On this episode of The Drill Down Insider, a Six Five Media production, host Cory Johnson is joined by Elastic‘s Ashutosh Kulkarni, CEO, for a conversation on the transformative potential of Generative AI and the Vectorization of Search that Elastic is pioneering.
Their discussion covers:
- The impact of Generative AI on search technologies and user experiences
- Elastic’s approach to integrating vector search with conventional search methodologies
- Future trends in AI and search technologies that businesses should watch out for
- Practical applications and case studies of Elastic’s search technologies in various industries
- The challenges and opportunities in developing AI-driven search solutions
Learn more at Elastic.
Watch the video below, and be sure to subscribe to our YouTube channel, so you never miss an episode.
Or listen to the audio here:
Disclaimer: The Drill Down Insider is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.
Transcript:
Cory Johnson: Generative AI. Vectorization of search. Elastic is the company in the right place at the right time. That’s why I talked to Ash Kulkarni, the CEO. Welcome to Drill Down Insider, where we drill down on the business stories behind the right companies at the right place at the right time. Hey, welcome to The Drill Down podcast. We are joined right now by Ash Kulkarni, the CEO of Elastic, an exciting company and exciting space. You’ve got search, you’ve got Generative AI, and you’ve got a growing company that’s had a really exciting year in the stock market. Some of it not all that much fun, some of it fantastic. Actually, outperformed the S&P 500 last I looked for the year, but in a bumpy fashion. So Ash, I thought what we’d talk about is, you’ve made some really significant changes in your technology this year and even in the last week. Let’s talk first about the last week and becoming more of an open source company.
Ash Kulkarni: Yeah, Cory, thanks for the question. Our ethos has always been open source. The company grew out of this incredibly popular open source project, Elasticsearch and our community has been a big source of our strength. The big announcement that we made last week was that we introduced AGPL as a license option for the free part of our technology, which until now has been licensed under SSPL license, as well as Elastic license v2 which was a very permissive license that we created.
And the whole idea behind adding AGPL to that was, AGPL is OSI compliant, which means that in all the places where our community, our users want to go and find a OSI approved open source license and get technologies that they can use, not just for search and log analytics and all of these areas that we’ve been in, but more importantly, vector search, vector databases. All the new kinds of capabilities that they’re looking to build, that’s what’s so exciting. So I’m thrilled about this. In some ways we’ve always been open source at the heart of things, but now we are able to proudly proclaim that we are open source again. And that was the big announcement.
Cory Johnson: How would that open up more opportunities for you? How do you see that playing out?
Ash Kulkarni: Yeah, absolutely. At the end of the day, our biggest strength has been our community. Every single customer that I ever visit and meet… I might be talking to the CIO, I might be talking to the CISO, but that customer, that company already has users of Elasticsearch developers that started using Elasticsearch years ago, they have built all kinds of interesting applications on Elasticsearch, and they might’ve been using the free version of our product, which has some basic capabilities, but they know us, they love us. And that’s been a big source of our strength. It’s that bottom-up sales motion.
So now as you’re seeing this Generative AI wave, there’s so much work going on around semantic search, around Retrieval Augmented Generation, every new application, every new business process is being automated with the use of large language models and vector databases are going to be at the heart of it. That’s what we really want to capture. So what I fully expect to happen is, more and more users are going to discover Elasticsearch as the vector database of choice. And as that happens, our market opportunity just grows, our total addressable market just expands. So this is a multi-year play for us, but it’s in a lot of ways coming back to our roots. And that’s what makes it so exciting.
Cory Johnson: Let’s talk about those two mega trends of vectorization and the Generative AI that results. Going back to Vectorization, talk about how that kind of folds into the history of this company and is a natural growth of it. Who knew we would be in this age like that?
Ash Kulkarni: Yeah, it’s fascinating. The company has always been about search. And where search is most relevant is when you’re dealing with unstructured data. Data that does not nicely fit into a schema, into a traditional database, old school database like a MongoDB or an Oracle. It doesn’t matter if it’s SQL or NoSQL, but all of those kinds of databases are really well-designed to deal with data that has good schema associated with it. When you’re talking about unstructured data…
Cory Johnson: Structure
Ash Kulkarni: Yeah, exactly. When you’re talking about unstructured data like log files where every developer decides what they want to put into it, when you’re talking about Word documents, PDFs, images, videos, there is no easy way to put them into a database. That’s really where search comes in, because we apply this notion of relevance to help you figure out what is most appropriate for the query, for question that you’re asking of that data. That’s how the company started. And one of the most amazing things that Shay, our co-founder and CTO did, was he made it possible for you to put any and all kinds of data into Elasticsearch. That was one of the key differentiators.
And what Vectorization is now letting you do is, take that one step forward where you can now vectorize just about any kind of data. Audio, video, graphs, images, doesn’t matter. You can vectorize it and in vector space now look for similarities. So you can basically say, and if I’m looking for cat images or cats jumping off tables, I want to look for something that looks like this. And you can search across billions of images and find just the images that have the exact same kinds of patterns. And that’s what’s exciting. So this is now bringing the power of search, something mean the relevance of search, something that we’ve always done with the intelligence of AI and now applying it to any and all kinds of data. And to me, that just opens up a whole big space of automating every business process, where you inherently have to deal with unstructured data in manual ways.
Cory Johnson: And it seems that it also allows customers who have lots of unstructured data, who suddenly have the capability to generate value from that data and to figure out, not have to have people searching it, but have a machine searching through it to find those connecting vectors that might draw value and use out of all the piles of data these companies are sitting on.
Ash Kulkarni: Yeah, and the use cases, Cory, are just unbelievable. Maybe I can talk about a few of these things that we have seen playing out.
Cory Johnson: Yeah, I’ve got one in mind that you mentioned in the conference call. Please give me what you got. Maybe you’ll to guess the same one.
Ash Kulkarni: No, there’s so many. So you have… The ones that I mentioned in the conference call, one was an E-discovery one. There’s a customer that is using us effectively for looking for case law examples that are similar to what they might be, that was the one working on…
Cory Johnson: That was the one. Giant law firm. So explain to me the problem that they have, how they were doing it before and how they’re able to do it now, thanks to Elastic.
Ash Kulkarni: So when you think about what matters when you’re working on a legal case, like legal precedence is extremely important. What are the core decisions that have been made on similar cases in the past? And what that means is, sifting through volumes and volumes of case law that’s out there, seeing the judgments that have been passed, and all of this is typically done manually. Because the only technique that worked until now was to effectively do textual search. So look for text patterns, exactly matching words and Elasticsearch could do that incredibly fast, and we were being used for doing that fast.
But when you’re looking for concepts, I’ll give you an example. You are looking for a contentious marriage case, a marriage settlement. A contentious marriage settlement, these three words back to back might not be the exact text that you’re looking for. But you’re looking for case law cases that might be related where there were some issues that were being debated back and forth. Now you’re talking about semantic search. You’re talking about looking for the meaning of words as opposed to the exact matches. And that’s what semantic search, that’s what Vectorization helps you do. I’ll give you some more examples. When you’re talking about…
Cory Johnson: Just stick with that one. So the law firm acquires your tools and your engines and applies it to what? And then they get what result?
Ash Kulkarni: So they suck in all the data feeds that they’re getting about case law that’s been decided. Every day there are courts around the country that are making judgments, that are passing judgments. All of that gets recorded, it gets transcribed, it gets published. They’re pulling in all of that case law on a daily basis. You’re talking about millions and millions of documents coming in, and they pull that into Elasticsearch. The traditional model then was, when they would search for something, they would be searching for exact text matches. Now when they search for something, they search for concepts. They’re searching for, “Is there anything similar to my case in concept?” And Elasticsearch now because of its vector database, is able to get them very precise matches that say, “Yep, here’s where there was a judgment”, and it can summarize all the information for them, which is such a powerful way to automate what used to take hours in the past.
Cory Johnson: So fascinating. It also strikes me that obviously these all kinds of other industries, I interrupted you before you give me another one.
Ash Kulkarni: Yeah, just take customer support as an example. I call somebody looking for information on how to configure my router or how to configure some new application that I’ve purchased. When you go through the mechanics of what a support engineer has to do in that particular scenario, they’re usually reading through many, many documents. They’re looking through information about the product itself, known defects, other information that they might have received in the past, related to that particular product. But you also have situations where that particular customer might have logged some issues in the past. That customer might be unhappy, because they’ve had to deal with that same issue over and over again. How do you quickly process all of that information and have it at the fingertips of the support engineer, so they can service you better? That’s been the holy grail forever. And the big problem in automating that process is, much of it is built on unstructured information.
It’s all built on data that’s in document form. It’s not in the database, so you can’t access it very quickly and you can’t get exactly the right meanings out of it. Now that process is being improved. In the past, I’ve talked about financial analysts using Elasticsearch and our vector database capabilities to provide the equivalent of a Bloomberg feed for you, related to your particular portfolio. Everything that’s happening in the world right now, that pertains to your portfolio and how it can affect your portfolio. Doesn’t matter if it’s exchange, currency fluctuations, if it’s the price of oil, if it’s the movement in and out of the tech sector of money, all of these things might affect your portfolio. How do they give you a very specific personalized set of information and answer your questions? Those are the things that have always needed human beings to sift through massive amounts of data. Unstructured, in documents and now we can automate all of that, which is so fascinating.
Cory Johnson: Also seems like I keep a pretty detailed financial model of Elastic, and it starts to explain some of the things that you see in the results you put up. So by my estimations, you saw 18% revenue growth, you put up a really good quarter, you lowered the guidance that talked about some issues that I’m going to get to. But so as you’re describing these technological possibilities, what it suggests to me is that the customers that have the most data and have only been dipping their toe into what Elastic can do, will benefit from spending a lot more from Elastic, as opposed to adding new, small customers in great numbers, you really have the opportunity to go a lot deeper with the customers you’ve got. And you saw that in the results.
You saw that the growth in customers of over a $100,000 in spend was three times the growth in customers outright and customer accounts. That makes sense now, and it makes sense why you would restructure the sales team to go after the big opportunities and make the big opportunities as big as they could be, instead of getting as many opportunities as there could be.
Ash Kulkarni: Cory, you got it. We are always going to get more customers that come to us because of the openness of our platform. But at the end of the day, the biggest opportunity for us in our business is to make sure that we really, not just land, but expand and consolidate. Get our customers to start using us for more and more use case. Whether it’s around observability or security or all these variants of search, around Generative AI. And Generative AI is such a huge opportunity for us. So the thing that we’ve been focused on is, how do we make sure that from a sales perspective, we are organized to truly go deep and get wallet share appropriate to all the value that we bring to our customers? That needs more customer intimacy. That means we have to organize our sales teams in such a way that they can service these customers better. And the change that we did at the beginning of this year was designed to do that. Where we had a misstep is, in the account transitions themselves. And we didn’t do that as smoothly as we could have.
And this is an area that as soon as we realized what was going on, we immediately all got into it. And the rigor that we’ve been applying is already showing the kinds of writing of the situation as I would’ve wanted to see. So I’m pretty confident this is going to be a couple of quarters before we are back to the kind of execution that we’ve always been capable of demonstrating and we have demonstrated in the past. But more importantly, with this new sales model, I feel even more confident that, as we go forward, the way we will be closer to our customers will allow us to expand even better, improve our net expansion rates, and that’ll be great for the long term for the company. And look, 18% growth in overall revenue in Q1, 30% growth in Elastic Cloud, we grew our operating income, we put up 11% in operating margins. And as I look forward, our story is both about growth and profitability. We are demonstrating that we can expand on both, and that’s really the story here. So I’m laser focused on this and feel pretty confident about the future.
Cory Johnson: Yeah, it’s got to make it crazy as a CEO to figure out how the opportunity is changing, where the technology puck is going to go, so you can be there when it gets there in the Gretzky metaphor, but also then to have the customers not get it right at first. And it’s why your job really is hard.
Ash Kulkarni: The way I think about it, I am humble about the fact that we had a mistake. There’s no denying that. I want make sure that I’m always, we are always doing the right things for our customers, for investors, for all our stakeholders. So there’s no doubt in my mind about it. But I’ll tell you, I’m extremely determined because we know the opportunity ahead of us. And my confidence and conviction in the long term is absolutely rock solid, just because of everything that I’m seeing. We are winning in these GenAI deals. Now we have over 1300 customers that are using us just in the cloud for these GenAI use cases. Over 200 customers that are in the 100K plus cohort that are using us for GenAI. So we are seeing that expand in a very nice and material way. We saw the acceleration of our search business, on the back of this GenAI momentum.
Even in a quarter where, like I said, the total number of sales commitments didn’t quite meet up to our expectation. And that just shows you that the core is expanding. The core is incredibly strong with people using us for GenAI, which is now really reigniting search. So lots of exciting things happening and you can’t help but be excited. Now when you have a misstep on the execution piece, the most important thing is, understanding what happened, being focused on fixing it with tremendous urgency and determination, and then demonstrating that. And as we do that, I know that our investors are going to be very happy, and that’s my goal to demonstrate that.
Cory Johnson: Ash Kulkarni is the CEO of Elastic. Ash, thanks for your time.
Ash Kulkarni: Thank you very much.
Speaker: The Drill Down Insider is brought to you by The Futurum Group. Where analysts, researchers, advisors, content creators, and marketing experts help business leaders anticipate and understand shifts in their industries and build strategies to leverage disruptive innovation. With deep analysis, Futurum Group’s extensive industry experience delivers reliable research and data, thought leadership and actionable advice to help you with your strategy and go-to-market efforts. Futurum Group.
Cory Johnson: All right, now to the Drill-Down bite. The one number that tells us a whole lot, we want to take a look at Elastic and I talked about how the large customers are growing at three times the rate as these small customers. Even I can do the math because 1% growth rate was the average customer growth, but 3% growth rate in the customers paying over a $100,000 dollars for Elastic, really shows that expansion strategy towards the big customers, who can pay and getting them to pay, at Elastic. All right, thanks for listening to Drill-Down. I’m Cory Johnson. Subscribe to us here on YouTube, check me out at X @CoryTV and on TikTok and Instagram @drilldownpod. Follow us, like us, dislike us if you must. Leave us some comments. Tell us what you want to think because we want to hear from you.
Author Information
Cory Johnson is the Futurum’s Chief Market Strategist and the host of the Drill Down podcast.
His peripatetic career has seen him in prominent roles as a hedge fund portfolio manager and investor, technology journalist and broadcaster. Fundamentally he’s an entrepreneur -- helping to start media companies such as TheStreet.com, the Industry Standard, Slam (the world’s best-selling basketball magazine) and Vibe. He was CNBC’s first Silicon Valley correspondent and later helped create the TV show Bloomberg West for Bloomberg TV and the radio show and podcast Bloomberg Advantage. He was a senior executive at the blockchain startup Ripple, a portfolio manager for Kingsford Capital and a principal at the Forensic Research Group.
Johnson is also an advisor to Braintrust, C3.ai, Prolly AI, Provenance Bio, Stringr and serves as a delegate to the Episcopal Diocese of California.
