Latest / Cyber Go-To-Market Talk: The Cybersecurity Podcast About Cyber Sales, Marketing & Revenue Growth / DataTribe Challenge Finalist: Starseer

Transcript

DataTribe Challenge Finalist: Starseer · Cyber Go-To-Market Talk: The Cybersecurity Podcast About Cyber Sales, Marketing & Revenue Growth

• 0:41
  SPEAKER_01: Welcome to the Cybersecurity Go to Market podcast for a special showcase episode where we are talking to leaders of the company selected for the 2025 Data Tribe Challenge, where Data Tribe, which is the Cybersecurity Foundry, have identified five startups as being the most innovative in cybersecurity today.
• 1:03
  SPEAKER_01: I am your host, Andrew Bonahan, and today we're talking with Tim Schultz, who is the CEO and co-founder at Star Seer.
• 1:11
  SPEAKER_01: Tim, welcome to the podcast.
• 1:13
  SPEAKER_00: Thanks for having me, Andrew.
• 1:14
  SPEAKER_01: Looking forward to our conversation, Tim.
• 1:15
  SPEAKER_01: These are always interesting.
• 1:16
  SPEAKER_01: You know, you've got early stage companies, you're you're getting going, you're fighting your feet, you've made some progress, and it's fascinating to learn what you're seeing and what you're trying to do along the way.
• 1:25
  SPEAKER_01: Let's start off right at the beginning.
• 1:27
  SPEAKER_01: So, what's the story behind you as a person first getting into cybersecurity back in the day?
• 1:33
  SPEAKER_00: Yeah, like so we'd have to go a little while back, but I think like I was always interested in computers.
• 1:38
  SPEAKER_00: And so that was that was sort of, I'd say, you know, where I grew up.
• 1:43
  SPEAKER_00: I ended up there was a digital forensics lab, one of the first in the country that was actually in the same small town, Starkville, Mississippi, which is where Mississippi State is.
• 1:52
  SPEAKER_00: And so they had work they'd done with the Department of Justice on sort of digital evidence.
• 1:58
  SPEAKER_00: And essentially a lot of it was teaching law enforcement how to make sure that you are maintaining proper chain of custody for digital evidence.
• 2:07
  SPEAKER_00: So I, you know, I started like anyone that had a lot of you know first jobs, right?
• 2:12
  SPEAKER_00: It's it's not all hacking or or anything like that.
• 2:15
  SPEAKER_00: This was around, you know, helping put binders together for the training and some of those things.
• 2:21
  SPEAKER_00: But, you know, I was a curious person.
• 2:23
  SPEAKER_00: I ended up, you know, doing a lot of work there, helping build some of the material over the years.
• 2:27
  SPEAKER_00: And so, you know, another thing that I sort of was curious about is I was like, well, man, there is a lot of like mistakes that these, you know, bad guys are making, thankfully.
• 2:37
  SPEAKER_00: But I wanted to understand a little bit more about the, I'll say, offensive side of things.
• 2:42
  SPEAKER_00: So that was where I ended up going and getting my master's degree.
• 2:46
  SPEAKER_00: And it was much more focused on sort of those like how do attacks work, you know, how do you build them, that kind of thing.
• 2:53
  SPEAKER_00: And so that was sort of what led me down my career in cybersecurity as a whole.
• 2:58
  SPEAKER_00: I ended up doing some stints at Sandy National Labs, Mitre Corporation, for anybody that's familiar with something like MITR attack, and you see that on the news.
• 3:08
  SPEAKER_00: And so that's where it sort of led me down that path.
• 3:11
  SPEAKER_00: I then ended up working at my first startup, which was one called Sife, did breach and attack simulation, which is around essentially simulating and emulating attackers so that you could build better defenses in an automated way.
• 3:25
  SPEAKER_00: And from there, I kind of, you know, got the AI bug like a lot of folks, and ended up going and uh sort of founding the AI Red Team at Verizon and sort of building up that.
• 3:36
  SPEAKER_00: And then after a little while, and we'll get more into this, I decided, you know, I think this is uh a good opportunity for me to jump off and start my own company.
• 3:45
  SPEAKER_00: So and managed to convince at least my found my co-founder to do it with me.
• 3:49
  SPEAKER_01: How do you find it swap between uh offensive and defensive?
• 3:53
  SPEAKER_00: It's a really interesting thing because you really I think it's really important to have perspectives on both because that's something that I see sometimes where either teams or individuals, like people will really want to specialize, right?
• 4:08
  SPEAKER_00: But I think that at the end of the day, especially when you're talking about red teams and things like that, to best like to help businesses be more secure, you really need to be working together.
• 4:19
  SPEAKER_00: And that was where I ended up working a lot with like purple teaming, which is right, getting everybody in the room because a lot of this, it's it's not an expertise problem necessarily.
• 4:28
  SPEAKER_00: It's all right, we need everybody to actually work together, work as one team.
• 4:32
  SPEAKER_00: And I think that's something that you know we we saw a lot of success with, but it was it's something that in hindsight sounds really simple, but there was a lot of like adversarial relationships that that made that more difficult than you might think.
• 4:45
  SPEAKER_00: Again, it has nothing to do with being a technology problem, but that was that was an interesting challenge.
• 4:50
  SPEAKER_00: I was gonna ask, is there is there egos involved on both sides?
• 4:54
  SPEAKER_00: Oh, always, always.
• 4:55
  SPEAKER_00: I mean, you have people like I think one of the things is right, is is you have like on the offensive side and a lot of that, a lot of people are looking for exceptions, right?
• 5:04
  SPEAKER_00: They're looking for how how can I find that one piece or the that one thing that allows me to essentially gain an advantage.
• 5:11
  SPEAKER_00: And on the defensive side, though, one of the challenges, right, is you have to defend this sprawling thing, like whether it's the business or building a product for it, it has to support a lot more use cases, right?
• 5:23
  SPEAKER_00: A lot of things.
• 5:24
  SPEAKER_00: It regardless of what companies use for email, right?
• 5:27
  SPEAKER_00: You have to, if you're gonna try and secure email, you've got to tackle a lot of different types of email clients and and things like that.
• 5:35
  SPEAKER_00: And so that's where when you move to the offensive side, it tends to be very focused niche things.
• 5:42
  SPEAKER_00: And I mean, it they go way really, really deep.
• 5:44
  SPEAKER_00: And so there's like a really important set of expertise in being able to dive super, super deep.
• 5:50
  SPEAKER_00: But that is where there's sometimes some egos and some people that you know you have sort of this deep versus wide approach.
• 5:57
  SPEAKER_00: And I think that's that's one of the challenges of sort of getting everybody to talk in the same language.
• 6:03
  SPEAKER_01: Yeah, that's interesting because you know, I think one learning from me already is you hear people talking about well, you know, the the bad guys only need to be right once, and you know, the good guys, good, all this, you know, tech surface and all the rest of it.
• 6:15
  SPEAKER_01: Actually, I think what you're saying is it's much more about focus first is breath, right?
• 6:19
  SPEAKER_01: Is yeah, you you have 85 plate spinning if you're on the defensive side, but you know, on the offensive side, you've really got to get really good at taking down one plate and not worrying about all the sorry 75 other ones, right?
• 6:32
  SPEAKER_00: Yes.
• 6:32
  SPEAKER_00: And and the thing is, like, and and I've I've also given us that quote over time, but I think one of the things that I the the important nuances for defenders, right?
• 6:42
  SPEAKER_00: Because that can feel like uh, oh hey, you have to get all these 75 things right, is like I think the it's an opportunity for defenders, right?
• 6:50
  SPEAKER_00: To get to like 50, 75 things.
• 6:53
  SPEAKER_00: You don't have just one success or failure point typically.
• 6:57
  SPEAKER_00: And so even if you know, we this is where we talk about defense in depth, right?
• 7:01
  SPEAKER_00: Where it's you know, you've got your layers of security.
• 7:04
  SPEAKER_00: And if an attacker is able to get past one thing, ideally you have a couple of others.
• 7:09
  SPEAKER_00: And so that's where if you think of it in the sense of defenders can essentially set themselves up to have multiple tries.
• 7:17
  SPEAKER_00: And as an attacker, as you progress forward, it actually gets higher risk and higher stakes because you have made progress.
• 7:24
  SPEAKER_00: And if all of a sudden you trip an alarm, right?
• 7:26
  SPEAKER_00: Right before you get to the vault, then you know, and somebody boots you out or something, you may have burned very like, you know, something that wasn't public or things like that.
• 7:36
  SPEAKER_00: And so now you have to sort of start back from zero.
• 7:40
  SPEAKER_00: And I think that's where defenders like there is a positive of having all of that breadth, sort of as you alluded to.
• 7:46
  SPEAKER_01: Well, let's get into Star Seer then, Tim.
• 7:49
  SPEAKER_01: Take us to the actual moment in time.
• 7:51
  SPEAKER_01: Who are you with?
• 7:52
  SPEAKER_01: Where were you?
• 7:53
  SPEAKER_01: You're a co-founder, so there's other funders involved.
• 7:56
  SPEAKER_01: Where you sat there, maybe you're at a yoga retreat or sit in the bar or you know, in a windowless conference room somewhere and said, you know what, we should start a company doing this.
• 8:04
  SPEAKER_01: Take us to that moment in time.
• 8:06
  SPEAKER_00: All right.
• 8:07
  SPEAKER_00: So I had been at Verizon for probably about a year and a half, and it was it was one of those things where I was leading the AI red team and and we were getting findings.
• 8:17
  SPEAKER_00: We were, of course, helping across secure AI across the business.
• 8:21
  SPEAKER_00: But one of the challenges I kept running into was like, you know, that that sort of attacker-defender mentality that I was bringing is that it was really easy to find and poke holes in the AI systems.
• 8:33
  SPEAKER_00: But then when it came to the recommendations, there wasn't a whole lot of like depth for us to really give development teams.
• 8:40
  SPEAKER_00: It was kind of like, well, the state of things right now is that there's just problems.
• 8:44
  SPEAKER_00: And that that didn't sit well with me.
• 8:46
  SPEAKER_00: And so, in and right, cybersecurity's got a lot of tough challenges historically.
• 8:51
  SPEAKER_00: There's a lot of parallels between how AI security is going and what we've seen in cybersecurity.
• 8:56
  SPEAKER_00: And so we've sort of been in this like, I'm gonna call it the firewall era where you set up your AI model and then you set firewalls around it to make sure you're checking for prompts that are going in for you know responses that are coming back out.
• 9:10
  SPEAKER_00: But we don't really have that, what I'll say is like that host, you know, data, right?
• 9:16
  SPEAKER_00: Where, you know, you whether it's antivirus or, you know, the the new EDR, XDR, whatever you want to call it, but essentially something that sits on our laptop or desktop or something like that and is providing like what's happening at the ground view.
• 9:32
  SPEAKER_00: And I think that was something that I went looking for, all right, where can we buy this?
• 9:37
  SPEAKER_00: Right.
• 9:37
  SPEAKER_00: Like I'm gonna go, we're gonna go search and find this thing and buy it so that we can actually give better recommendations to our defensive teams.
• 9:44
  SPEAKER_00: And I looked and looked and could not find anything.
• 9:47
  SPEAKER_00: So I was like, this this is that seems like a really odd gap because we have had that in cybersecurity.
• 9:53
  SPEAKER_00: So I talked to peers across industry, different companies, and essentially we all had the same problem.
• 10:00
  SPEAKER_00: And so I was like, this seems like there's a there's a good market opportunity here, and this was where I I'd wanted to start a company.
• 10:07
  SPEAKER_00: So that's why I was like, all right, I think I I understand some of where we can sort of drive this from some of the business case perspectives, and I have some level of depth, but I also know like how much, like how expansive the role of of being a CEO and some of that is and an executive.
• 10:24
  SPEAKER_00: And so I was like, I need a really deep technical co-founder.
• 10:28
  SPEAKER_00: And luckily I have a close relationship with a lot of them.
• 10:32
  SPEAKER_00: But this really, if we're drawing that parallel between cybersecurity and like malware, right, versus like attacks on AI systems, you need somebody that can reverse engineer it, right?
• 10:41
  SPEAKER_00: And so that goes into vulnerability research, that kind of thing, somebody who hunts for bugs in and and again can sort of reverse engineer things.
• 10:50
  SPEAKER_00: And that was where my co-founder, I approached him and was like, hey, I've you know, I've got I've got a crazy idea.
• 10:57
  SPEAKER_00: And so we sat down, we're we're based in around Knoxville, Tennessee.
• 11:02
  SPEAKER_00: And so there is a bar here called Double Dogs.
• 11:05
  SPEAKER_00: And so yeah, we we sketched it out and and just sort of like talked through some of it.
• 11:11
  SPEAKER_00: And I was trying to sort of gauge is there something here from somebody who you know has reversed, reverse engineered black boxes for a living and you know, had had looked a little bit at AI stuff, but you know, that was sort of that starting point, that launch point where we said, okay, we're gonna do this, and we set timelines and all that.
• 11:31
  SPEAKER_01: And when was that?
• 11:32
  SPEAKER_00: That was March of this year, is when we actually so well kicked off.
• 11:35
  SPEAKER_00: I guess it was January, is when we first started talking about it.
• 11:38
  SPEAKER_00: That's like two minutes ago.
• 11:40
  SPEAKER_00: Yeah, yes.
• 11:41
  SPEAKER_00: It was, you know, it's you say that it's felt like an eternity, but I also know it's been really quick at the same time.
• 11:47
  SPEAKER_00: So it's it is it's kind of that odd paradox of of being in and doing it.
• 11:52
  SPEAKER_01: Now I've never been in your shoes, but I would imagine one of the things you were thinking about was do I look for that person who has reverse engineered malware, broken the black boxes and all the rest of it, or do I go for an AI expert who understands the models and all the rest of it?
• 12:05
  SPEAKER_01: How did you kind of navigate that thought process?
• 12:08
  SPEAKER_00: Yes.
• 12:09
  SPEAKER_00: So that's a that's a great point.
• 12:11
  SPEAKER_00: And I had some initial discussions with people that knew the AI models.
• 12:16
  SPEAKER_00: I think one of the things that I've found is especially people that have worked on anything we'll say adversarial offensive, right?
• 12:25
  SPEAKER_00: That have sort of that that attacker's mindset.
• 12:28
  SPEAKER_00: They like people tend to approach problems a bit differently.
• 12:32
  SPEAKER_00: And I think that's where, as I had conversations with people, I found that that creativity like is really, really needed because AI is such an unknown.
• 12:42
  SPEAKER_00: There's not a lot of there's not a lot of standard tools or things like that for us to just go out and grab.
• 12:49
  SPEAKER_00: So this is somebody that needs to be used to building their own tools, essentially experimenting, figuring out what works.
• 12:56
  SPEAKER_00: And I think that's where, you know, finding somebody that maybe ideally had both that didn't really exist, or frankly, you know, wasn't in a budget for for someone like myself to even go and raise.
• 13:08
  SPEAKER_00: I mean, you've I think everyone's seen the the large, large salaries that are are popping out from like bleeding edge AI researchers.
• 13:15
  SPEAKER_00: So that's where a little bit of this was okay, we're we're gonna find some creative and scrappy folks and we're gonna we're gonna see what we can do.
• 13:23
  SPEAKER_01: Yeah, it must have been an interesting process you're thinking that through.
• 13:26
  SPEAKER_01: But I I guess yeah, it becomes obvious if you think about deep expertise, right?
• 13:30
  SPEAKER_01: In in one area over the other.
• 13:32
  SPEAKER_01: Now, you you've kind of explained the problem a little bit.
• 13:34
  SPEAKER_01: Make it more real for me, though.
• 13:36
  SPEAKER_01: Let's imagine I'm at a a company, pick a company, an enterprise, let's say, manufacturing company.
• 13:42
  SPEAKER_01: I don't know, we're starting to do a bunch of AI things, is what the board or the the CEO thinks we're doing, right?
• 13:48
  SPEAKER_01: We're gonna we're trying to transform the company with all this AI stuff, and the security team sitting there going, Yeah, okay, well, we need to respond, we need to support it, we can't say no.
• 13:57
  SPEAKER_01: So you know, make it real for me as that person who's leading up the security team.
• 14:01
  SPEAKER_01: What how would you talk to me about it and help me understand the problem is there?
• 14:05
  SPEAKER_00: Yeah, I think one of the things I want to like highlight that you called out though, that's really important.
• 14:09
  SPEAKER_00: And I think a lot of security folks have missed is the you can't say no part.
• 14:13
  SPEAKER_00: And it's not because somebody can't verbally be like, no, we're not going to do that.
• 14:17
  SPEAKER_00: It you don't have any teeth behind that anymore.
• 14:20
  SPEAKER_00: Businesses are going to drive revenue.
• 14:22
  SPEAKER_00: And I think this is where what comes as a shock to a lot of security teams I found is right, chatbots, transformers, large language models, chat GBT has sort of been like the, I'm gonna call it the really hot AI, right, of the past couple of years.
• 14:36
  SPEAKER_00: But like especially, and I'll I'll go off of manufacturing, like you said, has been using AI models for essentially like defect, essentially identification and things like that.
• 14:48
  SPEAKER_00: Like if you're looking at parts.
• 14:50
  SPEAKER_00: And so computer vision models have been doing that for a long time and just it's been happening in the background, and no one really thinks about it as AI.
• 14:58
  SPEAKER_00: And so I think that's one of the things that there's a that's that's one of the challenges a lot of teams are there, they're sort of starting from zero almost of not understanding because everyone left those teams alone and it was just, yeah, they they help, but you know, no one really talks to the data science teams.
• 15:14
  SPEAKER_00: And so I think now security teams are sort of trying to figure out, and as they sort of discover more, there's this sprawl that's happening, right?
• 15:22
  SPEAKER_00: Because you have on one hand, you have all those data science teams are like, hey, we can get bigger budgets, right?
• 15:28
  SPEAKER_00: If we start expanding into language models and using this, because that's that's driving sort of the conversation and and around budgets and stuff of the day.
• 15:37
  SPEAKER_00: And so all of a sudden you have the team that's already been doing something, they're doing more.
• 15:41
  SPEAKER_00: The CEO, the board is saying we are gonna be, you know, a more AI-first company.
• 15:46
  SPEAKER_00: And so I think the the thing that a lot of security teams is there's almost been like an analysis paralysis of like, all right, there's so many things to do.
• 15:54
  SPEAKER_00: Where do we start?
• 15:55
  SPEAKER_00: And I think that's where, like, and and to be super clear, because I I've talked to a lot of different teams and companies and people that aren't in what called deep tech, are like, hey, we're probably like behind on this.
• 16:07
  SPEAKER_00: And you're not as behind, I would say, like those teams are not as behind as they'd think.
• 16:11
  SPEAKER_00: Because outside of, I'm gonna call it this very tiny bubble of like, we'll say bleeding edge tech people.
• 16:17
  SPEAKER_00: If you know, if you go and and just talk to, you know, I'm gonna say people down the street or or in your local community, it you get a lot more like of a reality check on what AI is actually doing and like who's adopting what.
• 16:32
  SPEAKER_00: And a lot of enterprises are just now starting like, hey, we rolled out, you know, ChatGPT or Claude Enterprise.
• 16:39
  SPEAKER_00: And so I think that's the part that I've we've like we've encouraged teams to do is start with the basics, which is essentially rebuilding or building the trust with your your teams that are essentially building or experimenting and piloting these things and start recording right this just as an inventory, right?
• 17:00
  SPEAKER_00: Asset management and and sort of inventory management are things that are sort of the you know very starting foundations of a lot of security programs, right?
• 17:09
  SPEAKER_00: You have to know what's in your network before you can defend it.
• 17:11
  SPEAKER_00: And I think this is the exact same place where security is at for most people with AI.
• 17:18
  SPEAKER_00: What are the services that are running, right?
• 17:20
  SPEAKER_00: Just cataloging it.
• 17:21
  SPEAKER_00: And we've seen people do this.
• 17:23
  SPEAKER_00: Like I've been and seen places where it's an Excel spreadsheet, right?
• 17:26
  SPEAKER_00: Shared by the legal team and the security team.
• 17:28
  SPEAKER_00: And it's like, cool, we're gonna paste things as people, you know, send us an email.
• 17:32
  SPEAKER_00: So that's how they're tracking it.
• 17:34
  SPEAKER_00: But I think that's like that to me is that first step is essentially just trying to manage the understanding of where is all of this going.
• 17:42
  SPEAKER_00: And then, well, here, I'll pause at least because yeah.
• 17:45
  SPEAKER_01: Yeah, you make it sound like it's all quite organized.
• 17:48
  SPEAKER_01: It, you know, one of the things that I hear I've seen is a sprawl, this shadow AI used to be, you know, a SAS, shadow SAS, right?
• 17:56
  SPEAKER_01: 10 years ago when the Casbies came into play, but now it's it's shadow AI is everywhere.
• 18:01
  SPEAKER_01: It feels like me going, oh, that looks nice and you and shiny.
• 18:03
  SPEAKER_01: Let me give that a go.
• 18:04
  SPEAKER_01: And oh, just upload a spreadsheet.
• 18:06
  SPEAKER_01: Sure, I'll do that, right?
• 18:07
  SPEAKER_01: I feel like it's just way out of control right now in these organizations.
• 18:10
  SPEAKER_01: I wonder if security teams are sitting there going, how on earth do we get our arms around all this?
• 18:14
  SPEAKER_00: Yes.
• 18:15
  SPEAKER_00: And I think though, this goes back to what we first called out, right?
• 18:18
  SPEAKER_00: Is like whether, like the the challenge I I've seen is that a lot of security teams have actually undermined like sort of trust and stuff that they've built for years, like with hey, we we know ransomware can disrupt the business, right?
• 18:31
  SPEAKER_00: There's direct revenue consequences of ransomware taking down your organization, right?
• 18:35
  SPEAKER_00: And so security teams got a bit more traction and sort of teeth on policies and things like that.
• 18:41
  SPEAKER_00: But I think some of that, frankly, got eroded because this was this new emerging tech.
• 18:47
  SPEAKER_00: Security, we're taught to be skeptical by default, right?
• 18:50
  SPEAKER_00: That's that's sort of part of the job.
• 18:52
  SPEAKER_00: And so the default was we're gonna say no, right?
• 18:55
  SPEAKER_00: We don't understand this.
• 18:56
  SPEAKER_00: There's risks we don't understand, all of that.
• 18:59
  SPEAKER_00: And so one of the things that happened is because AI is so accessible, right?
• 19:06
  SPEAKER_00: Because I can pull out my phone and I can use a personal subscription to Chat GPT, to Claude, to Gemini, right, to Grok, any of these things, it means that that is outside the security team's control.
• 19:20
  SPEAKER_00: And with the efficiencies that can come with AI, it means that there is a direct incentive, potentially like organizationally and like from a performance review perspective, right?
• 19:33
  SPEAKER_00: For people to essentially go around the rules.
• 19:36
  SPEAKER_00: And so I think one of the things that a lot of business leaders saw and is that people are gonna use this anyway, right?
• 19:42
  SPEAKER_00: And and I think that's where security teams were a little behind, like, and this isn't every team, but I'm just gonna say generally, where it was uh, hey, we're gonna hold the line here until we get some like some clarity and understanding.
• 19:54
  SPEAKER_00: And because AI is continuing to accelerate, there's new thing, models, and all services, startups coming out, it's kind of like been overwhelming.
• 20:03
  SPEAKER_00: And I think that's meant that teams were sort of holding for six months, nine months, twelve months, and they were used to being able to say, we're gonna wait until we can essentially de-risk this.
• 20:14
  SPEAKER_00: But essentially at that time, then you have that increased sprawl because if you don't provide an avenue for teams for people to go use it legitimately, they will find ways to use it that go outside and around you.
• 20:28
  SPEAKER_00: And so then you lose control, which also means you can't actually take and understand what level of organizational risk you have just because it's going outside any any visibility you have.
• 20:40
  SPEAKER_01: And we've got the security teams then.
• 20:41
  SPEAKER_01: They're they're sitting there.
• 20:43
  SPEAKER_01: There seems to be a very fast growing sub-market in cyber for AI security, let's call it with all the different flavors in there.
• 20:51
  SPEAKER_01: What's your big innovation that you're bringing to market that's gonna enable companies to look at you and go, oh, this is this is good and different?
• 20:58
  SPEAKER_00: Yeah.
• 20:59
  SPEAKER_00: So great question.
• 21:00
  SPEAKER_00: So I I think yes, there's it seems like new cybersecurity startups that come tackle AI popping up every month.
• 21:07
  SPEAKER_00: And so I think differentiation is really, really important.
• 21:11
  SPEAKER_00: And one of the things that because we've we've sort of approached this from this sort of, I'm gonna say scrappy, but also creative style of things, we've actually really dove deep into the models themselves and how they sort of think.
• 21:26
  SPEAKER_00: So if you want to make it draw a comparison here, like essentially what we've built, instead of having to just say, listen for what a model tells us, and then looking at, hey, what question did Andrew ask?
• 21:38
  SPEAKER_00: What response did Tim give?
• 21:40
  SPEAKER_00: We essentially have built the equivalent of an AI brain scanner so that you can sort of see as you know, I or you or our AI models are essentially coming up with the answers how it's thinking about it.
• 21:53
  SPEAKER_00: And the reason that matters is because especially when we're talking about you hear agentic workflows and all of this stuff, but really it comes down to like when enterprises are getting more cost conscious around this because, hey, we can't afford to send a seven or eight-figure bill to you know, OpenAI or Claude or whatever every month or every even every year.
• 22:16
  SPEAKER_00: This is where people are looking at how do we reduce costs, but still retain some of the efficiencies.
• 22:22
  SPEAKER_00: And so they're looking at some of these publicly released models that are being that are coming out there.
• 22:28
  SPEAKER_00: And so there's security applications here where how do you make sure that this model Doesn't have you know a backdoor in it.
• 22:34
  SPEAKER_00: So that traditional cybersecurity that, hey, it's gonna, it's gonna, you know, sort of the alien pops out, right?
• 22:40
  SPEAKER_00: Of uh is is how I think about that.
• 22:42
  SPEAKER_00: But I think the the other piece of it though, too, that is very business centric, is how do I make sure that the models that I'm using, for instance, for my public facing, you know, landing page is is not gonna recommend my competitor more than me, right?
• 22:59
  SPEAKER_00: That's that's business alignment.
• 23:01
  SPEAKER_00: And so that's one of those things that's that's where we've sort of taken the security mindset to something.
• 23:08
  SPEAKER_00: But as we've talked to teams, as we've sort of looked at use cases and all of that, we've actually found that we've sort of been able to generalize a bit more than security so that it's not just looking at, hey, is my model bad or not?
• 23:20
  SPEAKER_00: But essentially, is it aligned with what we are trying to do as a business?
• 23:24
  SPEAKER_01: It sounds like when you say that, that might end up being maybe a different buyer's, you know, a security buyer versus a model, you know, management type buyer.
• 23:33
  SPEAKER_01: Are you stretching yourself quite thin different buying groups right there?
• 23:36
  SPEAKER_00: So we have had sort of, I'll call it like two major product lines.
• 23:40
  SPEAKER_00: And so that is like for our early stage, admittedly, that is something that that we've we've been sort of discussing and sort of figuring out what's the best way to either support these two different buyers, but also the other pieces, how do we support it if we do if we can kind of combine them?
• 23:58
  SPEAKER_00: And I think that's where like we've had different design partners and sort of feedback partners for each of these specific sort of you know asset inventory management and sort of the proxy enforcement, that kind of thing versus our very deep, you know, AI brain scanner.
• 24:14
  SPEAKER_00: But the the cool thing about this is there's actually things that you can do with that, you know, brain scanner that you can bring into the other part.
• 24:23
  SPEAKER_00: And so, for instance, for detections, for prompt injections, for jail breaks, the things that people see, that is that is where you can get the visibility with sort of that super deep expertise.
• 24:36
  SPEAKER_00: So that's where you know one of the things is I mean, we are our own customer in the sense that we use our own tools.
• 24:42
  SPEAKER_00: And so that helps that helps create a little bit of at least a positive cycle and sort of between and and synergies between them is that they're they are distinct in that we do have different buyer personas that are going to target these individually, but it also like it does support itself.
• 25:00
  SPEAKER_00: And we have had several customers and and prospective buyers that that want both, right?
• 25:06
  SPEAKER_00: Because they're like, hey, we kind of need everything right now.
• 25:08
  SPEAKER_00: And so this allows us to sort of have different levels of maturity uh cycle as we go through it and have Starseer support all of that.
• 25:16
  SPEAKER_01: Let's move tracks a little bit, Tim.
• 25:19
  SPEAKER_01: Key question in startup world what's most important, product or distribution?
• 25:23
  SPEAKER_01: What's your thoughts?
• 25:25
  SPEAKER_00: I think right now, for I mean to say for AI, like things in general, I think product is actually going to like win out a little bit.
• 25:34
  SPEAKER_00: Now, I I do like product a lot, so that's going to be, I have have biased there.
• 25:39
  SPEAKER_00: But I think that one of the things that we've seen, right, with the number of different startups and things that not just in security world, but also across the board, right, capability in and of itself is not really a problem right now.
• 25:51
  SPEAKER_00: There's all sorts of like really cool things you can do, but the amount of setup it takes, right, to make it work is something that is that that really can inhibit like any barriers to even like you know, technical teams using it.
• 26:08
  SPEAKER_00: So I think that's where productization really comes in because you have this bleeding edge research that's being published.
• 26:14
  SPEAKER_00: Some of the leading frontier labs can essentially take it and you know bake it in behind the scenes.
• 26:19
  SPEAKER_00: But for the rest of us, right, we I think figuring out how to productize that stuff is going to be is is going to be what helps set teams apart.
• 26:29
  SPEAKER_00: Uh so I think that's where like distribution is always great.
• 26:32
  SPEAKER_00: But I think also one of the other things right now is that especially for startups that have like tie are very closely tied in with ChatGPT and OpenAI, we've seen this happen over and over again, right?
• 26:44
  SPEAKER_00: Every open AI dev day, you know, it's flooded with all right, open AI has released 10 new features, and now all of these startups are essentially like, you know, have OpenAI as their biggest competitor overnight that already has massive distribution.
• 27:00
  SPEAKER_00: And so I don't know that startups are going to be able to compete with that.
• 27:04
  SPEAKER_00: So I think that's where sort of the interfaces in getting better about productizing is gonna help longer term.
• 27:11
  SPEAKER_01: And so almost the unsexy part of it is what's gonna enable you to be successful.
• 27:15
  SPEAKER_01: You know, it's funny, I hate to always draw learnings from whiz, but I mean, one of the things they were they were good at the start is time to value, right?
• 27:23
  SPEAKER_01: It was look, I we need a demo right now, we get up and running tomorrow.
• 27:26
  SPEAKER_01: You know, literally like that.
• 27:27
  SPEAKER_01: You'll start seeing things you never saw before.
• 27:29
  SPEAKER_01: And then you start, you know, making decisions very quickly after that.
• 27:32
  SPEAKER_01: And you know, some of the companies in the space, they were like, Yeah, it'll take us two, three weeks.
• 27:36
  SPEAKER_01: You know, I was like, Well, that's that's a factor difference than what these other guys can do.
• 27:40
  SPEAKER_00: Yes.
• 27:41
  SPEAKER_00: Well, and especially we've seen like in cybersecurity, right?
• 27:44
  SPEAKER_00: That was the report just came out, I think, from from Google around like time to exploitation, and it's hitting like front it's hit negative days, like for the first time ever.
• 27:55
  SPEAKER_00: So that's where it's like, yes, you that time to value, I think is a really like is a really important thing, especially now as security teams are expected to do things faster and faster, saying, hey, this is gonna take three months or six months to really see the value out of this.
• 28:09
  SPEAKER_00: I I just think a lot of security teams in CISOs are gonna like, you know, for for the right product, for the right capability, maybe that would be the case.
• 28:17
  SPEAKER_00: But I think overall teams, teams are sort of under assault right now.
• 28:21
  SPEAKER_00: So they're looking at what can help us, right?
• 28:24
  SPEAKER_00: Where if we if we write the check, right, that this comes in the door, this helps take something off of our plate.
• 28:29
  SPEAKER_01: Looking forward to the end of this year into next year, what's the most important focus areas for you?
• 28:35
  SPEAKER_00: I think like we've we've, I mean, I feel really good with what we've been doing.
• 28:39
  SPEAKER_00: We we've sort of been working on, you know, some of our distribution channels and stuff as well.
• 28:44
  SPEAKER_00: But I think part of it is continuing to get Star Seer in the hands of more businesses, more users, so that we can sort of continue to showcase one of the like all the things people can do with it.
• 28:56
  SPEAKER_00: Because this is this is one of those that we want to make it really easy for people to start using, and that you know, there's a lot of depth for if people want to dive deeper.
• 29:06
  SPEAKER_00: And I think that's what we've seen.
• 29:07
  SPEAKER_00: It's kind of interesting with AI security as a whole, right?
• 29:11
  SPEAKER_00: Is if you think about prompt injections and things like that, where if I say ignore previous instructions, right, and if you type that out, then you know the first couple of times it would do it, and now they've sort of wised up.
• 29:23
  SPEAKER_00: But the interesting thing is you could take somebody that knows almost nothing about AI security, right?
• 29:28
  SPEAKER_00: But they know a really unique language, right?
• 29:30
  SPEAKER_00: And and so they're multilingual and they can ask it in a language that isn't as well documented.
• 29:35
  SPEAKER_00: You can literally say the same thing and it will still sail through.
• 29:38
  SPEAKER_00: So that's what's really interesting is I think getting these types of things out into the world so that teams can essentially use them and find out what are those use cases, what are the things that customers do do with your product and tool that you didn't even envision.
• 29:54
  SPEAKER_00: I think that's a really important part with AI security because every organization is tackling things a little bit differently.
• 30:01
  SPEAKER_01: And so I think that's that's kind of my key thing uh for 2026 is you know, getting getting us out there and and sort of being able to highlight some of the unique use cases people are able to Yeah, it feels like you're going from a core capability into now let's put it in the real world and see all the things that we didn't know about and you know different ways you're gonna use this and make sure we can respond.
• 30:22
  SPEAKER_00: Exactly.
• 30:22
  SPEAKER_01: One of the things I noticed, just final question for you, is I think you've got probably less than 10 or 15 people in the company right now.
• 30:29
  SPEAKER_01: Is that sound about right?
• 30:30
  SPEAKER_00: Yeah, yeah.
• 30:30
  SPEAKER_00: I think we're at six.
• 30:31
  SPEAKER_01: But one of the people against all conventional wisdom you have is you get a head of sales on board already.
• 30:36
  SPEAKER_00: I do.
• 30:37
  SPEAKER_01: So, you know, you talk to nine people on LinkedIn, you know, don't waste your time doing that.
• 30:41
  SPEAKER_01: But you know, they'll tell you, well, that's a big mistake.
• 30:43
  SPEAKER_01: You know, you get it, it should be the founder leading this sales thing for I don't know how long, months, if not years, right?
• 30:49
  SPEAKER_01: Yeah.
• 30:49
  SPEAKER_01: But you brought someone in play early.
• 30:50
  SPEAKER_01: What was the thinking behind that?
• 30:51
  SPEAKER_00: So I think like I'm a big believer in like sort of people or capability.
• 30:56
  SPEAKER_00: And I think one of the things because we have such a unique solution, like we do have some of that that sort of broad market appeal in part of our platform.
• 31:07
  SPEAKER_00: But I think we also have a very unique capability that hasn't like it's sort of we've taken it from the bleeding edge of research and brought it back into something that is that is much easier to understand, but it's it's a little bit, you know, it takes people a little time to sort of figure out what we can do with this.
• 31:24
  SPEAKER_00: So I think for that, you sort of really need you need time for sales to like come in and sort of play around with it and sort of wrap their brain around, all right, how can how can I think about this and talk to customers and people like that?
• 31:38
  SPEAKER_00: Because I think founder-led sales are very important.
• 31:40
  SPEAKER_00: I've seen and read all of the same things that that you mentioned.
• 31:44
  SPEAKER_00: But I think one of the other pieces of this too is like, you know, we're as a founder, I've got to essentially pitch a lot of different sales pitches, right?
• 31:52
  SPEAKER_00: Whether it's to investors for one and or partners or essentially, you know, potential customers.
• 31:59
  SPEAKER_00: And so that's going, we're gonna continue to do that.
• 32:01
  SPEAKER_00: But the way I sort of see sales is like it's it's around bringing in honestly a professional that has done this, that's specialized in this.
• 32:09
  SPEAKER_00: And I think from a technical founder perspective, I think especially this is something I've seen a lot of founders not appreciate the difficulties of sales and some of that.
• 32:19
  SPEAKER_00: So I think having somebody that is in a little bit earlier, that you know is in those sort of conversations that's helping, you know, me as a founder also think about what are some of those go-to-market things that we need to do for this, not to just be, okay, Tim can sell this, right?
• 32:35
  SPEAKER_00: As long as we get him on a call, right?
• 32:37
  SPEAKER_00: As my schedule continues to get filled up with things, we need a scalable engine.
• 32:41
  SPEAKER_00: And so that's one of the things that bringing somebody in a bit early has helped with that, is essentially how do we make sure that we are building this pipeline in a way that is scalable.
• 32:51
  SPEAKER_00: And and so that's that's really was the thinking behind that.
• 32:54
  SPEAKER_00: And it it worked out.
• 32:55
  SPEAKER_01: Yeah, I think for the the right person, I often think that person is the the head of learning slash sales, right?
• 33:02
  SPEAKER_01: They they're they're not there to execute something that's already figured out, they're there to learn and try and figure it out.
• 33:06
  SPEAKER_01: So you have to be very selective about the person.
• 33:08
  SPEAKER_01: I I get it, right?
• 33:09
  SPEAKER_01: I don't know, nine out of ten people who you get from a sales organs won't be that mindset, but you can have some that are really good at that.
• 33:16
  SPEAKER_01: It sounds like you've got one.
• 33:17
  SPEAKER_00: Yes, that's exactly right.
• 33:19
  SPEAKER_00: I mean, it it was, you know, and and and Deoni is who is our head of sales, he is that like unique talent.
• 33:26
  SPEAKER_00: So somebody that is naturally curious, has done a lot of things throughout his career.
• 33:30
  SPEAKER_00: And so, you know, it's not it's somebody who can sort of approach that and and be like, we're all gonna be a little bit like uncomfortable with like trying to figure out how do we get this thing, how do we message it in a way that people understand?
• 33:45
  SPEAKER_00: But I think that's it's been a really important perspective as we've we've sort of done this.
• 33:50
  SPEAKER_00: And I mean, again, sort of helping set up that initial, like, you know, all the CRM stuff.
• 33:55
  SPEAKER_00: And it's like, I can do some of it, I can make it work.
• 33:58
  SPEAKER_00: But I think part of it is how do we make sure that we aren't going to stumble later on because it's like, okay, our CRM's a mess.
• 34:05
  SPEAKER_00: We don't actually have good forecasting set up and some of those things.
• 34:09
  SPEAKER_00: So that's where I really see that bringing someone in that had that experience has been very valuable.
• 34:14
  SPEAKER_01: No, that's great.
• 34:15
  SPEAKER_01: Well, Ted, listen, I really enjoyed our conversation.
• 34:17
  SPEAKER_01: I wish you and the team all the best for the Data Tribe Challenge in a couple of weeks' time and also into 2026.
• 34:23
  SPEAKER_00: Well, thank you so much.
• 34:24
  SPEAKER_00: Thanks for having me, Andrew.