or just go and say, I'm gonna use perplexity and I'm gonna basically pay them subscription and I'm gonna get through my firewall. So there's a set of use cases, employee access to ai ai trust, risk and security management shadow AI that we're also tackling. And remember when I said earlier on, we're not claiming to do all things ai.

We are identifying finite things where we can make a difference in ai. and that's kind of just a philosophical thing and saying, look, we're gonna focus. We're gonna do these. And so now big ID, and this is different from a year ago, could basically tackle these five very concrete use cases. So again, those are three things.

So we're doing operationally across the [00:50:00] business we're doing in the product. Copilots, AgTech, assistance everywhere, vectorization. And then thirdly, identifying areas that are important to our customers in terms of their AI journeys that we could assist them with. The last thing I'll just kind of mention for your audience in terms of AI is I still think that there is a general, oh yeah, I want AI everywhere, but not realizing, and I've been on a lot of calls with CISOs and CDOs and CPOs not fully appreciating or comprehending like, well what's what's involved in that?

Right? Are you gonna have to basically give the vendor full, full copy edit permission? Are you gonna be sending all of, you know, these things that you've been spent spending a fortune on DLP and other technologies to protect and now you're basically sending to a, to a third party forever to be memorized?

So I still think that there is a general poverty of understanding of how LLMs work, how some of the enrichment methods like rag and fine tuning work around [00:51:00] some of the things to ask vendors and users. I think everybody wants it, but not, and also what is the art of the possible, right? LLM is really kind of a particular type of transformer model.

It's a probabilistic model for looking at these kind of war tokens and kind of sequencing them to make them sound like, like human human speech or human writing. it's not like a, a one one thing si one size fits all. So I think understanding what the nuances of a deep learning model versus a particular type of LLM or SLM or NER and understanding, I think that understanding in among buyers is still very poor.

and I think it's also candidly very poor among vendors because vendors do a lot of jazz hands. If you're familiar. Maybe it's a little bit of an like, oh yeah, LLM, you know, like, yeah, we, we do it all. It's all LLM, everything's LLM. And of course, if that's true, then how is that differentiated?

'cause everybody basically can use the same open source models or commercial models. So anyway, so I think the fourth thing I would say is there's [00:52:00] still a little, a lot of confusion in the market. About ai. And the only thing that will kind of solve for that is just time. I think there's no other, there's no other solution.

Grant: Yeah. I mean, assuming, you know, some amount of education, right. Is the other part of that you're, you're probably in the market webinars, talking to your buyers, talking to folks, helping them see this and educating 'em. Is that, is that right?

recording-1_2025-07-29_07-40-35-1: We are in the market. So look, I'm speaking on it. I speak at conferences and so forth. We do lots of webinars. We obviously talk to lots of customers individually, one-on-one. But like I said, I think, look, it's easy for us as a business for, as a, you know, putting on my kind of operations hat to kind of say, okay, we, we wanna, we wanna create mandates or, or fixes to be more efficient, using ai.

That's one thing. In the product, we basically empowered all the various teams, product teams to say, how could I improve what I do in the product with ai? Maybe it's around how I prioritize risks, or maybe it's around how I search for data, or maybe it's how I, I get [00:53:00] assistance in terms of performing certain hard tasks.

So we've done that, and, and look, that's a, that's a journey. It's not a destination, right? And then wordly kind of helped them, but I think the education, the evangelism, it's not just us. It's like everybody, right? I think to some degree ai, we forget that three five from open AI came out in 2022 in the second half of the year.

So we're still in the very, very early innings of ai. And, you know, look, I, I would probably wager a lot of people are an expert on the internet, let alone ai and, and the internet's been with us now for. 30 some odd years. So I think it'll take a little bit of time and and there's no, you know, there's no way I think to forestall that or to shorten that.

Grant: Yeah, I mean, it, it does, I mean, one interesting insight, at least from a lot of the companies that seem to be scaling because of ai, I think there's like a handful, right? There's obviously like the cursors and these cogen, you know, agent coding tools. but also seems like there's also a lot of like the service [00:54:00] providers to the AI companies because these, you know, philanthropic and open AI are growing so fast.

Anything that becomes like a core part of their business also seems to be really growing quickly. 

recording-1_2025-07-29_07-40-35-1: Look, if you're fortunate enough, look who's that company that basically provides compute to two vendors? Microsoft. They just went public. But but yeah, look, I think it's great. And look, it also burnishes your image, right? Weights and balances, which got acquired by the same company.

Grant: Core weave. Yeah.

recording-1_2025-07-29_07-40-35-1: Core, weve, yeah, so Core Weve is a great example. They have like two customers. I think. One is OpenAI, one is Microsoft. Like that's basically their entire customer base. If you look at weights and balances, which got acquired by Core, weve. They, all the big vendors used them. But again, they weren't, they didn't have like great revenue because there's only so many companies that are building foundation models.

But you look at scale now, scale had great revenues, right? They built a fantastic model, a lot of its services, but they did have tremendous amount of revenue. But now that you know, the CEO and and some of the leadership team moved 

Grant: meta. Yeah.

recording-1_2025-07-29_07-40-35-1: [00:55:00] to Meta. Yeah. Is it meta or did he go to, I can't, can't keep track of who's hiring who, but did he go to Meta?

Grant: I believe so. Yes.

recording-1_2025-07-29_07-40-35-1: Yeah. So now that he's there, I guess he's running the Super intelligence whatever, or he's one of the 15 people running it. I think they've had Google and some of the other people that were using it. So look, I think it's still, I think it's great if you could get in one, in one of the foundations.

Certainly grow with them and certainly it's been fantastic for Nvidia. You know, as a shareholder, it's, it's fantastic news. and it's been fantastic for the people to build custom custom chips on Broadcom. So I still think you see huge benefits from the build out and it's frankly keeping the stock market going.

But I do think that there's the kind of the, like everybody else, right? Yeah. There's not a vendor today that isn't the way I described it, right? Thinking about AI proficiency, thinking about AI as a way to improve their product thinking about addressing certain AI use cases every single vendor is doing that.

So that means that you have 5,000 enterprise vendors that are now thinking about how do we do all three things [00:56:00] with ai? And so it'll be interesting to see how quickly they achieve the efficiencies the improvements to their own product but then also being able to quantify. What use cases are they gonna be able to speak to in that ai problem space, right?

Like, how am I gonna prepare data? How am I gonna label data? How am I going to ETL data, whatever that is. And so that'll be interesting as well.

Grant: Yeah. And I mean, I guess one of the, this is like a philosophical question. Do you think, do you think AI is gonna allow you to do more and expand your offering and make this sort of like a broader platform? Or do you think it's more about efficiency? Like how do you, how do you see

recording-1_2025-07-29_07-40-35-1: Yeah. So, so in terms of our product, so one is, it's gonna be look, I think from an efficiency automation standpoint. It's remarkable, right? So whether it's just from a big, big ID operations like sales, operations, marketing, operations, HR, operations, you know, like obviously [00:57:00] right? You're basically replacing certain manual tasks with some type of automation inside of the product.

I also like, look, it's gonna enable not only more automation, right? So for instance, we offer certain capabilities that today are done manually in the data governance space that require people to physically say, ah, this data is mapped to that business term. And that's not a scalable approach. So we do that automatically using ai.

So there's some of that, but it also allows, it also gives a couple of new interesting things that so people are talking about MCP server, we've introduced a couple and, and server side or, or product side agents as well. It first of all also downplays the necessity and importance of your user experience, right?

Historically, user experience is like everything, right? How easy is it to use the product? How clear cut. And it becomes harder as your product gets more, more complex, there's more things, and you don't want to confuse people. AI creates an [00:58:00] equalization, right? Because now you don't have to really worry about the user interface because now you're providing people a different way of interacting or interfacing with a product through this kind of AI assistant, whether it's a prompt or something more audible or something more visual.

So that's, I think that's really interesting from all vendors, right? So essentially you're almost creating this notion of a headless, product. Now, I don't think we're gonna see. UX disappear as an important principle, but it is a very new, if you think about kind of retail omnichannel, it's a very important and novel channel in terms of how people interact with your product.

And so we're spending a lot of time on that, right? In terms of building service side and MCP, 'cause we believe more and more customers really want to get to the destination. They don't wanna be clicking a lot of things to get there, right? I wanna be able to create a particular type of report. Why do I need to go, okay, I need to export the data to Snowflake and I need to be able to integrate it with Tableau and I need to be able to, or whatever that is, or I need to be [00:59:00] able to cleanse it or normalize it.

Wanna just say, I need the following data and I wanna make sure the data is accurate and everything just happens for you. That's what people want, right? They want to get to the destination. They don't really want to be able to figure out all the pieces of how I basically get through the journey. Right?

Turn, left turn. They don't want that. Which, you know, same thing as like, you know, driving, right? We want, we all want autonomous cars. I don't wanna like even look at the roadmap. The other thing that's gonna allow us to do is also this notion of Federation of a actions. So right now, usually we bring people back to our product, right?

Even if there's like, you know, an email you get or something, they come back to the product. Well all of a sudden, because there's gonna be agents kind of embedded inside of other products, you can say, well, you know what? I'm gonna operate, I'm gonna interface or interact with big ID through Slack. I don't need to come back to big id.

I'm gonna say I'm looking for this and I'm just gonna get an answer. And that's also a [01:00:00] really interesting possibility. Now, my last company is in the API space. I think APIs kind of promised, promised that reality, but they never really got there. That mechanism was still fairly brittle, breaks easily, and so you never got that true Federation.

APIs basically were just a way to kind of sync data. but here, because it's a little bit asynchronous, right? You tell your local agent I need the following, and then it goes and finds an agent with the right tooling and the right skills to go do the things it needs to do. So you're less concerned about and then talks to all the MCP servers to find out what, what they kind of expose.

So I do think that's another really interesting possibility in terms of changing the way we work. That a we'll be able to get to the destination much faster, but we'll also be able to work wherever we wanna work, right? Whether it's in email or whether it's in Slack or whether it's in Salesforce or, and, and there may not be one place or may be multiple places to be able to get to the same result.

So this idea of like historical life to jump to big idea, to jump to [01:01:00] Salesforce and jump to ServiceNow, to jump to, I think some of that is also over the next two years is probably gonna get is slowly gonna dissolve.

Grant: So, so instead of keeping all the context in your head, basically you have an AI agent that's like kind of jumping between things, but maintaining the context and like, yeah.

recording-1_2025-07-29_07-40-35-1: So I think there's gonna be two elements. So again, we're, we're right now building all this stuff in our product, but now we're starting to say, okay, you know what we want. The reality is sometimes people just want to say, okay, where do I want to get to? So, for instance, I want to prepare a data set for, for doing rag for training an AI model.

I wanna find everything that looks like a purchase order. Between the following dates, I wanna make sure that it's every, every file is different, not the same. I wanna make sure there's no personal information. Why do I have to go search for it in, like, you know, that's it. I, I already gave you the instructions.

Just come back when, when you're ready and bring me back the dataset. So big ID is doing that right now. Like, I think we'll actually have some of that by the end of this year. So we [01:02:00] already have all the tooling to do that, so we already have the mechanism to do that in the product, but now we're building the agents to be able to go orchestrate that for you.

The next step right after that is great. I mean, big id, I can basically get to the destination I want to get to by just telling it. The next step is, well, what if I don't actually want to be even in big I, what if I wanna basically be in ServiceNow? 'cause or what if I wanna be in Slack? Or what if I wanna be in GitHub, and just get whatever, kinda like the old adage if I wanna be, if I wanna look at data scientists would be in ju.

Notebooks, right? I just wanna, like, I have a place where I wanna work and maybe I have multiple places. And so that idea of being able to federate the functionality, again, APIs kind of promised that, but didn't quite get there because they were kind of very rigid. The thing that makes this kind of AI gentech model a little bit better is that you could, you only really are working with your local agent.

The agent is almost like the, again, some of your audience will remember this, but like a travel broker, it then goes out and figures out exactly. So there was a time we [01:03:00] didn't just go to booking or Expedia. You'd go to your travel agent, but travel agent didn't know everything, right? They didn't know where they were just smart enough.

They had a phone number to call American Airlines or United they could call Royal Caribbean. They get, they get ads from tour groups, so they knew where to go. Get all the information and compile it. And so to some degree that I think is gonna be ubiquitous probably by mid, mid, mid, or end of next year as well.

And certainly we're, we're embracing that as well.

Grant: That's amazing. Yeah, I, I think, yeah, it's a really, it's a really interesting perspective and I'm guessing that you're, you're doing the same thing that I am, which is just like pushing your teams to use all these things as much as possible in order to get those insights. So you said like you're doing it in your product, now you're building products and all people do it, and it's like the only way you get there is by, you know, having teams that feel like they're AI first, and this is just part of how they think about work.

And they approach things with like. [01:04:00] Instead of, here's how I used to do it. It's like, well, how can I do it differently now that I have AI around? Is that, does that feel right?

recording-1_2025-07-29_07-40-35-1: Yeah, I think, look, I think look, you wanna be able to take the technology. To its realm. And look, I, I think, so first of all, you know, I, I, I talked a little bit about and I'm just gonna talk very selfishly, and maybe I'm oversharing as I sometimes do, because I'm Canadian. That's what Canadians do.

We're so honest people. Look, I think just like I talked about what makes Americans, I don't dunno why I'm talking about the sharing is the Canadian, but, but now I'm thinking like an American, ambitious American. I wanna get the full flavor of ai, right? And I want that in my company. So I know that AI can help us operationally.

So I wanna make sure that we, we embrace that because I, I wanna learn from that, right? And I'm always on the look for tools that our teams could use. secondly, I wanna improve the the efficacy of people using our tool, right? So I know, so we've kind of made a clarion call to all the team leads.

Hey. Look for opportunities for you to take advantage of ai, whether it's [01:05:00] vectorization and search, whether it's, a copilot feature on documentation, whether it's using agents to be able to automate data stewardship, look for opportunities to take, take advantage of ai. Thirdly, we've as an executive team, a leadership team on the product engineering.

And obviously I, I kind of, stick my nose into it as well. As any, any founder in technology does,

Grant: Sure.

recording-1_2025-07-29_07-40-35-1: where are we gonna play in ai Now we take a, a little bit more of a deliberate approach. Like I said, I think a lot of my brethren, including some of my competitors, just say, oh yeah, we solve everything.

I go, to me that's like, you know, I roll my eyes and we're like, what if you solve everything, you solve nothing. And you know, you're basically like peanut butter. A very thin smear of peanut butter. So we've kind of identified, okay, we're gonna focus on, on these, set of use cases and we're gonna go deep.

And, and then we're gonna expand the set of use cases. And then the last thing, like I said, there's an evangelical element and the only thing you could do is do podcast speak publicly. I think to some degree you're not gonna bring, you know, the world [01:06:00] is gonna just mature at the, at the pace the world is.

But I do think that right now we're, we're at the phase where the general technology buying audience thinks that AI is, is literally cod liberal oil or some kind of like snake oil. Kind of like it solves all things, but I think there, there is some nuance, right? It, it's an incredible powerful forcing factor and it's gonna change the landscape and industry and everything else.

But it's also important to understand. Where it stops, where it starts and, and understand the various flavors of ai. I think more importantly, it's not so much that AI can't do certain things, but AI is not a generic panacea, especially LLM. It could do certain things, well, it could do certain things poorly.

Deep learning. Same thing. Cluster analysis, same thing. Graphs, same thing. NER, same thing. NLP, same thing. And so trying to understand what it all means is, is important I think, for the general audience. But, you know, there's only so much I could I could do there.

Grant: No, that's super helpful, Dmitri. I mean, like, I think you've shared a lot of really [01:07:00] amazing insights here, so, thank you so much. Anywhere where, where can people find you? Anything, anything else to add?

recording-1_2025-07-29_07-40-35-1: if you go to my Miami gimme a holler assuming I'm not traveling. But but generally speaking, you can find me at DDA at big id BIG id.com. Shoot me a note. Happy to chat and then be on that. Our website as you probably figured out, is BIG id.com.

Grant: Amazing. Dimitri. Thank you so much.