Name: Cymulate Customer Briefing - From Validation to Exposure-Informed Defenses
Uploaded: 2026-05-19T16:46:08.926Z
Duration: 50 min 50 s
Description: Cymulate Customer Briefing - From Validation to Exposure-Informed Defenses

Transcript for "Cymulate Customer Briefing - From Validation to Exposure-Informed Defenses": Hi, and welcome to today's customer only webinar, a customer briefing around Cymulate what's new and what's next. Our theme today is Beyond Validation, Agentic Cyber Defense Engineering. This is our approach to recognize that validation should not be done in a silo of proving what's what's with how what works. Validation is done for the purpose of engineering what's next, how to build better security. That's what we call cyber defense engineering. And hint, there's a piece of AI. It's the agentic cyber defense engineering. Today, I'm very excited to be here. I'm I'm Brian Moran. I I lead product marketing here at Cymulate. I've got Nir Krumer, VP of products, and David Kellerman, our field CTO. Nir, tell us about what we're gonna talk about today. AI. We're gonna talk a lot about AI. No. Just kidding. We're all gonna talk about AI. We're gonna talk about what's new in the platform. We'll integrate with AI. Lot of cool features that I wanna talk about, upcoming this launch. And we're also gonna talk a little bit about what's happening in the world with AI and missus, if you heard about it. So we're gonna have a really good time. Excellent. Thanks, Nir. David, what are you gonna show us? Because we're actually gonna show you the product. What are you gonna show us today? I think I will just follow Nir and I will say I'm going to show AI. I think that's, what's trending now. No. I'm kidding. We will also see all what Nir just mentioned in action. We will go through multiple flows, that our clients, just as the ambiance here on the call, will be able to experience very soon, with that. And we'll just see it all in action so you can get a little taste of all that. Gotcha. Alright. So we're recording this on Tuesday, May 19. You know, when can customers expect to see this, rolled out So within the upcoming days, probably early next week, they're gonna see it product? Okay. So so customers be on the lookout for communications in form of release notes, you know, an email from Simon Laid, and and then, of course, any questions, you're always gonna follow-up with your account team and your support team. product. And of course, we go through remember one thing, once you once we'll share that also, you will have an in app guide that will also walk you through the new things. Awesome. Alright. Cool. So, as we go through this webinar, couple here's housekeeping aspects. If you have a chat, please enter any technical issues you have with the chat, into the chat. Please use the q and a function to actually get any questions in. I'll be looking at the looking at that as we go. I'll be moderating that. If we have a chance to get the question into the conversation as we go, we'll address it, or we may wait to the end. Of course, this will be archive recorded, and everybody who registers, will be will be sent a a link to the recordings you can share with your team. Alright. Let's jump in. And I wanna set the stage by just, like, you know, what what are we what what are the problems we're solving? Our customers have long come to Cymulate to say, I've got I've got this expanding attack surface. I I don't know how good my controls are. We've delivered a solution for threat and control validation. Our customers have told us they have got broken, detection rules. I've got gaps in my coverage of of what I'm actually covering in in my with my detection. And so here we have SIM validation, and detection engineering. Of course, the the problems around exposure management have always been about overload of vulnerabilities, no way to prioritize or or a poor way to prioritize all those thousands of vulnerabilities that we're dealing with. So our solution there was validation and prioritization of those exposures. Now the new thing we're dealing with today is we have human speed defenses going up against machine speed attacks, and that's laying the groundwork for what our solution will roll out well today of agentic cyber defense engineering. Today's reality is that threats move at machine speed while the cyber defenses are constrained by silos and manual action. So, you know, of our exposures, we know that, hey. You know, 1% less than 1% are actually gonna get have be exploited, but the ones that are exploited are exploited within twenty four hours. More than more than a third of those ones that that do get exploited are exploited within twenty four hours. And then once they are exploited, there's no more hundred day dwell time in the hundreds of days. It's twenty nine minutes from exploit to, to breakout and actions on objectives. So super machine speed for how how fast these attacks are moving. Meanwhile, our defenses and our our cybersecurity teams are kinda still operating in silos oftentimes, where we have vulnerability management. It takes it takes a hundred and two days to to make a patch. You've got, you know, red team and threat intel team, you know, doing the controls testing doing the the threat intel, the testing validating the threats. You got security controls. You know, if you're doing this manual testing, a large mature team will spend more than forty hours a month, you know, testing controls if they don't have any automation there. To roll out a new, detection, you know, two days may be optimistic. You know, so our our SOC might take upwards of two days to a week to to deploy a new detection for for a for a threat, which you which you wanna detect. So this is a, you know, a reality of threats moving at machine speed with AI and that we're going up against, you know, the reality is our defenses also need to adopt AI, so it becomes an AI versus AI battleground. Now you're likely already doing AI. Right? If you're doing threat intel, security engineering, SOC, all the different tools out there say, hey. We have some AI to plug in here. And AI is gonna help you do those existing jobs faster. What it's not doing is providing context across those four silos. It's not necessarily providing the inputs of vulnerability management knowing what's truly exploitable from the stock, not necessarily providing the inputs of what are the exposures. And if if we remain in a silo deployment of AI, we get those fragmented signals operating locally, not having that closed loop of what was, you know, truly exploited. Was it fixed? Have we validated the fix? And SOC remains an at reactive AI triaging. We're just closing closing you know, responding events faster as opposed to building better security. And, Brian, if if I'll give one more one more, point on that. We're talking to customers if you go out to the previous slide, we're talking to customers and, like, I gave an example on AI SOC agents. And they're using AI SOC agents, kind of experimenting with the new startups that are now going live on how to better triage alerts. So they're, an incident. So we're they're reducing the number or they're handling a a lot of those, alerts, which is great. But the root cause is either the detection rules are, not configured properly or not triggering on the right things. So, eventually, we see a lot of the shift going back to how is our control really configured and what is really happening there. So doing AI in bits and pieces and in silos doesn't really help the full control plane. So it helps just in one area, but lacks context over over the overall chain and what's really happening in the organization. Excellent. So with Cymulate adopting, you know, AI and this AI versus AI battleground, Our goal is to provide that proactive AI driven security so you build secure build defenses using the latest threat intelligence as well as the full context of what are your exposures, what are your controls, their effectiveness, what can you do detection wise, telemetry rules, effectiveness, fidelity of those rules, and, of course, the latest threats, the TTPs, exploitability, the active campaigns. With this context, we now, in Cymulate, have this AgenTic security validation engine to provide the outputs of, hey. What are the how what are the improvements you can make to prevention and detection? How can you, you know, increase, you know, validated threat resilience? What's your proactive mitigation plan for your known exposures and your own gaps? Connecting the dots, providing context throughout the team so we're all working off a source of truth. Source of truth of what's exploitable, what do the teams need to do next, and how good are we today, and, most importantly, building stronger defenses. This future is Agentic cyber defense engineering. Again, this proactive approach that closes the risk to fix cap. Right? You've got security con security, you have exposure discovery and security posture management that tells you all the problems you have. You have controls that can potentially do the prevention and detection. You need to bridge those things together through validation of, you know, what can I what are what are attackers actually doing today? Let's tailor the testing to your environment. You know, let's let's give you the right testing for your your assets, exposures, and controls, your industry. Of course, it's the the safe, realistic attack simulation you know from Cymulate to give you proof and evidence of validation, what threats are prevented, what what's detected, the context of what to prioritize and what to fix, and increasingly making it easier and optimized and more autonomous to go drive the improvements. So improving your controls, providing the prevention, detection, and, of course, retesting continuously in this closed loop, closing that risk to fix gap. It's closed loop process of prove, prioritize, and adapt. Alright, Nir. So in the last, say, six weeks, we've had a lot of news in the industry about Mythos. What does this change for, for our customers, for these for the security industry? So for those of you who doesn't know MITUS or Daybreak from, from, CHaGPT, basically, it's trained AI models that are much more smarter, and they can find vulnerabilities much, much faster. With that and everybody's talking about how it changes the security landscape. Basically, what we've seen is that the discovery piece of finding new vulnerabilities, those days are are almost over. Vulnerabilities will be scarce. They will be everywhere. NVVs even saying they don't keep up if with even with today, vulnerability discovery, and there's gonna be much, much more. And when we have all of that, you can't just like build on the same processes. What becomes is that from humans that create those techniques and create POCs and creates all those, finding all those vulnerabilities, AI will start doing that, and it will do that in a much faster pace. And that shifts the ball into validation. The the the the data, like, doing it, once in a while and and and doing patch Tuesday, you'd have to start validating much, much more, constantly, the new threats that are coming in. There won't be a lot a lot of time to, like, analysis and all of that because, remember, in the past, once you there's a new vulnerability discovered, they go to NVD, they get this new CVE, then they or they and they talk to their vendor, the vendor issues the patch, there's, like, full discovery, and only then it's being released and there's time to, like, all the the patch, stuff. That time is now much more short much shorter. And security teams will have a lot more vulnerabilities that that they need to or exposures that they need to and, like, getting ready to. And that's why it's really important to validate what you really need to patch for and what you can put some sort of compensating control, in order to, prevent it. So and that's exactly, by the way, where where Cymulate sits. So validation is the new gate for all those different, attacks that are coming in. Cool. Alright. So the next few sections that you're gonna drive, so I'm gonna stop sharing, let you jump in and, share from your side and drive conversational. Yeah. Perfect. Okay. So in order to really change help the the the security team to change the control plane of cybersecurity, like, understanding how do we do AI versus AI, the AI attacker versus the AI of the defenders? How can we take the full the full control plane and and automate that as much as possible, like Brian showed on genetic defense engineering? It's not just about the control plane of identifying the the right threats, running validation, finding what are the gaps, and fixing them and iterating. It's about building a whole mechanism around it, and that's also what we're trying to do. It start with, like, each and every piece of the puzzle, there's multiple agents that we can trigger on those, or that we can generate for those that take basically the human activity of doing that specific thing and mapping it. So for example, we're creating a threat intelligence agents that go around looking for the relevant, threats that are relevant for your organization based on your security controls, based on, things that you've ran in the past, based on your industry and region. So taking all of those, identifying, and pushing it in. Then from all those risks, mapping it to the, the which scenarios from our over 100,000 scenarios that we have, what could be a good fit for driving that, simulation? From there, building the assessment, which environments is ready, which pre pre requirements we need, configuring all of that and running that, on on which agents to just select and running that. Then after we get all the findings, prioritizing which one is more more important, based on the impact, based on the security controls, based on a number of things, and eventually helping to fix a to fix those and issue reports, on demand. So that's kind of like we're building a lot of agentic, capabilities to help automate that flow, and we're putting in also new triggers that will will trigger that whole, flow. Like, there's a new threat intelligence. Hey. Let's, let's start that trigger. There's a new controller update. There's a new assessment result. There's a misconfiguration. So anytime there's stuff like that, we'll be able to, pull in and trigger that, whole flow. And that's kind of, like, what we're, starting to build, and you're seeing the the first stage of that in this upcoming launch. Let's talk about, what's new. So the first thing is Vero AI, and that's our, basically, AI name, which is the first agentic threat validation and defense engineering copilot. Vero, by the way, stands for truth, Because what we're doing here is we actually know the truth about what the environment which gaps you have in their environment, what are you really preventing or detecting and whatnot. And that will re enable the rest of the security team to know what they really need to focus on, based on what is really happening in the environment. So VeriEye is basically taking the old chatbot experience that we had, which is very siloed. You have to pick the right thing that you want and tell it what what you need to do. And it would come, like, spit up a template based on couple of, like, the old AI, like, v1 one of of AI. Now we basically built in an agentic approach where various layer of multiple agents, you talk to it with free language, and it will identify what do you want, what is your intent, will activate the right agents using their own skills and models that we already trained on our scenarios, how to run safe, repetitive scenarios in in in the environment, we'll choose the right tools and we'll run that for you within guardrails, of course. So it will, it's not only reactive, it's also proactive. If there's a new threat that is relevant for you, not just IOCs, but also IOBs, also behavior characteristics, it will trigger that there's new threats. It will trigger you to, hey. Would you like to run an assessment for those, or would you like to tell that I tell you if you have existing gaps in those, for that specific threat? Maybe you already ran some scenarios that will, help you, validate the threat that is upcoming. Again, in completing natural language, and it will know from creating new assessments to, looking in the knowledge base on how to do different things to get to creating different, dashboards using our, our dashboard creator. So it will, like, really create whatever you want based on the data in in, in the system. Besides having AI copilot in the platform, we also see a lot of customers starting to use their own, like, AI platforms, internally, like using Claude or Copilot or ChatGPT. And and when we're seeing those, they wanna interact with our system. So the way that now everybody's like, AI talks to AI is through an MCP server. So we're also launching an MCP server with Cymulate with all the tools, almost all the tools that we also have in the API. So now you don't have to, like, reconfigure all the or configure all those API interfaces with the developer, connect it to your, to your AI platform instance, and start asking questions, getting getting value from, from Cymulate. So you can harness your own AI on top of Cymulate data. So we, play nicely with all those capabilities as well. Now when we look at the customer base that we have, we have we talked about also, Brian talked about different pain points that we solve. We also see different personas. So there's persona that are really focused on security control validation. What they care about is having, running you know, they care about EDR. They care about WAF, and they wanna measure that specific thing. So we build them a purpose built lens for them that is taking them throughout the Cymulate journey, focusing them on security control validation use cases, filtering the templates that is relevant for security control validation, looking at the dashboard that's more relevant for them. So basically, focus zooming in on the thing that is relevant for them. For customers who are more threat validation, meaning they don't want to see, just one security control, they care about the threat, ransomware. I wanna see across my security controls, how how does the threat be, prevented or not, map scores to the MITRE framework. So that's the so customers that will wanna see a lens of threat validation would see more of the MITRE framework where the where you have gap from from a threat perspective. When they'll go to the, templates, you'll have bubble of the threat ready templates that are per specific threats. So it's a more tailored experience for your, different persona. We vulnerability validation, that's for customers. If if you are, an explosion management customer from Cymulate, that's basically what you have. So customers want to validate vulnerabilities and see, are they really exploitable or not? That's actually really relevant for for the for the miss for the customers who are scaling up mythos because they're looking at the vulnerabilities themselves and running simulations for them. They will have a focused view on the vulnerabilities that, the that that you have. And for the customer who are more detection engineers, we've also built a dedicated screen for them to validate their SIEM rules, and we also created a new package, for them. Both of the below packages, by the way, are add ons. So, if you're interested in in that type of value, contact your your, your account executive. But the two top ones are available for all customers. And if I'll go into down the detection engineering, we've noticed, as Brian said in the beginning, we detection engineering is one of the use cases that we've, that we're seeing customers use us for validating the same rules or the detection rules that they already built. And we once you know, we talk to a lot of customers who are using the using Cymulate to do that, and we believe that there's a better way for us to make that accessible for them just from selecting Cymulate from all of Cymulate scenarios and trying to understand which one is more relevant for the rules. Now we're giving a a a a a complete screen for the detection engineering use case. We call it detection studio, where we pull up all the detection rules from the customer. We map them automatically using AI, of course, to our, to our existing scenarios, where we say for that specific rule in your sim, that's the, that's the simulations that you need to run, and you can schedule those as well. So for every rule that you have, you can avoid rule rot and have those rule have those rules continuously validating. We have customers that says we have hundreds of rules or even more that we need to revalidate every quarter, and that's a really tedious job. With that, it really helps us streamline the whole, the whole flow. So that's Detection Studio. If I'm looking at environments, environments is is helpful for two use cases. One, for the CCEM customers who are more, vulnerability and they wanna attach the vulnerabilities to each to each environment. But also for the generic or for the for the more common, exposure validation use case, you also wanna create, an environment where you have the agents that is testing that environment. You wanna have the business context of that environment, the the priority of that, and maybe the associated assets integrating with your with the chasm, for example, to, to really understand what is protecting that environment. So, for example, when we will recommend you when Veera will recommend you on, which scenarios to run, it will understand, okay. These are the assets. These are the operating systems within that within, within the, the environment, and it will know to fine tune and that recommendation into the environment parameters. And we'll expand on that more in the future as well. We've, run a lot of simulations, and then you have the all the exposures or the findings, which is, kind of like a big database of all the different, explosion that you have. And we wanted to give an easier way for the fixer, the person who is taking those, all those exposures and wanna go and fix them, an easier way to handle those. So what we are doing is create we're creating a mitigation hub, which is taking all the exposures, and grouping them based on the environment, based on the control, based on the the the, the mitigation type. So it will be really easy for customers to say, this is one task with all the IOCs or all the EDR rules for that specific environment. I can go in either opening it in an ITSM or or push a a a a rule to to to to stop those in one go. For customers, it's really helpful for them because, instead of going one by one across all exposures, with one action, I can stop tens, even more of, open exposures. So that's the focus of the mitigation hub. And with auto mitigation, which is one of the add ons we have in Cymulate, up until now, we have IOCs where we would be able to push IOCs into the security control to stop common attacks. Now we're extending that into EDR rules. So when we'll have, exposures that we've seen that we haven't stopped them, and haven't prevented them, we will recommend we'll be able to push those EDR rules, automatically might, translate them into, Palo or CrowdStrike or SentinelOne. Those are coming with, this upcoming launch, and, might, translate it into the language of that vendor and push those, detection rules. Now pushing detection rules is not something you're gonna do just like, hey, let's push it in prevent mode. We're going to put them in detect mode or in alert mode. So we're not gonna stop, you know, cause problems in production. We're gonna see that, hey, we we we push that rule, there's detection on that, we're still not blocking. And only if you're getting for, let's say, weeks time, of detections and you see that there's no false positive, then just turn it on into prevent mode, and that will be, as set as done. And you already know that it only triggers on Cymulate simulations. So it's, it's good. So having all of those capabilities for us really helps complete that circle that Brian showed. How can we complete that circle for agentic defense engineering from understanding there's there's, a new threat that's relevant to validating it, to understanding what to solve, and to solve it completely. Kind of like that's our way to, wrap everything around in a bow. And, David, let's see it in Yeah. Absolutely. So first of all, that's. super exciting. I'm already, kind of excited to see it all in action with all of our clients. At the moment, all what I can do is, at least give you kind of a sneak peek of how it actually looks. So I'd like to cover, three main flows, with the remaining time and maybe leave some more time for Nir to talk about what's come what's coming next. We'll show how this Vero capability basically operates couple of microflows within that. Another one would be a seamless new detection studio, the detection engineering functionality. Lastly, if time permits, we will also take a sneak peek into the Mitigation Hub and the auto mitigation. Without further ado, allow me to real quick share my screen over here and we'll start with Devero. Many of you are already familiar with the Vero and what I would like to start is with a quite simple question over here. It is basically what is my thread profile? Nir briefly mentioned that behind the scene what Vero does, it builds for every client a threat profile. A common thing that our clients were struggling with over time is understand what do they need to focus on, what they need to test. If I'm a US banking organization, as I see over here, I'm able to see basically what is relevant to me based on my controls, based on the actual threat intelligence. Nevertheless, I might want to focus on a particular threat actor out of this profile. In my case, I'm asking the bearer to focus on muddy water techniques to test my EDR. What you will see is that this multi agent, I'll call it architecture, allows Nowvero to perform different types of actions derived from your natural language and basically guide it to what is the goal that you're trying to achieve in a very smart and intelligent way that is very contextualized to who you are, to your security controls, and to your threat intelligence. Now, the exciting thing about Varo is it will not only just go and create an assessment template for you as you might be familiar with the previous Ask AI capability that we had. It will now also tell you based on your previous assessments, how good are you doing against this particular threat. You can review the list of scenarios suggested by Vero or even choose them manually if you want to pick only a few of them. You can simply continue with the Vero Flow to actually run the entire validation assessment from just the same place where you all where it all actually started. Now, another thing that becomes a bit more interesting and we see across our clients, our clients deploy themselves across many different environments with many different, I would say, different test points. One of the things that Varo will also do is tell you what environments are a better fit for that particular case, for that particular test, for that assessment. You can see in my case that the default environment has the best match and 100% of the select scenarios will actually overlap with that. Moving forward, we will add more and more context to this environment, getting more context from your organization, from your business to really make meaningful recommendations. The next steps are a bit more operational about choosing what exactly you want to test. But the whole idea is that with Vero and also with our MCP and all these AI capabilities, you can complete a whole flow of actually interacting with the platform through a very simple AI native language chat where you can pick all the test points and eventually kick off an assessment. While we will be able to see the entire flow and see even the assessment launch through that, I would like to jump straight into the next flow with Vero. I'd like to jump to something that also very troubling to our clients. What mitigations should they prioritize? You launch a lot of assessments across many different security controls, but you don't necessarily know what to focus on. Today, Varo is able to basically do this analysis for you and give you a very detailed breakdown based on all the assessments you launched, what security controls you should focus on. Now, in our case, we can see that Varo prioritizes the e mail gateway and and the data exfiltration DLP. I want to ask it more. Show me the control weaknesses in more details. It will do another drill down into the control. Show actual percentages of penetration and prevention of the threats targeting these particular security controls. It takes you from just like another, I would say, very limited AI agent capability to a very broad multi agent capability that can give you a very detailed analysis, analyze your threat profile, and even take all that and use my next ask and create a dashboard for me that will represent all of my different security controls and their actual gaps in a very visual and executive ready format. While the dashboard is being created, I will just mention once again, if you have any questions, please feel free to use the chat, type in your questions, and Brian, Nir, and myself will be more than happy to go ahead and address all these questions. What we're seeing now is that our reporting agent, again, multi agent platform. The reporting agent now resolve the filters. In couple of moments, it will start placing all the widgets for you. Well, you don't need to do too much. You don't need to work manually in creating this dashboard and presenting your gaps. It will simply do it for you based on all the data that you have, again, in a very smart, advanced, and intelligent way. Finally, we will be able to get that dashboard and view it instantly over here. With that Vero demonstration, I'd like to jump into the next piece. After the Vero, another flow I'd like to show and it's something that we actually discussed together with Nir a couple of moments ago. It was both lenses and the detection engineering and detection studio. I'd like to show over here that once we jump into Summit's Lenses view, you can see all the dedicated lenses. As Nir mentioned, every lens like threat validation, for instance, will be able to adjust the view and the entire menu in the context of the platform to that particular use case, to that particular area of interest of the user, of the persona that uses the platform. Speaking of detection engineering, we actually introduced a new dashboard that is solely focused on detection engineering, showing you what is your actual detection breakdown, overall detection, and maybe even more insights that are relevant to that particular use case of validating your detections. But moreover, the more exciting thing, in my opinion, is the introduction of the detection engineering, the detection studio screen. What we see in here is basically the screen that allows you to see all of your detection rules. In my case, all the detections pulled from the Splunk is same within my environment. What it allows us to do, it allows us to focus on the entire list of rules that we have. We can see what's our coverage, basically how many scenarios are valid to test this particular rule. We can see what's our current detection ratio and have a very broad, perspective of all of our detections indicating more things such as the confidence score. And not only that, but we'll also tag all of your detections with what are MITRE ATT and CK, tactics and techniques that are associated with it, and also what is the actual severity of the threats that should be detected by this particular rule. More than that, if you click on one of the rules, you will see a more detailed breakdown including different details such as your rule ID, AI summary that will summarize what's the actual purpose of that detection rule so you can get more context before validating it What is the decision reasoning for why we think the correlated scenario that we'll see in a moment are actually matching that particular rule? For clients who actually do detection, validate their their detections, there are usually two approaches, validate if threats are detected, but also validate their rule health over time and ensure proper coverage and proper detection of their rules. Before having that, you didn't have one place, one studio to manage all of your detections and now you have. With a simple click of a button, you can see the correlated scenarios and immediately jump into the validation workbench and go ahead and create an assessment out of that. The last flow I'd like to demonstrate, I'll switch quickly for our unified view, is the Mitigation Hub. The last thing that I want to say is that what we saw from our clients that again, validating a lot of different things, but eventually want to be able to efficiently understand what you need to focus on in terms of optimization of your security control. Again, our goal is not only to prove that your security is performing well, but also improve that. The Mitigation Hub is basically a one stop shop for the fixers. That's the place where you can see all the different mitigations that you need to take, including the option to filter them, for instance, by IOC hashes. And you can see all the details of the mitigation that need to take place. Not only that, with the auto mitigation capability, you can very easily go ahead and mitigate that particular IOC within your environment. As mentioned, that capability is expanded also to EDR rules, so you will be able not only to filter the hash, but also to go and take a look at your EDR detection rules and also mitigate them too. I think a lot of exciting things are coming as you can see. I think on that note, I will turn it over back to Nir and I will say, Nir, maybe in the remaining time you'd like to share with us what's coming next or address maybe couple of questions. I see they are already addressed in the chat, so you should be good. Yeah. So let me just address the that one question a bit more elaborate on that a bit more. VerAI is available for our customers. We don't believe that there is an, Cymulate with AI or without AI. Cymulate is with AI in in house. You're gonna get that as part of the existing packages. There's no need for extra pay. And with the old Ask AI and the new Vero AI, it's except for the word AI that is there that both of them have, they're completely different infrastructure. The Ask AI was kinda like the early days of AI, which you kinda, like, had to learn what to do with various dedicated agents, And you have it was very, like, limited, and you had to make sure it's not, like, hallucinating. So now it's you don't have trick triggering for specific agents. It's one Vero AI, one like, assign you in in a way, but it's more than just ask. It will also provoke and ask you, ready made questions as well, and it will, try to understand what what is the intent behind the question and will activate internal, capabilities to do though. And I also see a question. There's, also how to prevent AI agents from executing unsafe and destructive actions. That is the Cymulate secret sauce, I will say. Because if like, I also got a question from, from a customer. Hey. There's this new mythos. AI can now do everything, so I can now develop everything with AI. Right? No. The whole point is not to, like, go go ahead and run AI, go and lose. We've trained AI to use only our capabilities in our infrastructure with our scenarios so it doesn't exit from the, the the the knowledge that it has. And that's why that's the Simonite advantage where it's always safe. It's always within the knowledge that we have. There's always the pre requirements and the cleanup, and, like, all those capabilities and frameworks are intact. So that's So, Nir Krumer, if I maybe summarize, how that yeah. we why you trust Cymulate and not some, you know, new, AI pen testing. Right? 100%. Cool. Let's let's take a look at what's next, and then, we'll we'll wrap up and, we'll take some questions at the end. Alright. Okay. So on what next? First of all, I had to put AI here as well. But, we talked about the MCP that we're exporting, that we now have an interface to. What we also want is to also give skills for, the different AI platforms that you have. So not only you would be able to interact with the tools that we have in our MCP, we'll all be we'll also help, be, we'll also wanna help train the AI mon that you're using or the AI platform that you're using, to, to interact with which tool to activate within our MCP. So the first two vendors that we're gonna choose is Cloud and Copilot. That's actually based on, the survey that we've made from our customers, that those are the most common one, being used. And from Vero, we're gonna expand on triggers to expand the capabilities of what we can do. We're gonna have more personalization to I'm more a CISO, and this is more what I want to have. I'm this is my role, and that's the that's the level of details I want you to share with me. So it will be much more tailored to what you need, and it will, provide you to do a lot of more automation capabilities, within the platform. Second step is on optimizing defenses. Once we wanna, you know, wanna optimize not just prove, but also improve our, our security. So ability to integrate with the ITSM, all the tickets that we had in, in the mitigation hub will be able to open and track bidirectional from your ITSM vendor, whether it's Jira or ServiceNow or whatnot. And also the auto mitigation is also going through an expansion. We've we've done it with EDR rules. We're gonna expand also on the vendors of that, but also on more and more, rules for WAF and SIM and and firewall and more. So that's on optimizing defenses. And the last one is on scale defense validation, which would help you, get the benefit of that platform more. The the first one is tailored threat intelligence, where we have the immediate threats, which is great and we get a lot of good feedback on. But customers are also saying, I wanna bubble that up and and decide and and see what is my threat view, maybe interact integrate with, other third parties. So we're going to, expand, our integration with or introduce our expand integration with CTI vendors as well, and kind of like bubble up the CTI interface within the platform. I will say if you have your preferred CTI vendor or Cyber Sight Intelligence vendor that you're working with, our team will be happy to know about it. So we'll be able to also, put that in our plans, as well. And the last one I wanted to talk about is, the track validation initiatives, which is something that we see, highly, from our customers that, okay, we have a lot of templates that we run, and we want to to connect them into initiatives that I'm that we have with the board. So, kind of like showing you from the R&D and d, the ability to see all the different templates that are, to track them basically as initiatives with their score and understand what is tested, what is partially tested, what's the schedule of those, maybe categorize them based on user templates or the best practices that you want. And once you go inside each and every one of them, you'll be able to basically track them as a corporate initiative. So seeing, for example, the trend of those, going into, with security control performance of each one, going through some insights for every for every assessment that is kinda like ready made and accessible, going through a summary. So all of those kind of like in a in a very nice way to track, a business initiative within the product. You're gonna see much more, kind of like droplets. It's not gonna be, like, in one go. You're gonna see, like, more and more, releases, taking all the templates of you and kinda, like, making them as an initiative, attract initiative within within your organization because that's where we're also talking to customers. A template is here to measure something across time, based on a corporate initiative. So we're gonna also address that in the upcoming, releases. Cool. And and wrap it up? Yeah. So, yeah. thanks for spending the last forty five minutes with us. We encourage you to explore what's new. So when this rolls out, next week, you know, there's a welcome guide. There's gonna be some in app messages. There'll be some you know, if if you explore new features, there'll be some guides there to to help you along. We invite you to try out Vero AI. Right? So, again, go build a custom assessment. So when you see a threat pop up of, hey. There's a guided threat, you know, check it out. You go go see how quickly you can build a custom assessment with your AI. There are play of resources for you. Go ahead. and and play around with it. Like, really, ask it questions. There's kind of, like, thumbs up, thumbs down. Use them. We we track those because we wanna see how you engage it. Eventually, AI is AI. It can be like it can have hallucinations and have a lot of different things. That way, we wanna see, your engagement and if you're seeing some stuff. So we're really tracking, if you're doing some up and dumb downs and and and why definitely, that will be helpful as well. We got the knowledge base updated. Again, the how to guides and such, product documentation, knowledge based articles for you. If you have questions, you you know, engage your customer success or account team. If you wanna deep dive on some features, you know, include you know, let's schedule some time. If you're interested in some some expanding capabilities around the new detection studio or looking at auto mitigation, the CTEM exposure management piece. Again, have that conversation with your account team, show that show off that capability and see what it means for you. We're rolling this out as a, you know, kind of a internal to our customers. You can look for more of a public announcement June 1. We'll be at the Gartner Security and Risk Summit, in the DC area, June. That's where we're gonna have some press releases and some blogs and kinda make more of a big deal about this on LinkedIn. So follow us on LinkedIn. Follow our news, for that. Okay. So, you guys have really stayed on a long time, the live. So we've got some, you know, you know, any other questions, you can get them in. We'll we'll we'll stay on for another three or four minutes to get any other, any questions that come in. Let me see that we have a lot a lot of talk about, a lot of talk about, you know, questions around around the AI piece, you know, just some questions around some guardrails, and we've got some good some great safety documentation we can we can include, make sure that customer gets that that that that, documentation around the safety and guardrails. You wanna talk about just briefly about the the model you're using and kinda, like, you know, you know, with the question around are we using our own or external LLM and just, you know, high levels the high level issue around what's what's that model look like? So high level, we're using, we're taking an existing model and we're tailoring it to our needs, so not reinventing a whole new model. There's much smart people that are building great models. There's no need to invent the models. There is a need to educate, on how does Cymulate work within the guardrails of Cymulate, how our scenarios work, what's a prereq, what's a cleanup, like, all that different knowledge on how do we work and what's our agents, and that's a lot of training that we do. And to understand intent and what to do with with the platform, that's a lot of guardrails that we, that we give. And, also, I see a question on autonomous remediation. Autonomous remediation is a nice one. I would say today is not so much as autonomous, more as automated because, we're not I I don't feel we're at the stage today where we can just, like, hey, give AI ability to do whatever it wants on the security controls. That's why on IOCs, it's pretty safe, so we have the ability to push. On EDR rules, it can be dangerous, so that's why we're doing it, in detect mode and not in prevent mode. So the again, we're pushing the rules just to make sure it's like there's no false positives, and then you have the ability to turn it on to prevent mode, but the rule is already there. So we're gonna put a lot of guardrails and gonna be a bit more monitoring on those, and here are the rules that we pushed into prevent mode. The ability to switch to to, to prevent within the platform. But, again, these are things that a man in man in the middle from a from a good guy perspective, where a human in the loop has to be involved, in really taking some things on the environment, which could be potentially dangerous. Cool. Alright. And what about, testing AI driven attacks? Are we I think we'll earlier this year, we introduced the LLM tester. Right? We have a full set of scenarios focused on, testing attacks against against an LLM, right, and and injection prompt and and such. Anything else around AI driven attacks you wanna talk about? So we'll have a lot of a we'll expand much more on, the AI, the AI attacks, not just creating AI attacks, but more on AI security, like LLM powered, applications, LLM copilots and assistant, and LLM agents. So we'll have a lot more scenarios on, on those. That's something that is really coming up, whether it's on the WAF, on the ASSISTANT, on digital twins. So, we'll have a lot of assessments on those as well. Alright. Okay. So it's, it's our audience here stayed with us a long time. I just wanna say thank you to our customers for, staying with us here. This will be an archive and and emailed out to everybody who registers. We're gonna get this in the knowledge base as well as an archive. We wanna thank you for for attending today, and, you know, hopefully, you've seen the excitement that we have for what's coming out. And, we ask for you to, again, get in the product, play around. Let us give some feedback. Give us a thumbs up, thumbs down, what's working for you, and we, and we value your input. So, Nir, David, thanks for your time. And customers, thank you. Have a good day. you, guys. Thank you.