DrZeroTrust: You guys have just sort of released a capability at Mars is what we're calling it. Minlo agent runtime security. So let's talk about what Mars is. We're not going to Mars, we're already there, apparently. Yeah. Right. So no, we're very excited about this, obviously with a lot. We are forecasting the next ⁓ users. We'll be Agents. So the next billion users, they're not gonna be the human workforce, they're gonna be the agent workforce, right? So we really need to put guardrails around that and control it. So being able to authenticate, because a lot of times me, if I'm going out and doing a lot of research, I'm gonna let that agent use some of my credentials. And I wanna make sure that it's going and getting the information that I need with my credentials, right? So that's a security risk potentially. So we have guardrails we can put around that. And in addition, if the you just wanna make sure that the agent also is not taking a different command because as we both know, right, the human eye can see certain things on the file, but the agent can see even more than that, right? It can see embedded And is that governed at the policy level within like the backside of Minlo? So Minlo, you're setting up the policies and then Mars is doing its thing and you're able to basically kind of put guardrails around what the agents are doing. And containerized application side of it and everything as well. Yeah, so like just like we do with the humans, right, where we put them into isolation with our technology, we're gonna do the same thing with the agent. We're gonna put them into isolation as well. Okay. So it depends on the type of The ⁓ if it's browser based agent, those would be detected through LLM channel. If it's ⁓ autonomous agents, we would be proxying them through the Menlo Secure Cloud Browser Platform. it might be other types of agents where we would proxy them through Menlo. When we proxy them through Menlo, we can provide DLP, what the agent can see and what it cannot see. we can prevent threats from coming in, whether it's being attacked with indirect prompt injection. We have file-based security where we can remove hidden content that might be tricking the agent and hijacking its its goals. Interesting. so we're very excited about the additional capabilities that it brings to table. And also, human in the loop is another aspect that we make sure because if you have capture, right, the agent will not know. what to do with that. So we will still maintain the human in the loop. It'll it'll know, hey, I need to go to a human to go get access to this additional information. Okay. So I mean honestly the the long hole in the tent here is like agents are the future kind of workforce. It's gonna be explosive ⁓ and you guys are already lining up on a way to put guardrails around that, but it's very native experience. It's not I mean I think that's where a lot of folks are getting a little bit kind of derailed on these things is is it going to be duplicative or is it gonna throw off the way that I do my work because now we're learning to do agentic whatever, but you guys are just saying keep doing what you're doing. Yeah. We're just gonna help basically put some controls that should be there around it. Exactly. And just think of it as like another person on your team. You're gonna be able to give it some commands, some things that you want it to do, it'll go do that and then it'll come back and you can make sure because if we put everything into isolation, it's not going to bring something back to the end user. So worst case like bad stuff will just happen out there on the internet. Which is fine. That's kind of the battlefield anyway. you're okay like the that's why I like I've been such a big call it proponent of like browser isolation, these types of things, 'cause I can live with my bad stuff happening out in the yard as long as it's not happening in my house. Exactly right. Because we have everything in isolation. Even when like say you they get a file that ha is contaminated to some d degree, when it comes into our isolation, we're gonna scrub it clean. So by the time you're able to view it, there's nothing there that's harmful. And that's l like a legacy. ⁓ thing that you guys have been doing for a long time, which basically de weaponizing the application stuff and downloads and all those things. Exactly right. And so with the acquisition that we did with Votero, right, with the CDR capabilities, lots of additional capability there. Where if you're looking at that file, most technologies tend to sandbox things. We don't do that. Okay. We will scrub it clean real time as it's a transition. And then let's just say you want to make like you have more access to everything on the file and I don't. I'm not allowed to see PII information. We can also mask that so the right person seen the right level of information and do the same with the agent. ⁓ really? So asking a redaction on the agentic level too. Okay. Wow. That's cool. So basically we were everything we were able to do for the human, we just paralleled that for the agent part. Which makes a lot of sense. I mean that's I think that's a really smart way to approach the problem is treat agents like people. Like people, or a member of your team. That's probably smarter than people in most cases and probably a more useful and they work longer and those types of things. They read twenty four hours a day. And they don't complain. And they don't complain. They don't get sick. My favorite thing about working with Claude right now is like I don't have to feel bad about telling it that it screwed something up. You know? I just like Claude, no that's wrong and it fixes it. Exactly. You don't have to deal with all the ⁓ hey would you please fix this? Right. Claude doesn't complain. Yes, but the thing with the age is they don't have logic. So with that human in the loop, that's still something that we have, right? Because you need like I know I asked I had some research that I was doing It came back and did that. I was gonna do a large ⁓ presentation and you can see some of the information that it had in there. I was like, wow, it's almost like I have to treat them like an elementary school, this agent, or I have to go and check its work. Right. So I had to clean it up a little bit. So think of agents as a member of your team, but still not as human as you are with your logic and capabilities, right? Yeah, I I've got a ⁓ slide and a research paper I'm working on that's literally titled ⁓ Agents Are Like Teenagers. Yeah. That's kinda how I see ⁓ right? Absolutely. Well, they think they're smarter than you. They're only gonna tell you what you probably, you know, will respond to and they're gonna kinda do their own thing based on the little bit of guardrails you give ⁓ So Exactly. Sounds a lot like teenagers to me. I've I've got a couple of those and that is exactly how they act. Yeah. So ⁓ for the folks that are listening, you can't see the visual here, but they actually have a pretty good graphic of the Claude AI agentic panel. it's nice 'cause it shows that it's active, it's running its analysis in real time. And then as you look there, ⁓ you know, from s like I do Claude stuff all the time, seeing that there's actually a thought log going on, that's super interesting because you're you're seeing how the agent's thinking. Yes. It bothers me that a lot of this stuff is black box. Right. You know, and you just kinda like, I really hope this is doing what I think it's. I think it's doing. But now you can see it. You can see the logic that's going behind it. And if you need to interject and put a different thing there, you can't tell it, ⁓ you know what, I want you to do this instead. Right. Or it shows something like, no, ignore that, right? Yeah. So yeah, I mean you could actually even probably get tailor your agentics stuff a little bit better because you can see that and it's not so black box. Cause that's where, you know, copilot I think is ⁓ it's copilot, right? It's Microsoft. Gemini's get it pretty good, but being able to tweak that kind of response logic feedback loop is pretty important. ⁓ yeah, for sure. Yeah. So we are very excited. We've doubled down on the agents because we recognize that's gonna be the majority of the workforce moving forward. And we needed to make sure that that was just as secure as we do for the humans. And our isolation approach, we're the only ones in the market that everything is isolated like that. And that's what keeps it off. By default, you guys take the position of isolate, right? So it's that's that's kind of a unique capability difference, right? Is ⁓ start with default deny and allow things on the other side. Exactly. Yeah, okay. And f you know, for the folks that do listen to this too, Menlo is actually a demo force customer, so With this new Mars capability, as soon as we can get it up and running on there, we'll share the link. And I say it all the time, like I just I love when companies are willing to let people try their software. So you guys have been great there. And ⁓ hopefully people can click on this and get a ⁓ insight into how it actually works. Exactly. Please go to demo force. It's a great way to see it live. And then also from there, we allow you to leverage the technology for three weeks. Okay. So we'll allow you to do that. You can See everything, we'll give you a report afterwards and we'll kinda walk you through the things that we found. wow. And it's very eye opening. Most of the customers when they experience that, they're like, ⁓ my gosh, we had no idea. Yeah, it's well people don't realize how much stuff that there's bouncing around all the time that they're, you know, not dealing with or they should be dealing with. So exactly makes tons of sense. And ⁓ what's your thoughts on the Gartner event so far at large? It has been awesome for us. Yes, we've had so many customers come by. ⁓ we have lots of large programs that we're we you know support with the government. Big in the federal space. ⁓ yeah, big presence in Fed. So ⁓ it's been wonderful. We've had people come over. A lot of the state agencies have also come over, which kind of surprised me a little bit. ⁓ once we're able to tell them all the great cross performance we've had on the federal side, they're like, wow, we need to look at that too. so I would say our state and local business is growing. very fast. Is that I was gonna ask so on the state and local side, is there a particular kind of vertical within that? Like is it education, healthcare, or is it just state and local in general? It's state and local in general. ⁓ some of the ones that have looked at us are like the city of blah blah blah. Right. ⁓ so yeah, it's state and local in general. ⁓ not too much on the education side yet. Yeah, education seems to be catching up. They don't want to take direct sales, but they also have got these kind of like closet relationships with resellers. It's it's not an easy thing. Yeah. Yeah. Very interesting. Well, so for those that are listening, this is Jacqueline from Minlo Security. Yes, hello. Nice to see all of you. And ⁓ Jacqueline Biggio. There you Biggio. I always screw up people's last names. So I was like, Bij Bij Big Yo. Big yo. Think of the ball player. There you go. Craig Biggio. Nice. ⁓ and folks, when you when you're trying this out, like go through, play with Minlo. You can't break it. have a good time learn these things. We'll hopefully get the new Mars stuff configured it up and running on there. ⁓ and you'll be blown away to see what's going on with these agent systems. So check it out. Thank you very much.