Transcript
Mike Matchett: Hi Mike Matchett with Small World Big Data and we're are talking today about AI, which is one of my favorite topics. Uh, we're talking about though, how do you practically use AI to do something useful in it? And we always talk about data protection. So we're going to dig in today on how you might productively use AI to achieve your data protection goals. And specifically we're going to talk with haiku here about using it for your avoiding the SaaS data apocalypse. That's a lot of stuff to unwrap. Just hang on. You love. Hey, Simon, welcome to our show again. Simon Taylor: Hey, Mike. It's great to be here. Mike Matchett: Now, I know previously we talked about the SaaS data apocalypse and and people going on there in haiku. Before we dive into that, though, let's just give us the elevator pitch for haiku. Where do you guys fit into this data protection market? Simon Taylor: Absolutely. Haiku is the world's fastest growing data protection as a service company focused on what we call protecting the modern data estate, that means protecting data across on prem. Public cloud and SaaS. We do that with the haiku Art cloud platform, which provides visualization of your entire data estate, the ability to protect all of your data through the haiku marketplace, as well as a low code development platform that allows you to actually build integrations to future proof your data protection strategy across on prem services, public cloud services, and SaaS services as well. We've got about 4000 customers in 78 countries today. Mike Matchett: Wow. So that was a that was a great, uh, great summary of what was going on there. Uh, last now, previously we've talked about this idea, and I believe you wrote, uh, wrote a significant book on this called the SaaS Data Apocalypse, or about the SaaS Data Apocalypse. Uh, and, uh, let's just start there. Companies are doing a lot more than just processing data in their data center and doing a lot more than just processing data by renting private clouds or colos and stuff. They're using more and more SaaS apps and more and more of that. Corporate data is getting distributed across third party properties. They don't even have control or necessarily even insight into. But it's corporate data that's still needs to be protected. Right? So so when you're when you're calling things in apocalypse, what does that what does that really mean to somebody. Simon Taylor: Yeah I know it sounds a little melodramatic doesn't doesn't it, when you use the word. But what we're really doing is we're calling attention to what we consider to be the most significant problem in data today, which is what we call the data silo sprawl problem. So today, um, you know, there are about 30,000 SaaS services in the US alone. I think the average mid-market company has their data in no less than 212 of them. So just think about that for a sec. You're not keeping your data in an old mainframe. It's not in some tape drives. It's not in even a, you know, some VMs on prem. And it's not even in the cloud. Your data is spread out across hundreds and hundreds of different data silos because of SaaS. You know, and if you just step back, if you're a CIO or you're a VP of infrastructure or a director of technology and you start to count. Oh, yeah, in finance, we've got these these apps, and in marketing we're using these apps and sales. We're using these apps. It suddenly becomes very apparent how many different locations your data is in. And the problem, Mike, is we talk about in the book is what we call the great myth in data protection, which is this idea that our data is protected in SaaS when all of these SaaS vendors actually use the shared responsibility model, which very clearly states that it's your data, it's yours to protect. Unfortunately, until haiku, there was no one, no one out there to protect it. And it really was leaving customers stranded. And that lack of protection has left 80% of the world's data without any kind of recovery mechanism whatsoever. That's the apocalypse we're talking about. Mike Matchett: And I know people who've watched, uh, some of my segments here on Truth in IT and on Small World Big Data. In the past, we've talked about, you know, hey, office 365, Salesforce, uh, in a in a couple just a couple little things. Maybe it's, uh, uh, you know, your, you know, Jenkins kind of environment or something like that. But, um, that's been the extent of it really. And when you say 200, when you say 212, I'm counting. Well, that was 2 or 3. Uh, and and what about the other 200? How do we, how do we back up those 200 things? They all have different APIs, right? There's all come from 100%. Simon Taylor: Well, so this is the problem. The first problem is that people don't even know what they have. Right. So so we've gone out and spoken with CIOs and they've said things like, oh yeah, I think we use office 365. That's about it. Until you say, well, does your marketing team use Marketo? You know, is your finance team using this? Is this team using Tableau? Is there Jira is there, confluence is there, Okta is there Salesforce. And it's sort of like you can watch the color drain from their faces when they suddenly realize, oh my gosh, I'm a victim. I'm in exactly the same situation as all these other people. So the question became, how do you go after protecting 30,000 SaaS apps? Well, if you were to use if you had unlimited time, money and resources, it would take you thousands of years to go out and build all of those. So we said, what if we kind of tear the top off our platform and create a low code development environment in which anybody can actually code integrations for what they've got themselves? In other words, instead of it being haiku coding 30,000 SaaS apps, what if 30,000 SaaS apps coded on haiku? And so we sort of reversed the paradigm. And I'll tell you, when we started this process, Mike, there were only five SaaS services that were protected by anybody in the Gartner Magic Quadrant. Five out of 30,000, not 5%. Five, five. Yeah. You know, and today, haiku protects 60. Eight of those services in our marketplace. So, you know, within just 12 months, haiku has literally gone from being, you know, a player with an idea about how to innovate the space to the most prolific protector of multi-cloud environments and SaaS environments in the world. Uh, and that's been a very, very exciting shift for haiku and a great evolution for the business. Mike Matchett: I mean, Simon, as good as 68 is, you told me you told us before 212 plus, right? Yes. And so if I'm doing my math, that's 144 more SaaS apps on average that are still hanging out there. What does someone do in that situation? Simon Taylor: Yeah, it's a great question. I think the first thing we did is we introduced something called our graph and what our graph does. It's a free tool. Everyone should try it. And you just turn it on and it will read your entire environment and let you visualize where all of your shadow. It is it's going to tell you where all of your apps are. All your databases are on prem, public, cloud and SaaS. But it doesn't stop there. Mike. What it does is it actually goes a step further and tells you what's protected out of the box, what's unprotected, and what's protected with haiku. If you see something that's unprotected, you can click on that app. It pulls up the marketplace. If it's not in the marketplace, you can request haiku or a partner to build that integration. Those integrations can be built in as little as three days. Uh, that was until, uh, our big announcement yesterday, which is that haiku now offers the world's first generative AI development platform, which will actually allow you to build those integrations in as little as 36 minutes. So, you know, it's built on the anthropic model. We've enriched anthropic, and you can now go in directly and leverage artificial intelligence to expedite the delivery of those integrations and really future proof and expand, you know, the data protection support that haiku provides. Mike Matchett: All right. So large language models, generative AI certainly a hot topic certainly hearing about that a lot of that. Now uh, what I, what I like about what you're saying is that it's not it's not a broad hey chat about anything kind of use case. It's a, it's it's a very focused use case where you can pre provide the context and sort of pre charge that model uh to know about how to do your low code environment, to know about what the APIs on the cloud platform. Yes. Um, to know what, what needs to be done with in general with that. And then I assume then what, what the what the user then brings along and says, hey, I'm using this SaaS application. It has this API. Yes. And so, you know, how do I how do I get there from here and that. And that's where it can provide a lot of help. Simon Taylor: You know, two things I would sort of mention here that I think the first is that, you know, we were very committed to delivering AI technology that was going to provide true customer value. You know, I think, you know, when we looked at the marketplace, we saw a lot of copilots. We saw a lot of, you know, you know, we call the website crawlers, you know, bot chat bots that will bring you to a website or a web page, uh, from the home page that to us did not really ring our bell. It didn't feel like that was adding true innovation and true customer value. The real value of the haiku AI initiative with anthropic is that because we had a low code development platform, because we offered this level of extensibility across our cloud, um, what we were able to do is very simply point the LM, enrich the LM with all of haikus, very specific security postures and data management and data protection principles, and then guide it to support the delta between the customer SaaS service and the instructions that were required to be filled out in the low code development platform. And so really what we're doing is we were able to create a finite problem set that the LM could expedite the development of. If you were just to sit there with an LM and say, go code a backup product for this SaaS service, good luck. Um, you know, we're not there yet with AI. And frankly, I think human supervision is still critical, you know, to supporting a safe and, um, enterprise quality, you know, AI tooling set. But I think the fact that we had this low code development platform and the fact we were able to enrich something as powerful as anthropic, uh, and it's Claude to model, you know, with all of haikus, sort of posture and data protection knowledge. That's the real secret sauce. We had the right level of problem in exactly the right place, so that I could create true customer value. And that value I can quantify for you, Mike, because we can. Now, we've brought the development time for a singular integration from three days to 36 minutes. And I think that is probably the first time I've seen an infrastructure software. You know, I really driving customer value, right? Mike Matchett: Because if I've got if I've got, you know, 200 unprotected SaaS apps and I'm spending, you know, three days to a week on each one of them, I'm just a never ending task. I'm never I'm probably never get them. But if I can, if I can get some smart people working on it and knock down ten of those a day, uh, within a couple of months, I can cover my whole my whole footprint there. I think you would call it your whole real estate. Right? Simon Taylor: I think that's it. I mean, we've been talking a lot internally about the fact that what we're really doing is de-risking innovation. Because if you think about the way companies innovate, they add tooling, whether it's on prem and public cloud or in SaaS, they add specific services that let them accomplish very specific problem sets and company goals. The problem is that innovation comes with a price, which is the inherent risk that's driven up every time we add a new tool to the stack. And so I think by adding haiku and our cloud and visualizing where that risk is, where the threats are, and then adding the AI development framework, we're actually de-risking that innovation for customers. And I think it's really the first time that data protection has gone beyond just backup and restore to really say, no, no, no, we're going to help you throw your arms back around your entire organization and we're going to make it safe for you. Mike Matchett: Yeah, I'm just thinking that CIO story you started telling where, you know, you ask them, you know, what SaaS service is this? We have three, uh, because now you can now you can help them understand where all the SaaS services are, what's covered and not covered on day zero. Uh, and then provide them some tooling and some ability to make progress against that in a very realistic and aggressive manner. Uh, to the point where, you know, if that's an initiative for a quarter, um, you can check it off and you can say like, hey, we went from not having any of the SaaS data services covered to having them all covered. I mean, 100%. Yeah, that's exactly right. Simon Taylor: That's exactly. And again, I think we should always look at AI as needing supervision. You know, I think the, uh, the CEO of anthropic said it really well, she said that, you know, that AI today is like, uh, somebody who's two years out of college, right? It is not a graduate student. It is not an executive. Uh, and I think if we look at AI through that framework, you know, it's almost like what haiku has been able to do. You know, if I think about my ten year old son taking him bowling, it's almost like we were able to fill the gutters with those, you know, long sleeves. Sure, sure that he can get the ball all the way to the, you know, hit the had to hit the hit the pins at the end. And I think that's really what we've been able to do. The low code development platform and the certification process that customers go through to certify these integrations, ensures our products integrity and ensures that the customer is safe. And ultimately, for us, that is the most critical aspect of running a data protection company. Mike Matchett: Yeah, so it's AI here, but it's AI to accelerate the how long it takes to turn the crank. But you're not saying, hey, we've used AI to do automatic validation, security and all the rest of it. You're still 100% involved, still heavily involved in that. Uh, which is very cool. I think this is a great application of, of generative AI for folks who are looking and saying, like, what is all this AI ops really about? And where can I really stick a toe in? It's like, this is a this sounds like a great use case. Uh, that's got all the, you know, rails in place to make it work. Well, uh, so, so if Simon, if someone wants to learn more about this, if they want to dig a little bit deeper, maybe, maybe, uh, uh, to have a conversation, what would you point them at? Simon Taylor: Just go to y Q comm. That's dot haiku comm. Mike Matchett: Ak khou.com. You can't miss it. Right. Simon Taylor: It's right there. That's right. Mike Matchett: Uh, all right. Thanks. Uh, well, thanks for talking. Uh, at some point, I want to get back around with you and find out what happens with this apocalypse. Did it did we avoid it or avoid it? Or is it still going to keep coming? Uh, but, uh, we'll have I'd be happy. Simon Taylor: To do that, Mike. And I certainly hope we avoid it. I think we're, you know, we've now got 4000 customers avoiding it with haiku. Uh, and I certainly think that number continues to grow and increase. Uh, always a pleasure to be here to chat with you. Mike Matchett: All right, take care, folks. Check it out. Bye.
