ExpertCast By ActualTech Media Securing Endpoints Through Effective Management, Solutions and Strategies

  • Future B2B
  • Automox
Securing Endpoints Through Effective Management, Solutions and Strategies

About This ExpertCast

Managing and protecting remote laptops, mobile devices, and server endpoints has become increasingly complex as threat vectors expand and hybrid work models persist. Led by a veteran cybersecurity authority, this ExpertCast delivers an actionable overview of modern technologies and operational frameworks designed to harden enterprise endpoints.

Participants will explore cutting-edge solutions for automated patch deployment, continuous threat detection, and centralized policy enforcement across distributed device fleets. The session examines practical methodologies for neutralizing zero-day vulnerabilities, streamlining vulnerability assessments, and reducing device maintenance overhead.

By focusing on practical tool selection and efficient management techniques rather than theoretical concepts, this presentation provides technical leaders with the insights needed to evaluate emerging endpoint protection platforms. Equip your organization with proven strategies to simplify device management, strengthen endpoint resilience, and eliminate critical security blind spots across your network.

Why You Should Join:

  1. Discover cutting-edge platforms to streamline distributed endpoint device management
  2. Evaluate innovative software solutions that automate threat detection and patching
  3. Identify advanced management tools that eliminate critical network vulnerability blind spots

Agenda & Speakers

  1. Keith Ward
    Keith Ward Webinar Moderator Future B2B
  2. Nick Cavalancia
    Nick Cavalancia 4-time Microsoft MVP Future B2B
Full Transcript Keith Ward Hello, everyone. Happy Friday, and welcome to securing endpoints through effective management solutions and strategies. Today's webinar is sponsored by Automox and produced by Actual Tech Media. For those of you who don't know, I am Keith. I am a moderator with Actual Tech Media, and we are a future b to b company. I am totally stoked to be here with you today. This is going to be a fabulous event. I promise. And I'm gonna start out by covering some housekeeping items you need to know about. We're going to begin our little micro tour as always in the chat tab on the far right hand side of your screen. This is our virtual Starbucks where we hang out during the event or virtual Pete's Coffee. Pick your your coffee shop of choice. Now this is where we connect with each other. We talk. We chat. We send emojis back and forth, thoughts, stories, anything you are excited about and wanna get out there. One of the best things we do is get to know our community a little bit better. Say hi. Let us know, how you are doing and where you are. I am gonna start right now by saying hi to Jordan who is in Colorado Springs, Colorado. Jimmy Arcade, my favorite name is in SoCal. Let's see. Steven says hola. Gang says hello, and Morteza is in California as well. Adam, hello there down in Atlanta. And let's see. Fahid is from SmartVision. I assume that's a company. Emily, you are pretty close to me over there in Columbia, Maryland. Good to see you. Brian in Porterville, California. Landis is in Bethlehem. I assume that's Pennsylvania and not the other one. Gilbert is in Irmo, South Carolina. I love seeing all you fine folks here. Juan is in Puerto Rico. Tom says hi. Andrew is in California. California is showing out today. So that is one of the things we do there. Those of you who haven't told us who you are and where you are, do that now. Again, let's get to know the community. Now the other thing you do with chat is you get help with any technical issues that you might be having. So if anything goes sideways, if you can't hear something, you can't see something, chat is where you go. First thing you wanna do is give your browser a refresh or two. Normally, that'll fix the problem. If that doesn't take care of it, my spectacular colleague, Irina, is standing by to help you out. So put that in the chat if something goes badly wrong. Alright. Now let's go two doors down to the right from chat, and you will see the letters q and a. And strangely enough, that stands for questions and answers. So that is where you put questions for our expert today, and hopefully, you'll get some answers as well. So we try not to put questions in the chat. We try to keep them in the q and a just to kind of keep them a little more ordered, so we know where all of that is. If you can keep those separate, that would be great. In the middle, you will see the word docs. That is where we've got resources ready to go. Lots of great free stuff for you to check out while the event goes on. And we are giving away more than just fabulous content. One of you is gonna walk out of here $250 in Best Buy cash richer courtesy of a gift card I am giving away at the end of the webinar. Now you do have to be in attendance during the entire live event to qualify for the prize. And in case you're wondering, yes, we do check. All winners must also meet our terms and conditions, which are in the docs tab. Scroll down to find them. It's incredibly boring reading, but if you're a bit of an insomniac, it might help you out. Finally, you can set the language and closed captioning in the lower right hand part of your screen if you need those functions. We have got multiple languages available. Please take advantage if English is not your native tongue. Okay. Now as you may or may not know, ExpertCast are product neutral looks at technology, but our sponsors make it possible for us to bring you this content. So our sponsor for today is Automox, and now it is time for a few words from them as the old time announcers used to say. Now for a few words from our sponsor. Cyber threats move at AI speed. Your response should too. Automox closes the distance between vulnerability discovery and action so IT and security teams can patch what's patchable and mitigate what isn't. With AI powered automated vulnerability mitigation worklets, Automox turns emerging threats into ready to use engineer reviewed automations even when no patch exists yet. How cool is that? That means faster mitigation, less manual work, and every Windows, Mac OS, and Linux endpoint hardened by policy every day. Patch, mitigate, remediate at the speed of AI. To learn more or get a demo, visit automox.com. And with that out of the way, it is time for the main event. I am completely thrilled to introduce you to the man who puts the expert in ExpertCast, Nick Cavalancia, a four times Microsoft MVP and CEO of whatever that company is that he's gonna tell you about. It has gone right out of my head. We're gonna bring him on now so he can tell you exactly… Conversational geek, that is your company. Right, Nick? I came up with… Nick Cavalancia Yes. Keith Ward You know, Nick Cavalancia You got. it. Well done. It came here. Keith Ward Nick, it is absolutely so exciting to have you on today. For those of you who don't know, Nick is a veteran speaker, and he is highly entertaining. So I am completely looking forward to getting educated by him today. Nick, I'm gonna see you on the other side for q and a. So folks, as Nick is doing his presentation, please remember to be asking questions because when it's all done, he and I are gonna be going over your best questions. Nick, I'm gonna turn this over to you. Please take the floor. Nick Cavalancia I appreciate it. Thank you very much, Keith. Everybody, glad that you're here today. I'm just gonna go ahead and make sure I get my screen shared out here so we can get things started. This topic is incredibly important. And by the way, Keith, I'm gonna bring you everywhere with me so you can introduce me when I walk into every room. I think that's better than Elvis has left the building is when you introduced me coming in. So you you got… You're hired. If you ever wanna retire from what you're doing, you can come just live in my house, I guess, and and announce me everywhere I go. But, I'm excited about today's topic because there's one aspect of cybersecurity that I I find when I'm talking to customers, particularly when it comes to endpoints. And the thought process is a lot more around things like I wanna detect when there's a threat rather than try to preventively mitigate the threat in the first place. That happens a lot through patching, through configuration management, things like that of endpoints, probably because the former is a lot easier to accomplish. You put some EDR software on an endpoint, and it has a whole list of either threats that it knows about Or it's got AI. Based detection or something along those lines because I I found something that's anomalous. We should do something about it. That's a whole lot more easy than it is to go and figure out how do I make sure that everything that's on the endpoint is properly secured in in a timely fashion as proactively as is possible, meaning before the threat ever even hits us. That… That's a lot more work. And so it's really important to have this conversation so that we can be thinking about how how do we actually secure our endpoints in a way that, that adds another layer so that something like, let's say, detection and response and so on, they have less work to do. The… With their… And that's the whole goal here is to put as many layers in the way of a threat actor and their actions as is possible, making it difficult for them to be able to accomplish their evil tasks, if you will. So, what what I wanna talk about today, there's kind of three questions that I'm gonna try to answer today. This is not an agenda slide. There's just the kind of things that I want to talk about. It's what what should you be looking for when it comes to automation when we're thinking about trying to secure endpoints from that proactive, perspective? How do you close vulnerability gaps that… When attackers use them? And we're gonna talk a little bit about what data sources you should be using so you actually get a sense of whether or not vulnerabilities are actively being used or not as opposed to, let's just say, one that's deemed critical because of the nature of a vulnerability, but is it maybe not necessarily being utilized or exploited just yet? And then what kinds of platforms exist? What should you be looking for in a platform that's actually going to secure a distributed fleet of desktops, servers, you know, tablets, whatever that… Whatever it is that you have for endpoints and whatever they happen to the line. So, first thing I like to do is… I don't wanna use this through. I like to use industry data as much as I can because it helps paint a picture of what's going on out there. Now I wanna preface this by saying, the goal isn't fun. I'm not trying to go, it's terrible. You've gotta do some… I I don't have any skin in that game of trying to make you be, you know, fearful or have uncertainty and doubt and so on. But I do like using industry data because, again, it it tells you what other organizations are seeing that maybe you're not. So So you can start going, think we should prepare for that and and that kind of thing. So throughout the presentation, I've got lots and lots of industry data. I've even borrowed some from Automox themselves and a number of other, industry sources here. And you will see at the bottom of every slide whenever I've got some industry data, I've at least got the name of the vendor or the report or whatever it is, the source that it comes from. So you know that it's not just, you know, like 87.6% of all stats coming up on the spot. None of this is made up on the spot. That's the point. So let's start with where the endpoint being the the starting point of a breach. And let's start with that fact here by looking at a couple pieces of industry data. And the first one here comes from us from the SANS report. We can see here that 86% of street operation centers have some kind of response to endpoint security alerts. It means there's a preponderance of endpoint related issues. Now this doesn't necessarily mean vulnerabilities. It could be that someone has already gotten infected the machine. It could be something like we're responding to a, someone who marked an email being phishing up on an endpoint. It just depends on whatever that SOC is using and what abilities they have to pull data in. But in essence, I'm trying to point to the fact that the endpoint is a material issue when it comes to security, and we have it demonstrated by almost every SOC has to respond to endpoint related security alerts. Then we've got from Sophos here, you see about a third or less than a third here, of exploits and credentialed and ransomware attacks start on a new device. And that just makes sense. They're trying to get a a foothold within an organization even if it's a logical foothold of someone's endpoint that happens to be at their house, and is it utilizing cloud based services, no VPN, no corporate network, it doesn't matter. None of that matters anymore. It's still the idea. They've got to get a foothold in the organization, and that initial attack factor in at least 30%, if not more, is always gonna be a user device. Then this is an interesting stat that kind of makes the case by looking after, kind of, after infection, after an attack. And you can see half of corporate users have had an info sealer in, in, infection, excuse me, on their PC, on their device. I shouldn't say PC anymore. They're not PC's necessary, but they're done their device. And that's showing you that fifty percent of people have had a… Some kind of attack on their own device successful of some… To some degree, an info stealer being, you know, the piece of malware that's gonna go through, data on the machine and they go through browser, router data, try and pull passwords, you know, all of those things and take that data and exfiltrate it. And so that tells you right off the bat again, the endpoint is a target. So we've seen so far on the left, SOCs are verifying the fact that, yep, we're getting lots and lots of endpoint related security events. We see that exploits and credential led, ransomware attacks are all utilizing endpoints as an initial attack vector. Half of users are actually experiencing attacks with info stealers. You can see here, that 46% of those info stealer attacks, the devices that they attack held corporate logons, you know, and on them and they're unmanaged. And so that becomes a real problem. And this is sort of the crux of where the rest of this presentation comes out is. If they're unmanaged, if they're not secured, but with by patching, by proper configuration, that sort of thing, then you get that one and two to the left of the question mark stat here of Infosys successfully gaining access to, endpoints. And then the last one here, you just see here, it's, 30% of breaches and this comes from, it looks like, Verizon's data. Here's the breaches start with an exploded vulnerability, and it came up from 20% the previous year. So we're seeing an increase in this, ability. And it's one of the reasons it is because it's becoming a lot easier to find certain AI. I'm I'm guessing none of you been hiding under a rock, and you heard when Anthropic released Mythos a number of months ago, and they they found, what was it, 27,000 exploits or something maybe? It was even more. It's tens of thousands anyways of vulnerabilities that had never been discovered before. Some are as old as, like, seventeen years, if I remember reading correctly. And so that's one of the reasons why those vulnerabilities exist. It's a question of finding them. And so that's why we had this. So the point we're making here, end point is where the breach is starting. Now there's kind of two clocks that exist here that are not in sync. There's the clock that the threat actors use. It's less than seven days as the mean time to exploit, and you can see twenty four hours is how quickly the… A zero day can be mass exploited at this point. It's probably less than twenty four hours, but that's a that's a good look…. Looking at seven days and one day. That's their clock. That's the speed at which the threat actors are running. Now in contrast, an organization's clock, not necessarily your specific clock, but an organization's clock. This is data that comes from, you can see Automox here. 0% of organizations reported there, there are twenty twenty six. It's… They call it twenty twenty seven reported looking forward, but it was just put out this year not too long ago. 0% of organizations can actually manage and deploy an emergency pass within hours, and almost all of them need at least three days. So we've got… Just stopping there for a second. Seven days, mean time to exploit, less than twenty four hours, it's it's, possible. And we're at with us as organizations. We have no ability to respond. In in zero days, if you will, it takes us three. And then the median, number of days is forty three days to remediate a vulnerability. So think about that. It just… Just even the dwell time alone. If that… If an order… If product is forty three days of dwell time within an organization, think of all the things that they can discover, the lateral movement they can do, the data that they can access, the permissions that they can gain, control over, the data they can exfiltrate, all of that within forty three days, particularly now that almost every ad… That is now just all… Fully automated, if not leveraging AI to make it autonomous as well. So we have a problem here. This means that response and doing the idea of responding to exploits, in a very reactive matter is not what's actually gonna be securing your environment. We've been thinking about this more proactively. There's also a stat here. I'm gonna bring this up once again later on during the presentation that you can see at the bottom here that, 58% of known loaner values, are sort of only partially remediated. And there's actually… I'll dig into that a little bit deeper. There's actually a worse stat about… That are never remediated, but I'll talk about that a little later on. So keep going. Why does it keep getting harder here? We all know you hear the things of, like, well, they're leveraging AI and seed of attacks and sophisticated. Yes. It's all very true, but it's like, why? And there's a couple things that have changed over the years. And I'm not gonna tell you anything necessarily new on this slide, but this is our opportunity together to take thirty, forty minutes and have you just focus on the problem and think about it. Think about all the nuance, all the facets of the problem so you can start figuring out when I get to some practical guidance, you go, okay. How would I implement that in my organization? That's what I'm trying to do. So some of this, maybe a lot of this, will not be earth shattering. Oh my gosh. I had no idea. But I wanna make sure it gets covered, so we walk through the logic together. So the first thing is your fleet of devices is not even… It may be even in a building at this point. There are some of you on this webcast today that don't even have an office anymore. Everything is purely 100% remote. And so the old idea of needing to connect back to a corporate network or existing within the corporate network may not exist, or it's very, very light or getting lighter as the days go on. Second one here is you have a need to support every operating system. I can remember the days… I go way back. I'm I'm in the in the tech space for a number of decades, and we had just had Windows. And it was like Windows n tied to Mac. You know, that was Windows and Mac. And now it's Windows and Mac and iOS and Android and, and Unix and Linux and all the different variations. So it's with every operating system and then many, many, many apps on top of that. And those applications, of course, have to be patched as well, and that becomes part of the problem of the third party software. And then the last one here is this idea here of AI finding the flaws actually that we can fix them. And that mythos story is one that, again, I'm assuming all of you have at least heard, if not know pretty well. And and so that's… It it it makes the case that, AI is gonna find a vulnerability much more quickly these days than it used to be. I remember, man, it was probably not more than, like, six years ago, maybe just for the pandemic. It was, like, a million dollar bounty that ransomware threat actors, because they're making so much money. They were doing vulnerability bounty. If you could give them a good vulnerability, they could take advantage of it. They pay a million dollars for that. And now you can just spin up AI to do it for you, yourself, in a couple of hours. My other world has changed. So then there's this concern that I have about automation. Automation's obviously part of the answer here. I would need to automate the securing of endpoints, automate the securing of configurations, and so on. And the paradox that exists is you're gonna see here… I mean, I'm gonna put up a couple sets of stats here. Is the team saying we're doing incredible. We're amazing. We're rocking, man. We're… This is great. But then I'm gonna show you what the numbers actually say what's going on, and we're gonna talk about why. So, the team say 81% of them say, input management, mostly or highly automated. And this again comes through Automox themselves, sponsored today's webcast. And so in this case, the team just saying, no. We've got this. We're just using automation. It's mostly being used or all being used, and everything happens on a magnet to anything, yeah, so to speak. And then 69% of them said their endpoint operations, they call it very mature. Now, typically, in any kind of maturity model, a level of maturity, let alone very mature, but a level of maturity usually means and indicates that you have a dynamic program going on. In program, I mean, the… Let's call it the endpoint security maturity model. The endpoint security program that the… We have methodologies, the processes, the people, the… Everything you have in play that secures the endpoint if it were a program that would be having a level of maturity. If one was to say it's mature, I would be assuming that it's completely dynamic and you don't have to do anything. It's gonna constantly be changing. It's running continuously, all of that. And so now I've got this very mature, which obviously is Automox's choice to put very mature versus mature. But I just love that they said very mature, because they're super confident about how awesome they are. But then the numbers from the same report show everybody spends ten hours or more responding to endpoint related issues. That doesn't sound very mature or completely autonomous or automated to me. I don't think it does to you either. And, again, almost everybody, they've got a mean time to patch of five days or more that is not very mature. Sorry. It just isn't. And this is not knocking any of you. It… It's it's tough. You're putting out fires every day. Your focus is constantly being pulled. It's tough to get these strategic initiatives done to a point where you can just put a pin in it and go with good. And so we're gonna talk about kinda how to try to accomplish that both from a process perspective as well as a solution perspective by the end of it. So this will not all just be… It's terrible. The sky is falling. I really wanna give you some practical advice. But I think we can agree that there's this paradox that exists. We know that automation is gonna be necessary in order to take advantage… Or some take advantage but address all the issues that have been outlined beforehand, where the endpoint has to be secured proactively. We can't just reactively secure it. But automation is gonna be the only way to do it, and yet there's this paradox. And there's a couple of problems. The first one is there's too many tools. You can see here, this can come from Automox. You can see that the majority of organizations are using, like, five different endpoint management tools. That'd be fair. I'm gonna assume because I don't have the report in front of me nor I think it probably can send it in the report. But endpoint management doesn't necessarily mean just tools that patch and or do configuration security or… It… Those kind of… It could be anything else. It could be an EDR or it could be something that puts out at Just, like, deployment of applications, like group policy, technically, sort of an endpoint management solution and stuff. But, I mean, it can do that kind of stuff. Right? But at the same time, if nothing else, I think we can all agree, we probably all have a a number of solutions that are doing something to manage some aspect of the endpoint. And according to their data, the majority of organizations have five or more, and then a little more than half have 10 or more. So that's one of the problems. And there's a couple of, outcomes from this. I mean, it's an impact that that flows from us. We see performance. There are portion of organizations can't even tell if it's driving the right outcomes that they need. We have over two thirds of organizations can't even measure ROI. And then a little over half saying reporting is actually very, you know, to be difficult. It would deliver the most value because it needs better report. So we have this problem here because we're using so many different solutions. We can't get one consolidated answer, and that's one of the problems here. The second problem is that the thought is that the gap is the automation itself. Like, if we could just automate this properly… Keeping in mind, really tough to automate if you have five different solutions because you can't just do one automation, if you will. And I know we we wouldn't just do one thing with one master automation. But in essence, constantly, we need to do one platform. There's five different platforms. Automation is very difficult. But the other one here is… And this again comes from all of us data. Only 9% of organizations are actually looking at automation as the basis for how they do things. Meaning, we start with the assumption we're going to automate this, not… You know, could we use automation to do this? That's different. That's the other 91%, but you can see that on the bottom there, bottom middle. They're using it as a tool when it's convenient or they come up with it. 9% are the old… Other people that are going, when we do x, it's gonna be automated. That's the default. And so the gap isn't the automation. Even if you have five solutions and let's say they have some automation, the gap still isn't the automation. It's probably the processes, the procedures that we're using, the standards. And there's there's some real, issue behind that. And, what I wanna do, I wanna talk about why and some of the trust factor in another slide. But before we do that, I wanna talk about sort of what's changing. So we're level setting, you know, this automation. It isn't just about automation because that's not what this entire session is about. But it's about if you were gonna do that, there's a a much grander sort of perspective you have to have on how that's going to play a role overall in your impact management. So I have this sort of yesterday and today, and truth be told, some of the yesterdays are like really really yesterday, a long time ago. Maybe it's yesteryear, yester decade. I don't know. I don't know. Make up words. It's a little long. I could do that because I'm the one speaking. But there's there's some… You see, like, yeah. That's old news. We don't do that anymore. Really, today and tomorrow, and that's fine today and tomorrow. So if you fall into any of the yesterdays, then you know you're definitely… What you're doing is not current. I'm hoping that most of you are somewhere in the today, if not tomorrow. I've got a smoothie, so pardon me. Sorry. Let's… Oh, here we go. Party. Sorry about that. Can't can't, help to be human. So the the first yesterday is thinking about, patch cycles. I put monthly. But does that say monthly, weekly, quarterly? It doesn't matter, but it's cycles of patching. That's the old thinking. The current thinking has to be it's continuous. It's policy driven. It's responsive, not in a reactive way, but responsive to, there's a a brand new CVE for this particular threat, this vulnerability. We need to respond to it now and and then coming up with that. And that's not thinking about it in terms of, like, a monthly cycle but continuous. Then the thing I've already kinda talked about, I'm gonna guess that most of you have a very distributed workforce, and the old line of thinking is it's not so much a VPN as much as it is that everyone connects back to a corporate network. And then… Because the thinking there, of course, then is we can run kind of an on prem focused solution. And then since everyone's VPNing in their logic, they honor corporate network, they're technically on prem, so to speak, we can push out and manage them. It has to be completely cloud native, works anywhere, anytime, doesn't matter if we're connected or not. That's important here. The OS patches… That's my yester decade methodology there that fits in. It's, of course, gotta be OS and everything else on the machine. If anybody falls in that category, application vulnerabilities are are giving, threat actors access in the same way that OS is. And that's… One of the reason this is actually so really critical as you think about this is just stepping back for a moment. Threat actors have a limited scope within which they can work. They have to work within the context of… Let's just pick one operating system for the sake of argument. Let's say just Windows. Yes. I know Mac. I know that… Then I I got it. I'm not I'm not saying those shouldn't be used or putting them down and picking one to make the conversation a little easier. If you just have Windows, they have to work within everything Windows can do and everything Windows will allow, and that's it. And so at some point, they might not find another vulnerability for the sake of argument… For the sake of argument within the operating system. So they're gonna go towards the applications that are supported by the application. Maybe I can do some kind of a direct memory injection through an application or something like that. Because the application doesn't have any security and the way that it's configured with the operating system, it can actually do direct memory injection of something because maybe it would do something like that itself to store something in memory. I mean, that that said, yes. So since code actors are limited in what they they can use, they're gonna try to expand that as much as possible which means applications come into play as well. An old kind of thinking which goes back to the concept of the monthly time cycle, which is you do a scan of a bunch of systems, you open up tickets when has to be passed, you wait until the monthly cycle and you do it. That's not it anymore. It's gonna be… I could find a fix and verify it to get it done. It's gonna be the moment you find it with a spot immediately, and that's the idea of continuous. Continuous, typically, when we say continuous, you know, we're all talking about, like, it's a process. It's constantly happening every day. Here's how my continuous from the the standpoint of, like, it's it's… In a timely sense, it's continuous. It's happening now, and five minutes from now, and ten minutes from now, and an hour from now, and at three in the morning, and four in the morning, and it's continuous. So it's happening. Yesterday's thinking is, you know, one off scripts, things like that. Now we're trying to do something that is shared automation. We're driving towards AI based autonomy as well with the Adjentic AI. And then the other part here is instead of waiting for the patch, we're trying to mitigate now so we can patch later. The mitigation could be a configuration change in an operating system potentially. Could be maybe doing… Maybe change in your virtual LAN configuration if someone was on prem maybe, maybe isolating an endpoint or an application or a sort of traffic or firewall. I mean, something. In, in the world of compliance, you'll have compliance mandates, and then sometimes organizations can't actually meet that compliance standard or whatever it is. And so they do something that's called a compensating control, which means I can't do that exact thing, but I can put all these controls around it or control around it, one or more, that are gonna ensure that at the end of the day, that thing actually does happen. But I can't just, like, check a box and it happens. I have to do seven things to make sure that happens or whatever. And the idea of mitigating instead of just going, well, there's not a vendor patch. I can't do much. You're gonna put a compensating control in place until there is a patch. Until such time as you can meet that need, you're gonna figure something around it. And then the last thing here is, obviously, like, spreadsheet. So I I I don't… I'm hoping nobody does that anymore, but I'm gonna think there's one of you somewhere in the audience that still does. Can't do that. This has to be live. You gotta go back to being continuous. It has to be something that's already approved, and you can look up and see the current state. Current meaning, like, right now, the current state of the order of magnitude. So now, the the practical stuff that I wanna give you here is, an operating framework. And I've got something like, I guess, about ten or fifteen minutes left along those lines. And Keith, you can know or give me a a number if you wouldn't mind just how much longer I have, and I'll double check. But I've got this kind of into a six part framework that I'm gonna cover over the next couple of slides. But, basically, it's know what you have to see. You have to prioritize what you need to patch, what you need to harden the configuration on, and so on. Third is actually automated patching. Fourth is enforcing the hardware configurations. Five is responding to a zero day and either mitigating it or patching it. And then this is proving that you've actually done this in state of the organizations and point security is harder than it was before and so on. So I'm gonna run through this just really quickly here. The first one here is just knowing every endpoint, and the first point here is just you have a a live inventory of what you have. You wanna know the devices, you wanna know the applications, and so on. And then there's also… With that inventory is the second checkbox there, which is a… An assessment. There's a continuous assessment, which goes back to what we don't think of, like, patches and vulnerabilities tied together as inventory. I sort of am in my head. It's like an inventory of what are the vulnerabilities that are out there that we found and what do we have to address. And so there's a continuous assessment done because the inventory itself, the physical inventory of hardware and software doesn't change as often as there are suddenly 600 new CDEs today. That's a very faster pace. And then the last part that's kind of important here is having an ability to have what I call a drift awareness. Meaning you're aware specifically with configurations, should someone drift away from that configuration so you can reinforce the secure configuration. Could be enabling JavaScript and PDFs for the sake of argument. Just something simple like that. That's bad. We wanna reinforce that. So we don't know if that happens. We don't know if the configuration drifts away. And then the last part here is I want you to think of this as kind of a detection. Typically, when we think of detection, we're probably thinking of, like, detecting with a… An EDR solution on an endpoint or we're detecting, an intrusion of the… Through maybe it's… Someone's accessing data they shouldn't be accessing or they're moving loudly with with the credentials or seeing what the credentials we may use and so on. We're thinking about actively a threat that's happening right now. But here, it means exposure. We're trying to detect. That's why I use the term inventory even for, like, missing classes and vulnerabilities. We're detecting that. We're detecting their exposure and their risk, and we're trying to do this proactively. So not reactive detection. We're talking about proactive detection. Second one is prioritize. We want to figure out, what's being exploited. Now if you're… Layer on the right of this, underneath this first… The the CVSS scores, those are scores that are telling you what the potential is for this particular vulnerability. It's it's critical because if this could happen, the, the thesis, the… I'm going blank. Knowledge based exploit vulnerable. I'm going, like, totally blank in my acronyms. It… It's their database of of of. I can't remember what it stands for. Somebody put that in the comments for me and save me. But, anyways, that actually tells you if it's actively being exploited. And that's different from a… Like like, the CVSS score, still good, but it's from academic in nature. The the the data is telling you right now it's actively being exploited, and then you go, oh, it's practical. It's application. I need to know that somehow I have to secure. And then you're weighing what that exposure… Excuse me, and the risk to your organization, and then you're trying to figure out how do you… You'll… Can you find a fix quickly? And that's what you're prioritizing. Then the idea here is I talked early on about that 58% of the moment was that that took a long time and got partially remediated, 60% never get remediated at all. Remember that. So there's some percentage that did nothing with that for most organizations. The third part of this, the patchable. We're thinking about this in rings. And by rings, I mean, sets of devices that could be based off the ones that are remote or the ones in the office or certain departments or whatever, and you're putting some kind of a pilot group first together or even a broader rollout. So you're always putting some kind of deadline with this whole thing through your patching. And, and then… Oops. I I went ahead and I didn't mean to. Let me go forward or back. There we go. You're thinking about the operating system and the applications, and then you also, of course, have to figure out ways to do this in a continuous fashion because you make a scenario when the sky is falling, when the sky is falling, and you have to be continuous all the time today in five minutes, ten minutes, and an hour. At the same time, you have… You just need to get work done, and you have to figure out how to balance that as well. And then this is one of the stats, I had put out as well earlier just to reiterate it. Ninety four percent of organizations are taking five or more days. So this is not happening for most of you. You're not doing stuff where you're trying to figure out, a a a time frame within which to remediate, and figuring that that comes in. Yeah. The last one here is… Or sorry. The last one here before we talk about this five or six is the third of the last one, is the enforcing of a a baseline configuration. We're thinking about defining a baseline. Certainly, there's a… What I'm gonna call it, baseline. Meaning, it's for everybody. But then there is maybe by role or by operating system or by department or by risk factor or something, and you're… No matter where that device is, you're gonna put put a policy based baseline and press that out, and then you make sure that there's that drift correction should they change something, going back to my example of, like, JavaScript enabled on a PDF or something. And then you see here, 40% of organizations are saying that the ongoing validation they could get, the configurations without… That being aligned would actually deliver value. So they're seeing the value, and that comes from AutoBox, their other report this year. But it's just basically demonstrating you that organizations do see that value, and I think you would as well if you knew. No one's tripping from the configuration. We know that it stays secure. We're continuously evaluating it and validating it and so on. Five is respond, and this is what you might look like for a zero day playbook. You wanna mitigate within hours. That means maybe a compensated control if you don't have a patch yet. You wanna verify that that super mitigation actually ran once they're the release of a patch of some time. You wanna make sure that it runs, and you see it done in the text there, an accelerated clock. And the last one there is a new cleanup. Any kind of workarounds, someone, that that couldn't get fixed on the unreleased, whatever the case is, but then also figure out how do we ensure this never happens again, and that it's part of now maybe the new baseline configuration for images for machines or whatever the case is. And then this goes back again to something I mentioned earlier, zero to eight. Service center of organizations can do it up and emergency patch in the balance. And so tell me this has to be happening so for you. And then, with this, you're gonna use automation, but there's talk about it early on. There's some reasons why organizations are automating really well. Couple of them are, you're worried about privacy, worried about security, worried about data governance, worried about transparency. Can you actually see what's going on with visibility? Are you… Do you understand what the automated decisions were made? These are all really real issues. And so most organizations, when I think about automation, it's the very same conversation I have when I'm talking about doing agentic AI and doing work autonomously within IT operations. You're gonna have to build some trust. The first thing is just do notifications. Go out and do this job. Come back. Tell me what you found. Right? And then the next one is, I'll tell you what I found, and I think you should do action, and you go. I approve… You become a supervisor. And then the last one is gonna be you're gonna put some limits around what it can do and anything outside of that stays a little approved, but you start automating. And you make it… And then we automate to be synonymous here with autonomous, meaning that you're gonna allow the platform, whether it's using AI or not or just through scripting or whatever, you're gonna let it work on a standard zone without being there. And, and then you can see this another standard from Automox here that GuardRash is… Would need that resistance automation. So the GuardRash will be the thing that are gonna help you. Alright. Step six. You wanna have four numbers to report. Mean time to patch. You start looking at that and start seeing how quickly you can get that down. Now we all know that the mean time to patch is something you can't always control. You can control your response. So I think the mean time the patch has to be from when it's released to when it's installed, not from when you detected in the thread or the CVE came out or so on to when it got patched. That… It's just when the patch is released, how long. You can you can control that. You can't control how long to a patch. Patch coverage, which is based to what percent of the organization is covered or not. Exposure window, and this is, this this is one where you're actually showing where you're not good. And it's important to do that. It's great to go. Our meantime, the patch is down 48% over last quarter. Sounds amazing. But what if only 3% of your organization's patch? Yeah. To the same argument. So if you really wanna have numbers report, you have to be a little vulnerable. That's the third one is. What's the exposure window? And then the hours are claimed through automation. And that one's really important when you're talking to, like, a board or something or a c suite, and they're trying to understand why this is important, why you need to invest more in automated solutions, and so on. And then, and then this one here is saying that the, the endpoint know how, 84%, again, both Monomox here. The know how round, the endpoint is only… It's somewhat distributed. And so that puts the organization at risk because everybody doesn't know how to do everything. It's another thing to make sure that with the data, you're gonna demonstrate you are actually getting better because the data points are gonna have to say that not everybody understands how to do this. And so if we're gonna actually get better, we have to be measuring ourselves. And then also here is another problem that organizations have. Making the case of why you need these numbers is 90% of organizations need three months just to prove that it's an actual repeatable workflow. And by doing and having some, metrics for yourself, that's gonna drive you towards repeatable workflows. And repeatable workflows give you, you know, the the the repeatableness, if you will, give you predictability, and predictability gives you the options that you want. Alright. So the last thing here, this is the six again. I just wanna make sure you see little arrows popping up when you think about this as a process. When you go through all these and you prove at the end the outcomes, you're also learning. Because remember you had that one the one that was on… During your your metrics, you're looking at how vulnerable you are and what the exposure looks like. That's what's gonna help you drive. What do I have to go back and see? I didn't see it last time. We're gonna improve this. We go on. So then the last thing that I've got here before just a couple of quick steps, I think I've got five minutes left when I'm done, is what to look for in a in a platform here. You wanna make sure it's cloud native on prem with a VPN. Just… I I think you can all agree that doesn't doesn't work because that everyone fits that mold anymore. Has to cover every single operating system we have, serve as well, no matter what what platform they are. You'd have a lot of depth around third party applications. It needs to have… This is important, extensible automation. That's an automation. Like, we've got these 57 scripts or these 57 things you could do. That's great. Every organization is different. That's one thing I learned when I owned my own consulting company. I went to every organization. It's like every company had its own flavor, is the best way I could put it. It did things differently because every company is different. Different platforms, different people, different thinking. And so it has to be sensible as well. Open integrations, so that way you can plug this into your other security workflows, processes, a soft, and so on. Guardrails and rollback, really important here because of the ideas. If you're gonna do automation with those guardrails should come the ability… I forgot to mention this on that slide. With the guardrails comes an ability to roll back. That way, if something does go wrong with automation, you should be able to undo it. And that way, you can go and kinda stop. That's your way of going, wait. Hold on, And let's figure out what went wrong, and then let… Then we'll put it back into production to be used later on once we've actually figured that out. But in the meantime, let's roll back. Audit ready reporting so that those metrics we talked about, you can dig in deeper, and so on. And then the last one here is just, agent and vendor trust. There's, there's an ability there for that platform to utilize, the agents within the organization. You have to get some trust back to there with the vendor. And so you're looking for some things around the application, like the updates are already signed. You've implemented at least stay up. You'll implement least further. They want least further to implement and things like that. There's attestations to the, the security that's implemented and so on. I'm gonna skip that thing there because I think I've already talked about it. I think this is my last slide. Is that right? Oh, I don't know. The first ninety days left, and then I have… This is your second to last slide. These are the kinds of questions you should be asking vendors, and it's like the first one is basically like, hey. I have everyone's everywhere. People don't attack their VPN. How do you support that? If I have a sale patch, I will roll that back. How do you find, a a fix or something without having to send you a ticket? What zero day patch zero day attack will no patch? What I do in the first twenty four hours of your platform? These are all sort of situational things that relate back to everything we've talked about. I don't think any of these are entirely, again, earth shattering like I mentioned early on, but it is… You want to be knowing these things. Things like number two, number six, Those are putting time frames on it for the vendor to go. You have this much time, twenty four hours. What am I doing with your platform? I have five minutes. What can I hand an auditor? Those are kind of important questions because those are realities for you as opposed to, you know, our product comes in 14 languages and seven colors. Good for you. That's not my world. I have to hand an order something in five minutes, or I have to have the system patched in an hour and a half or whatever your time constraints are. That's reality. So I think it's really explaining your reality and questions to see what they can do for you. Alright. So your first ninety days, the first thirty days baseline. You wanna figure out inventory that you have, wanna figure out your, your time to patch is, what your coverage is today, and so on to the pilot ring of devices that to do work with. Put some automation. Remember that… Those steps. Maybe you're first gonna just simply have it notify you, then you're gonna add approvals, and you're gonna automate and do some kind of level of automation here, start writing a run book. And in ninety days, you're gonna approve the security utilizing the baseline. And so now you're gonna have to do and do some drill as an example, reporting the numbers, and you get that prompt. So, I I hope this is better, you know, educational for you. I hope it gave you a chance to kinda stop and just think about this problem here. But the goals here are, you gotta shrink the window. You can't patch in days or… Effectively or weeks. It it has to be hours and so on. You have to look at automation. The guardrails are gonna be your friend there, but you wanna be thinking about automation, you wanna be able to prove the security of your records. So, again, I hope this is educational. Get something out of it. And, and with that, happy to take some questions, but I'll pass things back to Keith. Keith Ward Hey, Nick. Great stuff. Great stuff. I loved that presentation. Thank you so much for that heavy presentation full of shocking statistics and and outstanding advice. I think we do have time for a couple of questions, though. Can you do me a favor also in the meantime, Nick, and please close-up your screen share? That would be excellent. Alright. There. we go. Appreciate that. Okay. And we'll get this up in the meantime. Remember, he is a conversational geek, so you can hit him up right there. Okay. Time for just a couple of questions today. But that's okay because sometimes there's so much good information that you gotta get out there that that is where you need to focus, and I'd never mind that. So here's an interesting question from Rohan. Where do you draw the line between relying on automated AI driven detection versus human led threat hunting when managing hundreds of endpoints? Nick Cavalancia Good question. That's a deep question too, Keith Ward Yeah. Nick Cavalancia Rohan. Because the challenge… So I spend 70, 80% of my day, working in AI, both… It's for right to operations and for business processes, building platforms, all kinds of stuff. And one of the things that I have learned, and this applies back to your question talking about, like, AI based, you know, threat hunting versus human threat hunting. And it's sort of putting a little bit of work around, but taking that that is a a grain of salt, if you will, is that AI is the stupidest smart person in the room. That's the first thing I can say. And meaning if you don't know what you want it to do, it will do whatever you're told to do. And, yes, I I know that the models are fantastic. I mean, you've… The… Mythos found all those vulnerabilities. You know, Cloud just came out of Table five two, and that mean they get smarter and smarter. But even so, and I'm I'm going off kind of on a tangent, but I will answer your question. Even so, there's still lots of occurrences that occur. Even in my own life, I'm asking something that is not as critical as I need to to go and and patch and assess the vulnerability and exposure of hundreds of endpoints. I still will go, why did you cope with that answer? Because this is true. And it goes, oh, you know what? You're right. Sorry. My bad. And it's like, what? You're smarter than me. And you? Well, my… Yeah. It doesn't say my bad. But, you know, it it sort of does. And the the point there… Now coming back, I'll circle your question is, I think where you… It's not so much draw the line. Because draw the line sounds like we're no longer gonna let you cross the line, but where do you find the the break point between the two? I think what's necessary is to, have it perform very specific tasks that have been defined by the human experts. Maybe have it validate the tests that are gonna be done, seeing how it wants to augment those tests. So even before you ever do any kind of vulnerability testing and scanning and whatnot with AI, just figure that out is go through the process of I want you to do x. What else would you add? And then you go, okay. What do you mean by that? And you're doing a lot of back and forth. I know that's more work than you have time for. But that's not how I would approach that. So that way, the person that is in charge of, let's say, all your vulnerability scanning for the sake of argument actually knows the exact things that are gonna be done and then also asks for detailed reports of each thing independently, whether there's an… Whether there was something found or not. And the reason for that is holding AI accountable because it will also forget the context windows always eventually run out and you have to, you know, you have to clear out that your sessions and so on. But even if you're having it automated, making sure that it's doing the job it's supposed to do. I just had something today where I have something that's completely planned. It goes to a data source, it pulls the data, and it started reporting to me. You know, actually, that that data source stopped reporting on this one thing since, like, September. And And I just kinda let it go thinking it was right. I'm like, you know what? Wait a second. Spin up a brand new agent that knows nothing about anything we've done. Point it at that data source. Tell it you need to get this data from that data source and see what it collects. If it comes back and… Yeah. I go, you know, you're right. It was all there. There was a problem, and we didn't know. What do you mean you didn't know? That's your job. You're supposed to be. And so you you gotta verify what it's gonna do in the first place, and you gotta verify what it's doing afterward. And those are some of those guardrails. It's some human in the loop guardrails until you are comfortable enough that you know that there's a percentage of time… Or an amount of time where a percentage of its response is above some threshold you're willing to accept. Like, okay. It's it's it's right 96% of the time these days. So we'll just check-in less and less because we trust it. I hope that's your answer. It's a long winded answer, but I hope it helps. So… Keith. Keith Ward It it is… And it's such an important answer, Nick. I saw a statistic within the last week that up to 50% of output in many companies… A… Of AI output is not checked by a human before it gets into production or gets published or whatever. Now to me, that is absolutely mind boggling because any of us who have used AI for any amount of time have seen multiple hallucinations. They… AI gets so many things wrong. So why would we offload that job? And I think you… I I think it may just be… We assume it can handle it without checking on the AI or we use AI just to check on the AI that checked on the AI and the human gets more and more removed from that loop. Nick Cavalancia But but that's at least better. That's the… If if those of you that are are still trying to use I… AI for IT operations, even though this is kind of off topic, if let's say you do have it write a report, then go tell it, act as a CISO and review this report and and find anything that makes absolutely no sense to you or you're gonna give it criteria of where it should find things wrong and have that report back to you. That's at least a step in the right direction, Keith. I don't mind AI checking on AI, especially when you spawn off something that's brand new and you'd give it a different context and say, now go view this first thing and. tell them what's wrong with this first thing in this particular context. Then at least you're getting a better set of eyes. But… Keith Ward You you are. You are. But but what I'm… My fear, I guess, Nick, is that we are removing humans further and further out of that. loop. And in which case, it never really gets the kind of ultimate check that it needs just to make sure that that the AI did, did ultimately come to the conclusion that two plus two equals four because we have known AI to to get so many things wrong, you know, as, you said. Nick Cavalancia there is… What is it? It's four e's in strawberry or whatever. What was the what was the old chat g p t thing that was, like, it couldn't… There's some letters in strawberry that always got wrong, and you'd ask it which ones. Do you remember that? Somebody. in the audience remembered that. You put it in the chat, but it was, like… This is an early on model, like, two, three, four years ago. But you had asked, like, is it how many r's are in strawberries? I said, four. I said, really? Tell me what what letter positions. It would be like the third, s t r, if you got it. And then the the the the r in berry and the other r in berry. And, like, and and the sixth position, like, that's a b, you know, or whatever. Well, that's wrong. You just can't trust Keith Ward warnings. warnings that should be heeded, Nick, for sure, Nick Cavalancia Yeah. Keith Ward especially as AI does more and more of our work, you know, in organizations these days. So, Nick, I have to say this is my first event I've done with you. Of the many hundreds I have done, this is my first with you, and it was an absolute delight. I enjoyed it so much, and I wanna thank you for the great presentation, for your time in the q and a, and for joining us on the expert cast today. It has been an absolute pleasure. Nick Cavalancia It's Adana. Appreciate being here. Keith Ward Thank you so much, Nick. Alright. And there goes Nick, everyone. He was absolutely fantastic, as I said he would be. I had high hopes for him and he did not disappoint. Now before we do the grand prize drawing, which you know is coming up, we are first gonna open a quick survey for you, Irina. If you could launch the survey, that would be great. Now for those of you who don't know, this is a survey that you can fill out for us at actual tech to tell us what we do well and more importantly, what we don't do so well. Please let us know how we can make these better for you, the kind of things that you simply must attend. For instance, you may have feedback like get that clown, you know, in the striped shirt off. Please find another moderator in the future. I cannot stand listening to Keith anymore. That is the kind of feedback that is very valuable and that we will pass right along. So please take time to fill out that survey while, I give you quick little overview of the handout that Automox has for us. That is over in the doc section. They've got an informational site with a lot of different, stuff there, a lot of goodies for you to check out. One of them is a link to get a demo of how Automox works. They have got links to all kinds of good stuff. You can explore cross OS patch management, all kinds of information. You can explore turnkey results from their product. So lots of excellent stuff. Please check it out right now. Alright. And with that, I am now gonna ask Irina to close-up that survey. Irina, please close that up because we are ready for the last duty that I have on this expert cast today. And you know what that is, but you know I'm not gonna give away the prize until you all do something for me. So let's see it happen, folks. Uh-oh. I'm not seeing it yet, so I need to see everything. Are you guys, holding out on me today? Let's see. Let's see. I have got a gift card to give away. Let's see. Are all of you hearing me? Can you give me a thumbs up in the chat if you are hearing me? Okay. Excellent. Very, very good. So we have got the $250 Best Buy gift card to give away. And the winner of that for the expert cast is Mark Stobitz from Ohio. So congrats to you, Mark, and sit tight because we will be in touch ASAP about claiming your prize. And, folks, with that, on behalf of the actual tech media team, including the fabulous arena, I wanna thank Automox for making this event possible. Remember, if they aren't here, I'm not here and neither are you. And a huge thank you to Nick today for an amazing presentation. I really hope we can have him back very, very soon. I learned a ton. I felt like in some ways a kid in his first day of class. I think all of you learned a lot too. That is why I love these and I love being here with you. So until the next one folks, that is all. Have a great weekend everyone. If you can start it a little early, consider doing that. Bye bye now.

Please provide citations & attributions for any information you generate from this content, thank you.