Democratizing Cloud Security: Innovative Risk Assessment – Avi Shua, Orca Security

This is Textron tv. Hey everyone, welcome back here to Techstrong tv. I am happy to be joined by my next guest.

He hasn't been on in a few months. I don't know. He's been busy, we've been busy, whatever.

But it's a pleasure to have him back. It's my friend Shaw. Avi is the Chief Innovation Officer, as well as a co-founder at Orca Security.

Uh, you know, we've been following Orca, I think just about since they launched, and it's, it's a pleasure to see the success they've had in their growth. We've worked with them for years. Um, but not all of you may be familiar with Avi and Orca.

So why don't we start there. Avi, my friend, welcome to Tech Drunk tv. How are you?

Great. Thanks so much for having me. Pleasure, Avi.

Look, I don't mean to embarrass you or put pressure on you, but I gave you a big buildup. Share with the audience a little bit about your journey and a little bit about Orca security, if you don't mind. Sure.

So, uh, you can put me as the cliche of a cybersecurity founder, so to speak. So I've been obsessed with cybersecurity since I was a teenager. Try to see how to break into stuff and then I to protect against breaking into them, which is much harder job.

Usually at age of 18, I joined the Israeli intelligence a unit 8,200, spent a decade and then spent another decade in Checkpoint done doing various roles from a team leader to bring the chief technologist in the last four years. And four and a half years ago, I co-founded Orca Security with the goal to build a cloud security platform that actually works. And I know it might sound simple, but when we found that the company, there is no solution that actually works for cloud security.

And by works, I mean the simple thing, something that can cover all of the environment without the need to deploy agent and to have project that take month that can handle any type of risk, whether it's a vulnerability, misconfiguration identity, so so forth that is contextual based that can find the risk that actually matter, don't give you 10,000 issues that you don't care. And last but not least, it can be actually consumed. And you can summarize all of that to something that actually works.

And we've been fortunate to grow very rapidly. As you mentioned, we have hundreds of customers, anything from small, uh, startups all the way to Fortune 50, like SS a p and others. And we are proud to be protecting them.

Absolutely. And look, Orco was a, it was a security unicorn during Covid, right? And things really kind of took off for you with the move of companies to the cloud with digital transformation, et cetera.

But today it's all ai, right? Uh, AI consumes, it was consuming 40, 50% of all of the articles on our sites across the board, so much so that we created text strong ai just to kind of put all the AI stuff there. And it's still consuming all of our attention and, you know, sucking a lot of oxygen out of the room.

But you know what, there's a lot there. It's not just hype. AI really does have the promise to, to really make huge changes, Avi, as I'm, you're aware.

I know, but like everything else, you know, there's the reality and there, there's the hype. It's not putting security professionals at a business. It's not, it's not making cloud security autonomous, right?

Where, where it's, you know, self-driving and any more than we have autonomous vehicles other than in San Francisco, of course, that are allowed to drive, you know, without people behind the wheel. Um, but there's a lot of good it can do, right? And I, every company is being asked, Hey, what, what, what are you doing with ai?

What, what generative ai, how are you leveraging it? I know, you know, you're the chief innovation officer there. Orca just made a a pretty big announcement on this.

Share with us what's Orca doing with generator of AI here to, to help in cloud security? Sure. So first, I think you're completely correct with the self-driving vehicle analogy.

Uh, you can't just say, let's take AI through any problem in, in it, and it'll solve it in a way that you actually want to use that. This is not a situation, but it can do a lot of good to improve the lives and the economy of these technologies. One of the main things that we have been focused since we started the company is to make it something that can be actually used, that can be consumable.

And the reality is that most organizations don't have enough manpower to protect the cloud environment. Cybersecurity professors are extremely expensive. There isn't a lot of people that know cloud super well and everyone is using it more and more.

And you can't just add more and more people to protect. And we've been continuously using AI to make it slightly easier to make it more self-driving. So essentially the people that you have can do dramatically more with less.

I can give some examples. We created AI-driven remediation instruction. So given a set of issues, we use AI to come with very easy to understand recommendation.

We don't expect people to autoplay them. We still need an operator to review them in most cases and approve them, but it makes it much easier. We created AI driven identity and access management optimizer.

Think of a situation that you have a lot of entities, each one of them need permission to different set of, uh, resources. The AI driven mechanism can look at all of them and come with a recommendation for policies that make sense. I'll use an analogy if not from the IT world, but think of an organization, let's say a bank that have a lot of personas that each one of them need different permissions.

It can look at all the behavior and come with recommendation for the set of permissions. And now what we are, uh, uh, releasing is a mechanism that can translate free text speech to queries on a unified data model. So two years ago, if you wanted to ask something like, uh, in fact, let's take four years ago, if you wanted to understand something like tell me any, uh, container that I have that is exposed to the internet that run a certain type of workload.

You had to infuse data from several different systems and you never knew the answer for that. 'cause you needed to get agents data, C S P M data for many tools, it would be probably a project that, to answer it, we made sure that you can have all of this data in a unified data models. So Orca can answer all of these question about your cloud environment, anything from which systems have logged for J to what users exist in my environment that they haven't logged in in 90 days and have access to my sensitive databases.

But to an, to, uh, get the answer you needed to phrase the question in a way that the system understand, and sometimes it can take some time. What we created is we taught a model. We took a generative AI check G P t, uh, variant, and we taught him our models.

You can essentially ask question in free text something like, show me any virtual machine that exists in my environment that of health data and they will build a query for you. It's as simple as that. So instead of trying to build a query, you can do it by, uh, using AI and it saves time and it's, it has just another feature that make it closer to self-driving.

So is what is what's going on here is Aker is gathering all of the data about your cloud environment and your variables. Everything now is then taking that and putting it into sort of like a vector database into an L L M that sits on top of the chat G P T, so that when you input into chat G P T, it's actually pulling the, the information from that l l m or is Oracle putting all of that information in into some SQL type of database, right. Or some database.

And then when you ask the question, Aker is taking that question and converting it to a, a standards S Q L query into this database that aker's maintaining. And if, and if we're giving company secrets away, don't answer. No.

So you're not, uh, giving, it's more, it's closer to the second. So it's not sql. We have a unified data model with its own proprietary query language and essentially one, we don't share your sensitive data with any third party.

This is Oh, That's what I was getting at, right? Yeah. Okay.

So, uh, if, so essentially we take it, we store it within your environment, and when you come with a question, we use AI to translate that question to the internal model. But we do, we did teach the l l m the way that we structure things, the way that we store them, so you can build the best query for you. Got it.

Okay. That makes sense to me. Now, this is available now or are you just announcing it, but it'll be available?

What, what's the sort story? So it's currently in beta and we are releasing it this week. In fact, I believe that by the time that, uh, this, uh, episode, this will be released.

So it's a public beta for everyone. It's not like a limited release beta or something. Exactly.

Okay, very cool. Um, another question is, so is this like an extra charge, an add on to the basic orca? Or all or customers kinda get this?

No, it's if it'll be available to all Oracle customers. I believe that all of these features that are, I'll say the driver assist features are important to small businesses and enterprises alike. Absolutely.

So my next question, Avi, you know, and you may not know the answer to this, but look, this is a relative first step into generate using generative ai in this case with author and, and, uh, cloud asset search, cloud security. It's not costing anyone their job, it's making people's lives easier, right? But where do we go from here on that, do you think?

Where do, where, you know, this is, do we wait for the other shoe to drop? If so, when, what, how, So I think AI have done a lot of good things for a lot of industries. And generative AI is just one of the, and the toolset, what we found is generative AI is extremely good in anything, whichever human interface.

So whether outputting data to humans or translating data for humans, it's not good at other items. If you're trying to find anomalies, you don't want to use generative AI for that. There's other tools, but it's really good with anything language.

And yes, there's been a huge progress and we leverage this progress and like everyone knows these, uh, things enable to do more with less. I don't think it'll, it probably will take some people out of the job. I don't think in the cybersecurity industry.

Think in the cybersecurity industry, it'll allow people to do more with less. And there is such a deep shortage of people that I don't think that we will see people losing jobs over that. I I did, I agree with you.

I mean, look, I've been in, I've been in technology a long time and I've been in security almost as long. I've never seen a technology and they've all come out and said, oh, this is going to take away people's jobs. I've never seen one that made less jobs.

They make more jobs. There will be more jobs here, um, as a result of this and that I'm, I'm certain of. So I, I don't worry about that.

I don't fall into that camp at all. Um, Avi you know what we didn't even give, and my fault, we didn't even give people where they can go on the web to, to get more information on this. Sure.

So the nice thing about ORCA is that it's as easy to deploy as connecting your account in three steps and getting the visibility for your environment so you can get more data. com, get a free trial, connect to a W s Azure G C P environment and see the risks actually literal in minutes. Absolutely.

So just go right to the website. It, it's on there. Um, it sounds great, Avi.

I mean, this is a great, I think this is the, a great use of, of what a, of what generative AI does best, right? As you said, it's that great front end to, to go into it. It's not meant as you, you know, to do vulnerabilities and, and stuff like that.

Um, I, I do think, you know, we had a hackathon here a couple weeks ago at the office with some pretty bright people using AI for the operational, for operations and, and security. I do think we're gonna see more and more of it helping security professionals be more effective. But this is a, this is a tangible, put your finger on it, you know, first step, great use of, of the technology within the context of your existing product, right?

You're not out here throwing a hail Mary pass trying to make something all new. This is, Hey, you've got our product. This is, and this is a better way, or a good way of using this technology to be more effective.

And I think those are the kinds of solutions that win in the end. That's what people want. I totally agree.

I'll share another example. We spoke recently with many customers, and we heard that they're spending so much time just pinpointing the origin of an issue. So they see an alert in the environment, a vulnerability and misconfiguration and the SOC analysts that see it, or the, the vulnerability team or the defender essentially that sees it can spend hours to find the developer that caused that issue.

And it might sound a lot, but in large organization, it takes time. So we created a mechanism that can automatically find the person who created it. It's not, I'll say a technological breakthrough, something that, no, you know, that, uh, changed the industry.

But if it saves hours every time of the person, it can do more. And this is at the end of the day, what matters for our customers. Agreed.

Agreed. Avi, we're outta time. Thank you so much for coming on.

As I said, very exciting, very exciting to see real tangible generative AI solutions built in here. Um, keep us posted and we'll have you back on soon. Definitely.

Thank you so much. Thank you. Avi Hua, co-founder, chief innovation officer at Orca Security out with their, uh, ORCA security, AI driven cloud asset search users.

Uh, chat, G P T interface. Check it out. It's, it's available right now.

We're gonna take a break. We'll be right back here on Textron.