Video: Get Your Nights Back: How Actionable Data Intelligence Reduces Risk, Cuts Noise, and Accelerates AI Adoption | Duration: 2712s | Summary: Get Your Nights Back: How Actionable Data Intelligence Reduces Risk, Cuts Noise, and Accelerates AI Adoption | Chapters: Welcome and Introductions (27.499999s), Data Security Challenges (327.599999s), Policy and Alert Quality (997.059999s), Classification Insights (1463.609999s), Building Cross-Team Collaboration (1728.339999s), AI Security Concerns (1880.059999s), Practitioner Insights (1977.375s), Policy and AI Transition (2293.3999989999998s)
Transcript for "Get Your Nights Back: How Actionable Data Intelligence Reduces Risk, Cuts Noise, and Accelerates AI Adoption":
Welcome, everyone. My name is Christy Hart Smith, and this is my, cat pilot, Danger Mouse, who is here with us today. The title of our session is get your nights back, and I really kinda mean that literally. I I picked my cat up because she sleeps in the window next to me waiting for me to get off of work so that she can have, s n a c k s and, and get a lot more p e t s. We have to spell things because she's a very smart cat. But we are also here to talk about, data security, and better data security should not mean more work and more noise or more late evenings. So Jennifer Glenn, who is joining us from IDC, is gonna tell us a little bit about why IDC found across nine different organizations using Cyera. And Carlos, if I move over here onto our next slide here, is going to tell us what things look like on an ordinary Tuesday. So, I kind of keep asking myself a question, as I've been thinking through this material on, like, what changes when the information that your team has is something that it can act on. Now a quick transparency note as we get started. IDC ran this research independently. Saira sponsored it, and IDC doesn't endorse the sponsor's products. But what I would love to do is allow each of these two amazing speakers to give a brief intro. So, Jennifer Glenn, you are IDC's research director for information and data security. Your work covers things like DLP and data sovereignty, trusted data for AI, and you coauthored this, business value of Cyera with Megan Surly of IDC's business value practice. Welcome. Anything you'd, like to add? And, actually, I'll throw my icebreaker in here on the, getting the plates back portion. What do you like to protect your nights for? Well, let me just... And if I add to my own introduction, I've been in the security space for for several years, even though I've only been with IDC for five. But, really, what I like to do in my spare time is not. I like to spend time with my kids, spend time with my husband. My, you know, my kids do ninja warriors. I like to watch them do that. It's pretty cool and I'm kinda jealous. And also just, you know, I like to actually try to coax my garden to kinda grow a little bit. That's what I do in my spare time. Amazing. Carlos, Carlos Escobar is VP of information security at Revolution Medicines. You have nearly two decades in IT, most of it in bio and pharma. You know, I, noticed that you like to think of this job as using technology to help the business drive results, and that's a pretty big lens today. You've got a lot of, collaboration with the different roles that you've had and and and the team that you're leading today. So my icebreaker to you here as well. Share a little bit more about, your background and what you like to protect your nights for. Sure. So, yeah, my background, as you mentioned, is working in IT and cybersecurity for a couple decades. That sounds like a lot. But, you know, really focusing on, especially in in the realm of, biopharma technology. I So, Jennifer, I'm gonna start with a couple of questions for you. You know, you... I think you probably have some talking points around these data elements that we brought here, but I'm I'm interested in your perspective as an analyst. Like, when you look at our market, you talk to a ton of practitioners and you're you're looking at, like, what makes data security materially harder right now. Yeah. So... Yeah. I mean, just, I think about the conversations I have with, users who call with questions about, you know, things that they're looking for, the pain points that they have. And I look at our syndicated research and, you know, the syndicated research that we do outside of, you know, this this custom survey. And one of the things that, one of the big things that I have seen in that syndicated research, which is is interesting, is that security and compliance budgets are competing with AI now. I think for many years, when I looked at spending and budgets and resources, security was always the number one, you know, this budget is not going to be cut. This one's going to grow. This is where we're gonna invest. And now, I mean, it is competing pretty heavily with with AI, you know, tooling and and initiatives, generally speaking. I mean, just a data note here. I have, from our CIO sentiment study, 35% of CIOs said that AI and automation will receive their most investment in the next twelve months compared with only 31% that said, cybersecurity and risk management would get the most of that investment. So it's just really sort of an overarching theme that I think, is notable and important. Yes. Yeah. And and the other thing that I see, is just data volume, I think, is a huge, huge problem. That probably is one of the things I hear the most. We just have too much data, and trying to manage it and control it and see it all is is just so... Such a problem. And AI is just making that worse. I wound up fielding some internal questions from the Sayuri team that supports DataWatcher, which is our services team that we use to extend our own clients' capabilities. And some of the questions that came up were around, like, the the FTEs, and and salary avoidances that you noticed there. And, really, what the topic sort of boiled down to and and what you're getting at here is, like, the amount of data that is being generated keeps exponentially increasing, especially now with AI tools both consuming and now creating their own data that, like, it's not necessarily that... It's not like you're not gonna hire somebody. It's that how do you keep up with the volume of data that your organization is creating without without having to keep hiring additional staff to deal with the influx of data volume. Yes. I... 100% the biggest problem. I mean... And not only is there more volume, it's obviously in more places and it's moving all of these different different ways. But, it's just, the amount of data is probably the number one thing I hear the most, and then followed by, you know, the need to to do something with that data. Right? Use it for compliance purposes, which is just a number one driver for any security, of course, is meeting compliance regulations, but also, again, just keeping any sensitive data, from being exposed or or exfiltrated. Yeah. Carlos, I'd love to turn the, the mic over to you for a moment. So if you think about the challenges that you've been facing, how much of the challenge has been finding the risk out there in your environment versus giving the people who own it enough context in order to act on it meaningfully? Yeah. That's a great, question. And so, like, finding the data has, has always been a a main sticking point for me because I think understanding your threat landscape is really important. And so finding the data... But not not just finding the data, but also classifying the data, finding who has access to it, it... It's such a big piece. And, Jennifer mentioned that the amount of data being generated now has immensely increased. And so, that leads to at least my security team dealing with alert fatigue because of all the different types of data that's involved in, securing that data and keeping it compliant. So, it... It's had a huge impact. Yeah. I can I can absolutely imagine that? Let's, I'm gonna shift over here into our next slide that, pops up a number of numbers that came from the, the business value paper. So this is nine organizations across a range of industries and sizes. And so, you know, Jennifer, I'll start again with you on this one. How should somebody read an average, like, 441 or 3,300,000? And, you know, how how does that, just, like, maybe help a practitioner make these numbers a little bit more credible to what their own organization might be facing. Yeah. I think, well, I know I look at some of these numbers. I I looked at some of the the bigger percentages because I always think that, any sort of big percentage or small percentages is always the best ones to talk about. But when I talk about, you know, when you see 68 of more actionable information, I think, you know, what what we hear... And and Carlos actually mentioned it too. It's it's having the ability to not only see the problem, but to, you know, reduce that fatigue, to make sure that when you are acting on an incident or, a particular risk that you know with, you know, a relative amount of confidence that that is an actual risk, and you're not, intentionally or inadvertently, I suppose, blocking business from happening. And I think, the problem with the teams that we talked to just coming from this this exact survey, it wasn't, a shortage of data. Obviously, they have a lot, and it wasn't a lack of tooling either. It was really, just making sure that they had that confidence that they didn't have to keep checking and rechecking to make sure that they had everything right. You know, there's better... You know, they needed better classification. They needed this ability to just, again, be more confident that the actions they were taking were the right ones. I love that point about confidence. And, really, I think back, like, very early in my career where I realized that the kind of information that I was providing during, you know, systems integrations and mergers and acquisitions were were things like decision support. I wasn't necessarily making the decisions myself, but I was providing the support for the individuals who would say, yes. This is the path forward. Yes. So I was the one that was providing them with the context and the confidence in order for them to feel good about the decision that they're making. Was there anything of the findings that you had across this, Jennifer, that surprised you? Or was there anything maybe around the teams that had the biggest gains that maybe they had in common that you noticed about the the organizations? Yeah. I think one of the things that I thought was was interesting, it was this 51% efficiency of teams. And one of the questions that I I typically ask when I get calls coming in and say, hey. What, you know, I I really need a DLP. Is it for, you know, any... Somebody who maybe hasn't had DLP before or something. They will call and say, well, you know, I I need something. Can you give me an idea what that is? And my first question to them is always, before I, you know, before I can answer your question, I need to know, are you going to be able to hire somebody? Or do you have to stick with the team that you have? And I think one of the things that came out of the findings here was, there was an average of five fewer full time employees worth of manual scanning and classification, false positive investigation. And so that was time that was redirected, to other things versus headcount being cut. Like, I thought that 51% was a really interesting a really interesting stat. Yeah. I can absolutely see that. And, you know, the report itself was also, you know, getting into, like, what does it take for a team, to operate. The report talks about how vendors can supply things like visibility and context, but not the organizational alignment. You know, maybe what I'll do is bridge this over to Carlos here, like, organizational alignment, collaboration, the quality of the information that you're putting in front of the team so that they feel that they're competent, so that they trust you, that you're providing them with information that they can take action on. Like, what does that look like for you in practice? Yeah. So, I see this with the d... DSPM and also with Omni DLT, where I've had to collaborate with other functions of of the company, including compliance and legal, to to really sift through what's... And really cutting out the noise, sifting through actionable items that have been found within the the alerts within Cyera. And those those actionable items have led to, like, true investigations for for us to, like, dive deep into. But not only has that led to, like, looking at investigations, but then it leads to the development of policies or written policies within the company, which then has an impact on how we secure and manage our data. So that that is a big piece. We we see this all the time with the, the collaboration with our our functions in in general as we see those incidents. They could be actually doing, true business related activities. And so, that may come off... May come across as an alert, but then it's working, within Cyera to ensure that we, like, we modify our policies, ensure we understand what those business processes are so that, we're not identifying them as potential threats. And so then it's less noise for us to deal with later. So, getting that date... That time back again at night, to do those important things. So it's it's really kinda cutting into that again. Yeah. That... You brought up a number of things there. Policy is one of those areas where, like, I didn't think I was gonna be as passionate about policy if I look back in my career. And then now all of a sudden, I, you know, I developed just this really deep appreciation for the consistency of policy. I mean, for those of us that have been in the security industry for a hot minute, you know that you've got a DLP with just about every type of a tool you bring in. Right? If you have an email provider, you have a DLP. If you have a cloud provider, you have a DLP. You've got a DLP in all of these different places. All of them have their own policy engines. None of them necessarily understand the data itself really well, and so it's hard to, like, get quality results coming up out of those that a team can trust to take action on and, like, believe you that this is something that's worth spending their time to investigate. And, also, you're dealing with this legal and compliance issue of the fact that data... The same type of data is not treated the same across various different areas of your environment. And so having that consistency and the quality of information behind it really can make a difference in how the teams can function well. So I can see I can see that, quite clearly from my own perspective. Carlos, if you look at the numbers that are on this particular slide, are there any of these outcomes that maybe feel closest to your own team's experience? Yeah. I mean, I would say that that the 68% of more actionable information is is is really truly there. We see this with, like, OmniDLP as an example where, it it can really cut through the noise, and we can really dive deep into, whether it's a higher critical alert and then understand how that really needs to be dealt with. And and, like I said, moving moving to to to collaborate with legal and compliance and quickly... Or or legal IP and understanding what the sensitivity of the data is and ensuring that it's being, processed appropriately. And so we we understand the impact. So I I think that is a definite, like, key number for me and then, obviously, the efficiency and the security teams because, I I want my teams to be focusing on on on true actionable items and not chasing down false positives at at a constant rate. So the more that we can move towards those actionable items, then the less that they're having to to waste the time. And and it's just not fun chasing false positives around all the time. So it... It's really just, ensuring that they have, like, a a good working environment to work in. You know, I was thinking about, your, like, before versus after on DLP alert tuning, as I, pull together this, this slide here when the one last alert is actually the last alert. Right? The the quality of the alerts really make a difference. And so, you know, we're gonna move into a little bit of understanding just your own story, but, maybe just on this particular meme, like, is there anything around how alerts have become more consumable, that has maybe made a difference or an impact on you or your team's ability to enjoy their evenings off? Yeah. The alerts, just at face value, they're they're so informative. They give you the AI summary of of what, is happening within the alert. It gives the the the the user that's involved in the alert, what actions are being taken, and then you can drill down to the evidence real clearly without actually moving to the actual d s... DLP platform. So you can see all of them on a single pane if you have multiple d l p... DLP platforms. So having that is is just a key piece to get that information quickly. So, you know, data is is great when you can make decisions. And so if you have that data there and you can make that decision quickly as to whether or not we need to move to an investigation or not, it it... It's it's valuable. Great. Well, let's, you know, kind of along those lines. Right? I am gonna move into this next portion of our story, which is, like, talking a little bit about the customer story that is is coming up behind some of the numbers that we've just looked at. So, you know, I'll... I'd like to spend a little bit of time on this one. And the the questions that I pulled together are ones that really look at, like, the before stage of what you had been dealing with, you know, what you changed, how... What kinds of changes that you made, and then what it enabled, like, the outcomes, the the getting your nights back kind of a session that align with, you know, the theme the theme of this. And so, you know, I'll start off, Carlos, with the question. Before Cyera, where was your team's time going that maybe you'd really rather it had been going? Sure. So it... It's hard to fathom what that life was before, truly, like, there's... There was so much information being... That we would be bombarded with as far as whether it's DLP alerts from the platforms themselves, with little context. And so having to sift through them and read through them and and really decipher what was happening was, just overwhelming to the point where, you know, we really had to almost pick and choose alerts as they came through. And so Yeah. There was... There was just no... There's no form or reason. It was just like, I'll get through it if I can get to it. And then really on the DSPM side, it... Gathering a list, an an AI based preview of of the classification of the data. I had... I didn't have that. I I had spreadsheets of of, like, what classifications of data had, and it was impossible to keep up to date, with the growth of the company. So, it it was never really accurate, or or very little inaccuracy. So having that at... Now at the DSPM level just just change the game and then being able to then apply, good security practices and and security frameworks around the management of that sensitive data. It it just it just all bled together and and and led to really a quick identification of of issues when we first brought on Cyera, to, like, remediation within weeks, you know, of of really identifying true true concerns. So and and totally, again, valuable ones more. Yeah. I, I can't even... Well, I guess I can because when I go back in my career and I think about all of the different systems that I've had and the impact of all of those noisy alerts, like your phrase of just, like, having to pick and choose the ones that you guess are going to be the ones that maybe are the most important and then the blind spots of not really noticing or understanding. Like, you talked about the value coming out of the the the classification side of things. But was there anything that surprised you about the kinds of learned classifications or things that might have been unique to your business that you just might not have found otherwise? Yeah. I'm... And and, I went through an exercise with working with some of our key functions and, you know, exporting a list of those classifications with them and and going through the functions. And it it was really helpful because a a lot of them were justified, in those learned classifications, which is really helpful. But then we could also customize and build our own classifications based on our business. And that that took things to another level because then we could really identify new classifications and new access points for that type of data, which which leads to better security. Right? So I think that was a key piece. One of the things that I love having spent a lot of time, building and explaining policies over the years, is the way that you can not only create those classifications, but you can group them into topics, right, and then create the policies based on the topics. It's a way to, like, get the same kind of a treatment for a a variety of things where it's helpful to understand the granularity of it, especially if you're trying to quantify risk or, you know, dealing with an incident or something along those lines. But then, being able to have, like, just the ease of creating, a single policy that can handle over wide variety of those classifications across multiple enforcement points, that, that speaks really clearly to my own to my own background. We talked a little bit about the alerts, as being more consumable, and we talked a little bit about what that has meant, for your team. Did you find that that improved... Like, what was the experience like with... Did you get reactions from individuals as they started seeing these? Like, you know, what did that... How did that change maybe their willingness to participate in your program, and maybe just the overall reputation and, brand of your program within the company. Yeah. That's the alerts were were eye opening to a a lot of different people and, well, starting with the security team itself, really, truly, like, getting the identification of the users involved and the con... Context of, the data and this... The the sensitivity of the data involved within these alerts. Having that information was just completely invaluable, and and it really made it so that there was a lot less of this, you know, what I like to call is, like, IS security digging and and trying to, correlate information between our different security tool platforms. And so having that all in one place was key. And then to the to the functions that we're collaborating with, you know, I would get comments from our compliance teams regularly. Like, we... I've never seen this before. This is amazing. We're protecting the company so so easily, and and we can... We have the information available on this. It's leading to written policy. So I think it's just... They they they found so much value to it, that that it was, just overwhelming. And then I would say the the the people involved with trying to understand the program, let's say, the the people involved within the alerts, it was eye opening to them to understand that, like, hey. We are... We're we're serious about protecting our data. And it it... And a lot of times, it wasn't in a in a... An approach of, hey. You did this this thing, and we're gonna slap you on the wrist. It's more of like, hey. We want you to understand how sensitive the data is that you're working with. What are the potential threats that are involved with with with, with not purposing that that data appropriately. So I think it... That's that's an eye opening piece that is just all around happening. Yeah. And I find this, I find this is true for just about every role I've ever held that being able to create great working relationships with this diverse of a group of people across my organization really improves my ability to get my job done. I mean, a little bit, it's the function of the role I have now as analyst relations. And so in a way, I can kinda, like, get into everybody else's business about, you know, what they're doing and how well, how well they're growing and what kind of success stories they have and where they're... Where they wanna learn. So I find that that makes my own program more successful. And and I also can see how from a security perspective moving into a world where everyone at your company could be building and managing agents and thinking about the ways in which data is flowing to be able to have a culture and a collaboration across the board with your different teams where where you're a trusted entity, where you're helping them make good decisions about things that they otherwise maybe wouldn't care about so much. But, like, certainly, legal and compliance are are going to care for sure. But that really, that's really a testament to the program that you've built. No. Thank you. With all of the time back on this, is there anything that, that your team got to do instead? Were you able to prioritize your team and achieve anything beyond what you would have otherwise spent your time doing? Yeah. I mean, I've... I find it amazing that I haven't talked, like, delved deep delved deep into AI within this conversation but within, like, thirty... The thirty minutes we were discussing. But, you know, there's there's the the always there type of security items, type of security concerns that we always have around security and access and DLP. But then, obviously, it's the the new things around AI and MCPs and and all of the AI assets that that you may need to have concerns about. So my team has more time to really focus on understanding the tool sets that are out there, with with... But also the new potential threats that AI is is now introducing into the landscape. I think having the team focus on that has really been helpful because we can get what feels like ahead of the game, although it never feels that way. But at least you feel like we're we're we're looking at some of them... These new potential threats and feeling like we're we're jumping to that new, realm. Yeah. I can see that. It's, the mandates for rolling out AI tools, I mean, it's become just for all organizations, just a competitive mandate in order to be able to compete in today's world, right, to to leverage some of these new technologies in responsible ways. I would love to flip the conversation over, Jennifer, onto your side and sharpen things a little bit because you talk to hundreds, if not thousands, of practitioners a year about topics like this that go... That goes above and beyond the the conversations that you held for the creation of this paper. So thinking about what Carlos has shared and, like, this customer story behind the numbers that's here, how how does this compare with patterns that you hear? Are there any places where experiences that you've heard from other practitioners might differ? Oh, I don't know about differ. But, as Carlos was talking, I was sharing... Or I was writing down some of the things he said because I said, oh, I I have something I wanna say about that because, the first piece was on classification and being able to get, yeah, I know Carlos was, you know, talking about your product and getting classification sent sent to him. But in general, classification is one of the biggest problems I hear, with with DLP, data access governance, is having really good accurate classification. I think it's something that a lot of companies struggle with. You know, it's, and the challenge comes down, I think, to to a couple of things. But, I... If I'm if I'm paraphrasing or, you know, kind of giving you an average anecdotal view of what that looks like, you know, it's like different data may have different meanings to other people, and you wanna make sure that as an organization, you're getting all of the data classified appropriately so that, you know, everybody can get their their data in the way that they need. Right? So I think some of the the classification piece is huge. That's something I hear, again, so so often. And then the other piece that I wrote down and I wanted to talk about because, I mean, again, this is one thing that I hear, a lot is, the need to communicate cross functionally. To be able to use all of the information that you're getting from, you know, your security tool and be able to reframe that in... Reframe that risk in a way that's valuable and, you know, actionable, I suppose, from from other groups, you know, from legal or HR or compliance. That is, you know, one of the things that, again, when I hear and I talk to practitioners, that's what they're looking for. So... That's, that's really... Those points are really well made. And your comment on classification reminds me that there was a few years ago at Black Hat... If I TikTok, which I don't because I'm in security, but if I did, this would have been hilarious, but I went around to every booth on the vendor floor that had posture management in it. And I was like, hi. Excuse me. Can you... I see a posture management on your booth. What's your definition of posture management? And then the the answers I got back were completely crazy. But, like, classification or maybe DSPM has become, it's something that exploded onto the scene, like, around 2020, and then now it's almost like a sigh of fries. Like, I see so many different vendors out there saying that they have DSPM. Now my own experience with these, tools and certainly, like, my delve into it leads me to believe that not all of these classification capabilities are built the same. So, your point on classification is being, really important is... It's also important to note that not all classification engines function the same, just like DLP has been a dirty word for a long time because it hasn't really delivered on the promise of what it could be. And I think a lot of that is because the DLP doesn't understand the data that it's trying to protect. One of the other points that you made, Jennifer, and I and I just really wanna give you a shout out on this because I see this in the way that you approach all of your interactions as an analyst because I've seen you with practitioners, and I've seen you work with vendors as well. And, you do focus teams on not just the technology that they have on hand, but really thinking through what are the people and the process that are needed in order for this to be successful. And I know you and I have also talked about the information that's provided. I mentioned decision support earlier. Right? But, like, what is the information that's being provided that allows these processes to function in these people to do a good job at their jobs. But... So thank you for that because... Thank you for being that champion. Are there any other, points on this customer story behind the the numbers that we want to hit on before we before we knew... You move next? Let me just go make sure I checked off everything on my list here. But, you know, again, I think, some of the other things that were were sort of interesting were 47% less time to resolve and remediate. That was sort of an average of what, companies had. I think as we're talking through this this particular topic of get your nights back, just, again, having the ability to, you know, having the ability to spend less time. And I know Carlos, you could... You talked about that quite a bit. Just being able to shorten the time that you have to do all of these things as... That 47% was another interesting, was an interesting statistic I wanna talk about or not. Right. An interesting finding. Interesting finding. Excellent. I'm looking over some of the other, questions that I've had come in and and some of these that I jotted down in my in my other notes. So, you know, Carlos, maybe this one for you as we were talking about policy a moment ago. Has there been, like, an incident or a particular investigation? You're talking about both legal and compliance being really delighted with what they were able to find from this. Was there a particular situation that really changed a broader policy conversation that, you know, led to, a significant improvement or a change in some policy that you had or your ability to enforce it? Yeah. One of the main things that we've been seeing in some alerts are related to, what's being captured within, unsanctioned recording devices, which is, I think, is a hot topic right now because I think that's an area that is really changing, because recording devices is no longer just like a a handheld recorder. It could be a wearable, eventually, an injectable, what... Whatever it may be. Recording devices, maybe recording information within the company, but then also there's AI being used as a note taking device to to then process that information that's being recorded and then being used. Right? So I think that is a hot topic right now that we're we're we're looking to define in policy because I think that that's something we wanna get ahead of because we know that with the the new technology coming down, it it... It's just something we we need to address now. That's, that is a that's a big one. I think I was reading there was some legislation in in EMEA where they were, like, talk... Talking about blocking, wearable devices in public spaces, just out of concerns for for other people's, privacy on that. You know, a lot of the other questions that I have here are, coming in from the word of the year, AI front. So let's move over to, the next slide here, which is a nice segue to talk a little bit about moving from data security over into AI confidence. And, you know, if I look at research and comments across the board, you know, like, AI ready is then this phrase that I see tagged on to just about everything. Is there a minimum of what a team should be able to show about its data before it scales AI? Hang on, Penny. I'll I'll take an answer from... Actually, I'll take an answer from both of you on this one because you might have some similar answers, but you would have some for... Like, some, nuance twists on this one. Yeah. I think... I mean, it... To me, it boils down to just four questions. Right? It's the ability to answer what data do I have. You know, is it restricted? Is it confidential? Is it sensitive? You know, what what data do I actually have? You have to know this before you start AI. Right? You have to be able to know where your data... You know, what data you have in order to be able to getting value from it, as well as protect it. Right? So what what data do I have? Where does it live? And that way you can understand where it's coming from to know whether whether it's legitimate or, trustworthy, as well as, you know, you know where to start putting protections as well. You know, who has access to the data, I think, is another key question. I think that's a that's a big one because it's no longer just, a person anymore. It's, you know, what agents have access to that data. And then the fourth is, is that appropriately protected? So if you can talk about... If you can answer those four questions, that... Those are, to me, the four the four basics of AI readiness for for data. Amazing. I I don't disagree with you at all. Carlos, do you have anything to add there? No. I I think those are, like, I I would totally totally agree with those four. I mean, I think I would maybe highlight the the access, like, the role based access is a is a key piece on the security front for me. Whether that's the agent's access and enroll within the function of of AI being able to use that data, or the user. But eventually, right, it may... It's gonna be agents. So that... That's really a key piece that, we wanna keep a a spotlight on. Yeah. I can see that. And and I had a couple of questions around, I'll just maybe, like, I'll broadly lump it in on identity side. But to be maybe a little more specific with this, you know, like, what happens in situations when tools or agents are inheriting access to sensitive or overexposed data. Right? That's, that's a big piece. And so, kind of curious about, like, in getting around the... Getting your arms around this kind of a problem, you know, what, what kinds of information do you need to understand? Or are there some some controls or safeguards that you need to put in place in order to address it? Different. You wanna go, Carlos? Yeah. I'll I'll touch on this a little bit because, with the use of DSPM, it it it comes back to the data classification and having that visibility in the data classification has really led to understanding who should have access to it. And it really has led to discussions with the functions about, really defining better who should have access to our data, defining that a little bit better because, eventually, like you said, AI will will need or want access to a lot a lot of data for it to to do its purpose.