Cribl Copilot with Dritan Bitincka

This is Textron tv. Hey, everyone. Welcome back here to another Textron TV interview.

You know, I've got a, he's a first timer though. We've covered the company since I think the day they launched almost, or the day they announced. I've never had the pleasure of having him on the show before.

Let me introduce you to Ton Ska. Uh, ton is the co-founder and head of product at cribble, and that's, uh, C-R-I-B-L. If you're not familiar with Cribble, standby, we'll get you familiar with Cribble.

But first Ton welcome and, and thanks for being on Textron. Thank you for having me. It's great.

So you are the co-founder, one of the co-founders. Clint Sharp is the CEO co-founder. We've had the pleasure of having Clint done more than several times over the years, and we know Clint was a former, I, I think he was CTO or something.

Uh, executive at Splunk left. Splunk founded Cribble a better approach to some of the same problems Splunk was looking to, to help with, with, with security and stuff. But let's hear your story a little bit.

Now. We got the other side. There's always three sides, right?

But let's hear your side. Fantastic. Yeah.

So, uh, so Clint and I, and my other co-founder, uh, Lyon, we all met at Splunk actually. Um, I was, uh, I was there for six years and so was Clint. Uh, prior to Splunk, I was a practitioner.

Um, Clint was a practitioner in it as well. Um, we spent, uh, over six years working together, uh, in various areas. Uh, he was a product manager.

He was building products, um, uh, ion, the other co-founder, he was a, a engineering architect. He was actually a hands-on keyboard. Um, and I was out in the field deploying, um, Splunk, um, software and services to all the customers.

Uh, after about six years, and this is early 2017, we all decided to, um, um, to leave and, and try our own luck. Uh, and we formed, uh, actually, we, we didn't form ble then. We formed something entirely different.

We managed to successfully fail for a year and a half before we pivoted to what is scribble now. Uh, and then in the summer of 2018, we, uh, you know, we started kibble. We sprinted for about three months.

We released our first software, uh, log stream, as we called it back then. We called it Stream now, uh, in October of 2018. Um, and then about six weeks later, we had, um, a call out of a small company, um, Australia, of all the places they're, Hey, we're interested in purchasing your software.

We met you at the trade show. Uh, and we were not prepared to, you know, accept the transaction. We had no pricing.

Uh, and I remember I was in New York, you know, driving across town, talking to c Clinton and sort of arguing and debating about how much to charge. Um, and we, we just, um, we sold $5,000 worth of enterprise software for unlimited volume. We never made that mistake again.

Yeah, you know what? So look, and then I, I've founded Co-founded or founded five, six companies over the years going back to the nineties. I love these stories, right?

Because that's exactly how it happened. Someone wants to buy the software, they're gonna gimme $5,000. We don't even know if it works.

Take the five grand and then you look back and say, I gave away the house for $5,000. Um, but that's, but that's, yeah. That's what found, you know, founders do, and that's how startups work, right.

Until you get kind of that act together, what a great story. Absolutely. Of course, cripples come a long way since then, right?

You guys, uh, recently announced, uh, a, I think it was Series E Round, is that right? Correct. Series E round of financing.

And, well, I don't wanna steal your thunder. You tell us about Series E round finance. Sure.

Um, yeah. So, um, so a, a few weeks ago we closed our, uh, series E, uh, that was led by, uh, uh, gv, Google, uh, Google Ventures. It was one of their largest investments ever.

Um, and, uh, it's kind of interesting 'cause in the era of this sort of massive founding rounds, uh, of, uh, mostly AI companies right now, uh, it showcases that we've been stable and the importance of the, of the enterprise software, the problems that we're solving are real, uh, some of the problems that our customers actually have, um, out categorize them as, um, basic in terms of what pain points customers have. But you know, they are, so the problems that they, the customers are, are dealing with, deal with, um, tedious amount and voluminous amount of data. And so they just need, uh, the support from, uh, software that is built by practitioners for practitioners.

Um, and, uh, agreed. Yeah. So, um, it, it's, it's very interesting.

Uh, you know, we got to know the, the GV team over, um, you know, a few months prior to the round. Uh, we were super excited to get, uh, Mike McBride as part of our, our board as well. Um, he's been an incredible, um, uh, field person.

Obviously he's led, um, companies in the past, like, like, uh, GitLab, he was a CRO at GitLab there as well. He brings in a wealth of experience, um, and we're going to lean on that for, for advice and, um, and all the subsequent, um, sort of deals and, and all the great things that we have in the, in the hopper for that. Absolutely.

Just to add some color, and if I got it wrong, please correct me. So this series e valued cribble at like three and a half billion, right? Correct.

Yeah. Uh, super excited about that. 5 million, which is A huge, I mean, in today's market, you know, during covid we was crazy.

Numbers were being thrown around, and we knew they were, everyone knew they were all a little crazy. But a lot of those have come back to earth in a bad way now, right? People aren't raising up rounds.

Um, AB Absolutely. Yeah. Yep.

So that's quite an accomplishment. Congratulations to you and Clint and the whole Crile team. That's, uh, that's great.

And of course, having Google in your corner doesn't hurt or doesn't suck, as they say, right? And, um, yeah. Good, good for you guys.

Good for you guys. Um, you mentioned ai. Yep.

And it is sucking a lot of oxygen out of the air. I mean, here at Techron, we Yep. As much as we try not to be too AI heavy in our stories and our coverage Yep.

Everybody's talking about ai. What are we supposed to cover? Like, make stuff up?

I mean, you know, and we use AI to make it up, so we're still using ai. So, you know, but I think there's a little irony here. You guys come from Splunk, you know, Cisco announced in the, in the middle of all this AI spending and all this AI interest.

Cisco laid off about 8,000 people the other day. A lot of friends of mine who've been in Cisco security years. And, um, one of the reasons they gave was that they're gonna invest that money into AI initiatives.

So basically replacing, you know, some of the people with ai, Hey, there's a lot of good people out on the street. You guys got some money, could be a great opportunity for you. But with that being said, you guys have your own AI initiatives under underway.

Let's hear about that a little bit. Yeah, absolutely. Thank you for, for that question.

It's, it's actually a topic that we keep discussing, obviously, uh, internally quite a lot, as you said. Um, it, it's all over the place. Uh, lately.

One of the things that we do well here at cribble is, um, we try to solve problems from first principles. Uh, this is something that we've been, um, sort of systematic since day one. And, and by that, what we mean is we go and ask customers and ask our field folks about actual problems that they have.

We try to understand the why they're trying to solve it. Uh, and then we build the how on top of it. We're try to do the same thing with ai.

I mean, in the, to your point in this era of, um, AI in your face or AI all over the place, um, generally what I'm feeling when I talk to customers, uh, is this sense of AI fatigue. Just yesterday, for example, I was one, um, I had a lunch with one of the largest financial institutions in New York for an hour and a half. Uh, the topic of AI only came up once, and I found it interesting.

Um, That is, Yeah, because what we spend the rest of the time was the actual problems that they had were, um, more basic than the AI could actually, um, pretend to solve. And so our, um, our initiative here internally is called Crile Car Pilot. Uh, it's, it's a feature that cuts across the entire Crile suite.

Um, and the way we're positioning it is a, um, is a feature that enhances the capabilities of the operator. So, in other words, it helps the operator that is driving the software to do things faster and get the job to be done quicker. Um, and in this day and age where, uh, the data volume is growing at 28% year over year and budgets are not growing, uh, the need for, uh, getting stuff done faster is, is the fundamental need that most customers have.

Um, and the types of problems that we're solving with that, uh, deal with, with actual, uh, data transformations and data parsing. So in the world today, gen AI typically is, is super great at, at, uh, uh, predicting the next word. So you can generate text very, very quickly, but when it comes to transforming data or parsing data, you have to be super precise and super exact.

So the problem is not particularly the same. Um, and one thing that I tell my product managers is that we have to build the right, because if we don't, we have a reputation risk. And the moment the customers cannot trust the copilot, they will never use it.

And so, uh, so far it's been great. Uh, the adoption has been in incredible. Um, it has short circuit, a lot of conversations with, uh, with our support folks.

For example, uh, customers have questions about solving different problems that they have with our product. Instead of coming in and reaching out and talking to a particular support person, they could just ask AI and get an answer right away. Um, they'll also ask co-pilot to generate code, generate pipelines, generate, uh, query language, uh, strings.

Uh, and it's getting very, very good with it. Uh, and so one point I want to make sure, Alan, that we we get through, uh, to the audience is that we're being very intentional and we're being very careful with how we apply it. Uh, the last thing I want is, you know, AI all over the place, uh, with low confidence, it will do the job to be done.

Right? Absolutely. You know, I was talking to another founder from a different company today, AI security company, surprise.

And, uh, you know, he was saying, look, generative AI isn't the solution for a lot of these security problems. They use more predictive ai, which you wanna call it ML ops or whatever you wanna call it. But predictive ai and generative ai, as you said, is great for a help desk situation where you wanna ask a question about something.

But when you do need precise and you can't afford to fudge, you know, to have an answer that's kind of fudge. You know, some people are gun shy on, on generative ai. Now what helps us with, with these things, well, better training, better LLMs.

Yep. Right? So is that kind of the, the Crile recipe is, are you using some sort of custom LLMs here?

Is it better training? Right? I would imagine so, but Yeah, absolutely.

Um, I mean, you hit the nail there, e exactly. So we, we are, uh, we're using better training data. Um, we actually are in the data path of hundreds of terabytes of data per day, uh, that customers, that's true transit through us.

So, so we get mm-Hmm, uh, we get the, the opportunity to take a look at that data and then use that, uh, to then build, uh, better models and build better, uh, uh, you know, outcomes out of the, the copilot. Um, you know, it's getting, it's getting very interesting when, when you talk to customers when they want those predictive capabilities. Um, and, you know, thank you for, you know, this, you know, sort of explaining the difference between gen and, and predictive ai.

A lot of people seem to mix both of them. Um, oh yeah. And, um, you know, gen AI's very easy.

You just, you know, predict the next token. Off you go. Predictive AI is where, where, where the money actually is in my, in my view.

Um, and when we're building technology, actually we do have a quite a significant amount of features that we're planning to land next year. Um, with respect to that. Um, you know, things like, um, you know, you, you transit terabytes upon terabytes of data through us.

I, I would like to know about the shape of the data, the volume of the data, the content of the data. Has it changed? Does it have a, any seasonality over it?

Um, you know, customers, for example, would like to know when devices that used to emit data no longer emit data. Uh, can we predict that? Can we give users some sort of confident level of confidence that say, Hey, maybe on Wednesdays, uh, you know, the devices go go crazy.

But then on the weekend, you know, they, they sort of go down and what can we do to help customers get that confidence that the software is operating correctly? And, um, the quality of the data is what they actually expect it to be. Um, and, you know, a number of features are coming up, and I'm super excited about all the work that the team is doing here.

Absolutely. It Look you, and if I did this, I apologize. I don't mean to throw a wet blanket on ai, right?

It, there's a reason why everyone's excited. It is game changing potential stuff. Um, and I think you recognize that too, otherwise you wouldn't be putting all these resources into it.

Right? Um, but we do it, it's still, you know, the as fast as we're going, it's, it's still a little too fast to make everyone feel really comfortable. Do, do you know what I mean?

Exactly. And yep. That, that's it.

But nevertheless, I, I wholeheartedly with my whole heart believe this is, you know, game, game changing stuff. Um, Yep. Question for you on the co-pilot, the co-pilot is available now, it's in ga.

Correct? The co-pilot has been available for a few months already. Yes.

So is it, and, and I've been asking all the companies who are coming out with similar kind of, well, not, you know, with co-pilots, basically. Yep. 'cause it's not just Microsoft, right?

It's not a trademark name. Correct. It's called an AI assisted.

Is it an extra charge to use copilot, or does it come with the base package? Um, we have chosen not to charge extra for it. Uh, we believe that making, um, customers get to the job to be done as quickly as possible is part of, uh, the user experience that, that we're building for them.

Um, that's not to say we're never gonna charge for, for it, you know, small amounts over time, but, um, you know, for the foreseeable future, it will just remain, um, you know, part of the, of the entire suite. Um, look, Alan, I think, um, from our perspective, uh, if the customers find the software usable and the experience delightful, um, I, I don't care. We'll build, we'll build it for them.

Even if, even if it costs us the AI training, uh, I want customers to be active in our, um, in our software suite and get value out of it. Mm-Hmm, absolutely. And look, you owe it to your customers really to stay on top of the latest technologies.

I mean, these are real, your not just your customers. Yeah. But the enterprise space, and when we look at security, we look at data analysis and data management and application management.

These are real world problems. You know, I was, I was talking to someone the other day, he said, yeah, no, when I was a vendor, it was all fun and games. Then I went to work for bank, it was bank in New York Mellon.

So that gives you an idea of when it was, and we had an outage and they find us $200 million because of the outage. It got real. He said, that made it really real.

And those are the kind of real world. Yeah. You know, that's what your customers live in day to day.

Like he said, you went for lunch, right? An hour and a half. If they were down an hour and a half, man, they'd be, It would've been painful.

Yeah. Yeah. Well, You, You're right.

You can remember that. Yep. Uh, re real problems, um, you know, being tackled by, by real people.

Um, I mean, I was a practitioner in the past and, uh, and I can empathize with, uh, with, with those, uh, personas, the IT and security operations people, um, very, very tightly. Um, you know, one of the, the, the things that we say internally is we're we're building, uh, software, uh, we're adding delightful features like the copilot, uh, for the personas that for the most part, you know, nobody thanks them. Like, for example, when was the last time you thank, uh, the person that runs your mail server?

You know, it's like, I, I, I don't even know who runs that anymore. You know what I mean? Um, So yeah, well, we're a small company, so I Right.

We, we sa a lot of it, but believe me, I think the person, so we have a guy who does a lot of the it, the mail and the, the slack and, and those things. And as the CEO founder here, I thank him all the time. 'cause I know, right?

And I'll tell you what else, when he can't fix it, I'm the next line of defense. So I'm like our best IT guy. So, yeah.

I believe me. I appreciate it. Yeah.

Uh, it's, it's, yeah, it's a job. It is. It's a job.

Yeah. I, I was just gonna say, I mean, for example, in my career and Clint's career in the past, um, you know, operations, security teams were, you know, folks, you know, in the, in the back office doing all the, all the things to support the business. And, uh, uh, for, for the most part, um, they ended up getting, um, you know, they landed with enterprise software.

It was a socket business. I, our goal right now is to make enterprise software stock less, if you, if you wanna call that. Yeah.

Well, no. You know, when I was running my own security, I co-founded a couple venture backed security companies. That was our motto.

We, we suck less, right? Yeah. Um, and so the, the more features we can add, uh, to support those personas, the, the better our user experience, the more delightful, um, you know, the way that we think the job to be done is, uh, even if it's got AI on it, we'll, we'll get it done.

I mean, we, we fully support these personas and, uh, um, you know, I'll never forget the days where I was a practitioner, not a vendor. And so I can, I can relate to that very closely, Absolutely. To your time.

We're about at, we're probably over time. But hey man, thanks for coming on today. Appreciated the conversation.

Congratulations on the amazing raise and having Google on board with you. Um, and congratulations on the copilot launch, right? This is, this is what customers want and expect today from their enterprise vendors.

So continue success, continued success. Keep us posted on this. 'cause obviously this is something we, we like to cover and hear about.

App. Appreciate the time. A thank you so much.

Yep. Trian, we didn't mention the website. com?

com or cripple? DO io. Yep.

com or io. Check it out. Draton ska.

I hope I didn't mess that. Upton Ska Co-founder head of product here at cribble. We're gonna take a break on Techstrong tv.

We'll be right back in a minute.