Elastic N.V. ($ESTC)

Good afternoon, everyone. It's Blair Abernethy software analyst with Rosenblatt. Thanks for joining us. Today, we're happy to have Eric Prengel with us, who's the Global Vice President of Finance for Elastic. And we're going to spend the next 45 minutes walking through Elastic's current business and how they're positioning themselves for this AI rapidly changing AI features that we have entered into in the last couple of years. Welcome, Eric.

Thanks for having me, Blair. Really appreciate you having me and everybody who's joining. Looking forward to the conversation.

So, maybe before I ask the first question, I just want to suggest to anyone who might have a question for me to -- actually, if you could just e-mail it to me at babernethy.rvlt.com, and I will pull up and filter that into our conversation that would be appreciated.

So Eric, why don't we just start with -- at a high level, some context for people that may not have looked at Elastic in the recent recent past, just give an overview of sort of where the business is at today, the market or the problems that you guys are addressing and then a little bit about your background as well.

Yes. So why don't I start with the business, and I'll end with my background and if I forget to anything just double click on anything, please ask me. So Elastic is the world's leading data platform, which is specifically designed to index store and analyze massive volumes of messy unstructured data. That's what it was originally created it for -- and what has happened over time is it's been evolved to be used in 3 core use cases. So the first 1 is search. And historically, Elastic was something that you could build search applications on top of. search has turned into AI pretty quickly. AI has kind of been used as a search problem. So now Elastic is increasingly the platform that people are building AI applications on top of -- we offer vector databases. We have inference services where you can bring large language models into Elastic. We have multiple models to Gen AI through a reranker model, an embedding model. There's hybrid search capabilities that we have that allow people to build AI applications on top of Elastic. And that has been a big driver of growth for our business. It's expanding the TAM in search, where people are using Elastic to build applications. That's 1 of the core pillars of Elastic. The next pillar, I'd say, is probably observability, where the ability to index store and analyze these massive volumes of data has been applied to first logging, where you can index all log data and parse through it and use it really for observability. And that's where we've got a ton of strength in logging -- and we also have metrics and traces, which is APM capabilities at Elastic. But long means kind of the core functionality. And I'm sure we'll talk about it later, but there's a lot going on with Metrics Elastic that we're super excited about and that's a new place that we're headed. And then the third capability is security. And it's starting with SIM security, the SIM capabilities. And that has evolved. We also have endpoint functionality and some other stuff. And that's effectively a logging use case where people are ingesting this log data and finding security issues with in a elongated parsing through it. And that's a place where AI has been tremendously valuable to Elastic. And where we're able to -- we've created a bunch of things in the security business where we were first to market. We were first to have an AI assistant in our business. We were first to have an Agentic SOC, which is this attack discovery capability that we have. which sort of auto makes the passing through of data and alerts and triage is it down that the SOC analyst doesn't have to do that manually. And so that's been a tremendous benefit that we've seen in securing our security business is just doing really, really well. And so taking a step back, and AI has changed our business in the last 3 years. It's been pretty massive. So for search, AI has expanded the TAM where we used to have a narrower TAM with AI. We're now being used to build applications. That's been huge -- and then with observability and security, I think what's happened is it has AI has necessarily expanded the TAM. But because we have AI capabilities built natively into our platform. We've been able to be first to market with a lot of AI functionality. And we've got a road map which is really opulent to our customer base around AI. And so that's helped us become more differentiated. That's a high-level big picture elastic story. I don't know, Blair, do you want me to go into a little bit kind of the update on the quarter? Would that be helpful or?

Sure. Actually, given your role as Global Vice President in Finance, absolutely give a snapshot of your last quarter, which was -- and maybe just.

I'll give you a little context on what I do. So here at Elastic, I run FP&A. I have strategic finance. I run IR. I have procurement, have enterprise data. So all of the business-facing finance functions are kind of where I spend my time and I partnered with Naval and a lot of that part of the business. And so I do have a pretty good sense of how the business has been performing. -- we had a really strong Q4 to end the year. We ended the year with CRPO from 20%, Mario growing 28%. We just had a tremendous amount of long-term commitments that we saw coming from our customers. And when I saw that data, I said, "Well, this is remarkable. Let's double-click into it. Is it that are we doing more discounting, and that's what happened. And I looked into it, and we're not doing more discounting than we have historically. Was it more 5-year deals or some outlier 7-year deals that we think out signed that RPO. And no, it's not the case. It's all 1-year deals and 3-year deals, which is the same tenure -- same duration that our deals have historically been structured to have. So our business was just performing really tremendously well from a commitment perspective, and we're very happy to see the way that the business was performing as we exited the year. So that was really positive and a lot of excitement here at Elastic around how our business is doing.

Yes. And Eric, you also -- just for some of those that may not be familiar last fall, October time frame, you sort of gave a longer-term or medium-term view on where you guys are going. Maybe if you want to just recap that and sort of your progress towards those goals. .

Yes. And we've actually -- we've updated those levels. We've taken those goals even higher than they were before, but we had an Analyst Day in October of 2025. And at that Analyst Day, we walked through what we're seeing in the business around AI, the traction that we're seeing with AI. And as we take a step back and look at our business, the growth that we think that's going to drive. And so we were comfortable to tell people that we think that we're going to get a 20% plus on our sales-led subscription business. And sales-led subscription is the total subscription revenue, excluding the monthly cloud, which is more of a month-to-month business. And so we feel very comfortable that we're going to get to that 20% sales-led subscription growth. And we also talked about a Rule of 40 type metric whereby our Rule 40 is going to be 40-plus -- and we talked about operating margin being in excess of 20% non-net operating margin being in excess of 20% in the intermediate term and the intermediate term for us FY '29. And we updated that a little bit at the Q4 earnings, where an involvement has said that we actually now have the saying over 25% the non-GAAP operating margin, we now expect that we're going to be in the 25% range for non-GAAP operating margin in that time frame. So a lot of positivity and excitement around where the business is going. .

Yes. And a significant amount of your growth in the next few years seems to be coming or you believe will be coming from AI? And maybe we can shift back to that for a minute. On your core search business where you guys came from where you have literally millions of downloads of the open-source surgeon, but also thousands of customers -- how is -- how are you approaching AI in your installed base in the search side? What are you providing for them? What's the value that Elastic can deliver and keep yourselves relevant or sticky within that search installed base?

Yes. So we've got a lot of functionality. We've got a vector database that they use to build the effication on top of. We've got reanchor models. We've got embedding models through GenAI. We've also got inferences of service where people can bring LLM usage in inference usage from different LLMs into our platform, and so they can use all of that as well as through hybrid search to build applications on top of. And there's a couple of reasons as to why we really think that we're doing so well. So one of the issues is data gravity. People aren't going to put all that data into the LLM, -- they need to bring the LLM to the data instead of the reverse. And so Elastic allows them to do that. Elastic has retrievable capabilities that bring the data that's relevant to a question to the LLM and we had a blog post recently where an application without elastic -- sorry, an application with Elastic was 70% more efficient in terms of token usage than 1 that wasn't using Elastic, because the last thing is using that retrieval augmented generation, bringing the data that's relevant to the query to the L or bringing the specific details to the LLM and making the much more efficient. So .

Really using traditional search capabilities and you're highly indexed -- your rapid scalable, highly scalable indexing to find the information of the relevant information instead of waste, not wasting, but tokens are expenses. .

It makes faster, it makes it more efficient. It makes it more accurate, all at a lower cost. So if you're building an application with Elastic, it's tremendously impactful and can reduce the cost can increase the efficiency can increase the velocity of the application that you're building. And so that's the kind of thing that Elastic can do for people who are building applications. The other thing just to give you a little bit more context, and this is an example that we've used historically. But if inside of your inside of Rosenblatt, there was some sort of search thing. And you want -- you went to -- you had an internal to answer questions like an internal like Chat-driven interface. And you said, "Hey, I just broke my arm I hope you how do we do a -- but hey, I just broke my arm, I need to go to the doctor, what can I do? They're going to -- through Elastic can provide the underlying capabilities that say, "Hey, this is Blair Abernethy. He lives in this location. He has this medical coverage you can say, Blair, you should call this person, you should do this person because they're in network. It can give you all of that because it has context around Blair, Elastic brings that Blair context to the bigger picture LOM. And so you're able to build applications like that with Elastic. .

Yes. Then you guys were talking about that a lot last year, context engineering was starting to become -- the word was coming up. And I guess -- I mean, part of the challenge here is that unstructured data is 80-plus percent of the 85% of the data out there, right? And it's -- a lot of it's unindexed. So are your customers recognizing this context engineering capability that you guys are bringing? .

Absolutely. And that's why we're winning these use cases, and that's why we're helping them to be more efficient. We're just we're seeing a ton of customers who need this context. Like there's a lot of stuff we do. We've got an AI solution which help -- we've got customers in a big document company. And what they do, we allow them to acquire across their data, we bring the context to them. So it's not just that LLM has to read every single document. Elastic finds a document and then the LLM can read the document. Think of -- if you said, "Hey, I'm looking for contracts that have been signed with big automakers, and I want to know ex about them." You asking on that, they have to look at every single contract in the database. Think how expensive and time consuming that is. With Elastic, Elastic brings the 3 contracts that are with automakers -- it points them to what they need. And then the LLM can analyze that data and use all the intelligence that LMs are great. They're super powerful. They've changed the way that we work. but they don't have the ability to parse out which 1 they should be looking for. And Elastic can bring that to the table, make them much more efficient, much faster. You're getting better results at a lower cost. -- kind of how -- and so customers are wildly excited for that. .

Eric, from a go-to-market standpoint, how are you guys taking this to your search installed base? Like and how are you -- I mean, there's a lot of as we were talking before the call, there's so much innovation happening some many changes happening so rapidly, how does Elastic on a break above the noise of it and say, "Hey, we have -- we can solve some of these issues for you"? .

Yes. We go to our customers. We understand what issues they're having. We work with them to tell them where we can help them. We've put a lot more capacity into our go-to-market recently. It's -- I know that in Q1 of '21, there were some issues where I go to market we resegmented the go-to-market, and we kind of narrowed the scope of reps who are covering higher-value customers and rework to go to market there. There were some issues with that. for that first quarter. But since then, the go-to-market has really been humming where we see tremendous success. We've seen a big uptick in productivity. And because of the uptick in productivity -- or strong uptick in productivity that we talked about at the Financial Analyst Day, because of that strength that we've seen in productivity, we put more capacity into the model to really go out and chase the business. So we've been super excited for that. And the way we're doing it is we've got some specialists who are search-focused who can give people expertise around AI and bring that to bear, and we're going and talking to customers and seeing what they need. And it's -- I'd say that -- because of the strength that we have in AI, we've been able to up level the kinds of conversations that we're having where we're coming in because -- I mean, Asia board level initiative. And so the most senior levels at the company are engaging with Elastic because of that. And even if they're not necessarily purchasing AI capabilities today, we're seeing customers who are making commitments and longer-term commitments with Elastic because of the AI road map that we have and because of our ability to partner with them over the long term, to really be the AI road map that they are excited for.

Is -- maybe we can shift a little over to observability because it's been 1 of your observability solutions and security solutions has been important drivers for you in the last 5 years. Where are you at on observability today? It's -- there are some -- a number of competitors in the market? And then maybe talk a little bit about sort of AI within your products or what you're doing to stay relevant in that space? .

Yes. And so the biggest thing for us with observability is there are kind of 3 pillars of observability. There's logging, there's APM or traces and there's metrics or infrastructure monitoring. Historically, Elastic has been super strong in logging. We've had -- if you have a big logging use case or a log use case, Elastic and maybe 1 other vendor you call. We are are in that space, and we feel super confident about our ability to get into most of the long opportunities and to win most of the long opportunities. We've got an incredibly attractive price-to-value ratio, we were price to technology ratio, price to innovation ratio, where we kind of best-in-class there, and we can bring best in class to bear at a reasonable price. With that being said, metrics is probably -- or infrastructure monitoring, the same thing. It's just -- it's probably been the fastest growing part of the observability market. And it's been growing so fast because a lot of this cloud infrastructure that's been coming online becoming a bigger and bigger part of the technology ecosystem. And we probably haven't had a best-in-class solution in metrics historically. We've got a competitive solution, but it's not best-in-class or it hasn't been. And we spent a lot of time in FY '26, reworking that solution, making it more compelling, more competitive. We made it meaningfully more efficient by putting a column or data store on the back end, where we've reduced the footprint of storage for metrics significantly by using a different storage ecosystem than we use for a logging. And because of that, when you benchmark us against our competitors, we're now meaningfully more efficient to them, have a much better price-to-value ratio. And so we think the big opportunity for us in observability and potentially 1 of the big opportunities for us all solutions is this metric opportunity, where we now think that we can compete and win against the leaders in that space in a way that we couldn't previously. And it's probably going to manifest itself with us initially rolling out as an adamant play where we try and sell metrics to customers who are already using us for login. But ultimately, we think that we can lead with metrics and go on with big. And it's something, Blair that a lot of the leaders in the field have been asking for, and we're excited for. And so when our team got on stage, and our sales kickoff at the beginning of May and told everybody in the field that we were going to have this best-in-class metrics offering and walk through some of the benefit. There was so much excitement. It's been something that our field team has been dying to sell. And it was funny, somebody said, "Oh, are you going to have any incentive for the field team to sell metrics"? I don't know was laughed out loud because they've been wanting metrics so badly. It's like you gave them exactly what they wanted. They don't need incentives, something they've been wanting to sell forever, and they've been selling dense, but now with this reworked architecture with the back end with the efficiencies, it's something that they can sell so much better than they've been able to sell for. So yes, there's AI capability. We have an AI assistant. We have other capabilities that we have that are AI-centric in observability. But I think what's really going to move the needle is the Matrix capabilities that we're bringing to bear. And like that's going to be just a complete game changer for our business and for the metrics business overall.

So that's GA now. .

That is being sold. It's in production. I think if it's not GA, it will be GA very soon. I think just in terms of -- we also just because I don't want to short change out that we're super excited about metrics, but we also have an AI SRE where we're going to have bring a lot of AI to SRE and like we're instead of relying on engineers to manually do all this work. It's something where you're going to have AI that can automatically organize these unstructured logs, discover patterns figure out root causes, and this is going to make things so much more effective similar to what we do with attack discovery in security, which has been just a -- the security business, I can't talk about that, but we've seen so much happening positive with the security business. I think that that's the AI capabilities that we're bringing in for the SRE are going to really change the game for us. But domestic is probably the bigger opportunity even than that AI opportunity, to be honest with you.

Yes. So I mean, you've had a lot of traction with logs in the last couple of years. So I guess a lot of those customers are probably coming up for renewal over the next 24 months. So that's a perfect opportunity for you to sell metrics.

The field could not be more excited about Metrix. So hopefully, we'll see some impact from that in coming quarters.

This is not 5 years out -- is this some...

We hear the thing. So we just bought it to -- we're just going to market in Q1. People have to get enabled on it and they have to go and sell it, they do when they get the deal signed and then it takes time turn to revenue. So if you think about the revenue outlook that we have for FY '27, we're not assuming a massive revenue impact metrics. I think it's hopefully going to help us win commitments, get a deal signed in the business, but I don't think that it's going to be something where we're going to say, "Oh, look, a meaningful force revenue is specifically attributable to metrics I think it's FY '28 that the revenue component really starts to have a bigger impact. I think right now, it's going to be bookings is probably going to be the bigger deal. .

Got it. Got it. Before we go to security, just with metrics help us to understand how your pricing works for observability for logs and so forth. .

Yes. Right now, the way that we price the solution is definitely based on ingest and compute. And so that's going to continue to be how we think about things. There might be some tweaks around the edges, but that's typically how we price things at Elastic. And there's going to be further serves offering. It's slightly different in how we price things versus ECH, and it's a little bit more based on outcomes versus -- well, outcomes. It's not just based on storage and compute. There's other complexities in there, but it's typically storage and compute is the best way to think about how we price and it's consumption units. So we have consumption-based pricing, and that's going to continue to be the cloud. It's going to consume less storage because we're not storing it much more efficiently with metrics. So it's going to compress the pricing that people see and make us much more able to compete with other vendors. .

Right, right. And the -- just on your observability installed base, is it -- maybe help us understand how much of that is sort of elastic cloud versus on-prem or self-hosted -- is it mostly cloud? .

Well, I think first and foremost, important to realize that the self-managed is not necessarily on-prem. A lot of people are taking that self-managed business and deploying licenses in their cloud ecosystems. And so they're not -- if you think about self-managed versus -- if you think about self-manage versus cloud. It's not legacy as self-managed mode is cloud. We're seeing some tenement use cases happening in self-managed environments. -- where customers are taking our licenses and deploying them in their AWS, GCP or Azure instances and running very modern AI use cases, but where they're buying from us in a self-managed way. To your other question, how do we think of -- I'd say that there's not a big distinction where 1 solution skews more towards the cloud than others. I'd say U.S. public sector tends to be where we see more self-managed business and a large 1 way I'm looking for highly regulated industries tend to have more self-managed, but I don't think that necessarily aligns with a specific solution where I'd say like, "Oh, observability is very self-managed and search is very cloud-centric." I don't think that's the case. And even as you think about search and the motion we've seen towards more and more AI. I don't think even that has been more cloud, for example. We're seeing a lot of customers who are buying self-managed licenses. -- and then using those managed licenses to deploy AI solutions in their own cloud environments or in their own on-premise instances. .

Right. Right. And that's part of your part of your value prop is that you have the flexibility to allow customers to do it any way to run it any way they want. Let's just shift to security. Maybe go back and just sort of high level. So here's what our security offering, where we came from. And then we can talk about some of the new stuff like the AI site live engineering and AI assistance and what progress you've made there on the product side. .

Yes, I'm very excited too. And so logging turn into SIM because that's based on log data and using that engine to part data from a security perspective and to identify threats and anomalies. That's what the SIM solution does. We've also added XDR into the business. And we -- that came through the Endgame acquisition where we have end points. And we actually -- 1 of the $220 million plus deals that we signed in Q2 of FY '26. One of those deals was very XDR focus, which is a tremendous win for us that we're very excited about. And we've also got each builder on top of the security solution, agent builders or a solution that allows you to take the data that's inside of our ecosystem and engage with it in the same way that you would AI chat interface, if you can just ask questions directly on the data build different outputs with that data and do a lot of products. So do a lot of automation with the data that's in our ecosystem already. So you can create agents that are utilizing the data directly with Elastic in a way that we're pretty excited about. And so those are 2 things where we were first to market on. And then also attack discovery is something that's been huge for us from a security perspective. And what intact discovery does is it allows you to -- without needing a ton of manual work to parse through the different alerts that you're receiving and figure out which ones we should be spending more time on, which ones are less serious and just sort of automating that triage to prioritize the alert to spend time with is what a tax gain does. And it allows people to reduce the amount of time that SOC analyst spend doing that manually. And it's been a tremendously positive thing where we've seen a lot of traction around it in our business and a lot of excitement. So A lot of people are using it, but there's also a lot of people who are excited about even if they're not ready to yet, I think it's part of their road map. And so when they're making their purchasing decision, it's something that comes to bear. And I think that security overall, we've just seen tremendous momentum in. And as we think about FY '26 as a whole and how we exited '26, a lot of the momentum that we're seeing in our business was security. And I think the reason that -- it used to be that our reps were saying like, "Oh when we're in a deal we do well, but it's hard for us to get in deals." I think that Elastic Security is getting stronger and stronger. We're getting in more and more deals. And when we're in deals, we're really winning them. I ran it to 1 of our 3 geo leaders at our sales kickoff. I was talking to him for a while. And he was just -- I was talking about the whole business, and he was just so excited about what's going on for our security business. And just the capabilities that we're bringing, our ability to win when we get it deals because of our technological differentiation. He's like, "look, my job is to get us in deals, the product team's job is give me a product that I can sell once I get in deals and the product team is 100% delivered." We're ahead of our competitors on AI. We're excited about our ability to win we're head-to-head with competitors. Maybe people can -- there's -- they can discount their product to a point that it's untenable. But generally, we're really excited about our ability to get in deals and then to win those deals once we're in the deals. Just it was really powerful to hear his excitement when I was with him at our sales kickoff.

Are you -- is Elastics -- if you think about your marketing and your profiles, you've been around for a long time, and you've had extremely well known on the search side of things. How are you raising your profile? Where are you investing to raise your profile on the security side?

Yes. I think there's stuff that we're doing for branding. I think there's stuff that we're doing in terms of going to conferences, raising our visibility. We've got search specialists, AEs is kind of a field goal and SA is an overlay where they bring to bear specific domain knowledge and are able to engage in those security conversations and have relationships with the appropriate security buyers in a way that we might not have had a year or 2 ago before we had that functionality. And so just from a security perspective, we've made a lot of investment. I think A lot of it also is word of mouth and the fact that our solutions are winning and that they're being trusted by some of the biggest companies in the world. The CISA deal, we haven't talked about it yet, but that was a huge deal for us. I mean we signed a $26 million deal with the U.S. federal government through the CISA agency, where they're providing SIEM as a service to the civilian agencies in the U.S. public sector. And that just shows a massive trust in Elastic. And these are big, chunky deployments that are really important to agencies that are critical to the U.S. government, and they're trusting Elastic to deploy SIEM as a Service. And so stuff like that is just -- it's a great -- that kind of commitment from the U.S. government is phenomenal and speaks to the capabilities that our security offering has.

Yes. And that's -- you also just in the recent months, got FedRAMP high authorization, which is great. Is this...

Long ago, right? Then we got FedRAMP high. Yes, very exciting.

So that, combined with this CISA deal is pretty important. How about other non-U.S. federal state level and maybe talk a little more about internationally, how is your profile? How is business trending over there?

It's doing great. I mean the system just start to stand out that I mentioned it, and I'm so excited by it. the traction we saw there was game-changing. But internationally, the business continues to do really well with the public sector internationally at the state level, at -- with the Department of War and the different aspects of the non-civilian U.S. public sector, we've seen a ton of strength. I think that from a public sector perspective, we're very happy with the way the business is performing. CIS is the bellwether and it's actually put a pretty big uptick in terms of the cloud component of the public sector business, which isn't typically as much cloud and CIS has like change that a little bit. But across the board, we're seeing strength in the public sector. There's a lot of excitement there.

Eric, how -- maybe if you can comment from Elastic's perspective on Agentic AI adoption within your customer bases. What -- how would you sort of characterize what are they doing? What have you seen that's working effectively? Were they leveraging your platform? And maybe just give us some sense of where we are on this -- on the customer journey in enterprise.

Yes. So we talked a little bit about the document search that we're seeing, but there's also an AI-native customer who we have on the music generation side and what they're doing is they're using Elastic search for vector search and beddings, a lot of detailed model stuff. And what that allows them to do is they can index over 1 billion songs and then kind of create this retrieval and matching and generate Lurex to create this -- I think we see a lot of these AI generated songs. They use Elastic for that, and that's really a great use for us. You're also seeing customers using us for Agentic. There's a global supply chain they're leveraging us for the vector search features and embedding to embed AI into their products. And they've got a Agentic product that isn't released yet, but it's on the way up. I'd say Agentic is very top of mind. It's still in the phase of becoming more prevalent, I think we're going to see more and more that gets deployed in the real world at agenetic over the next 12, 24 months, but it's still earlier days for Agentic than it is for some of the other AI use cases is my sense. But we're definitely seeing more and more of us with this agent stock with some of the stuff that I just talked about on this music company and then some of the stuff that we're doing in observability where people can use agent builder to engage with their data in a way and then ultimately, the goal is then to be able to create agents that can act in -- can automate some of the things that needed to be done manually in an Agentic capacity. I think that's the direction that we're going and then we're seeing more and more customers start to if they're not deploying it, some are deploying it, but even if they're not deploying it, they're starting to work with it, they're starting to think through how they get value from this Agentic approach to AI.

Interesting. And the -- your products have traditionally been pretty technical and for customers. And so it's always hard to -- or difficult to have the staff that the teams available within an organization to really understand how to get the most out of the platform, the components of your platform. Is AI helping customers, helping to make Elastic easier to use. .

Oh, absolutely. It makes a huge difference. I mean, migrations off of other solutions on to Elastic are much easier with AI, tooling. The AI assistant enables people to do things in Elastic that needs to be very manual and can be totally automated or simplified now. AI is definitely a big tool in terms of how much more efficient it is for people to get up and running on Elastic and to reduce some of the complexity unusable .

Yes. Yes. And then what about internally and maybe just bring us up to speed on what -- how Elastic has been leveraging AI within your operations? .

Yes. In terms of our operations, I mean, there's a lot that we're doing in terms of engineering. So in R&D, we're heavily utilizing AI coding tools, all the ones you think about from OpenAI, Anthropic from Google, and that's definitely accelerating the amount of code that we're deploying that we're writing, we're creating much more efficient ways. In terms of go-to-market, the field is using some AI modules to automate seller onboarding and to make things a little faster. We're using it for feature testing and engineering. And in terms of in my function, there's a lot that we're doing in AI in the finance function, where there are things that used to be very manual and we're experimenting with ways to automate them. I'm not going to get too into detail with that because some of them were still figuring out. But they kind of sold in the last 3 months -- not going to say 6 months in the last 3 months that we started to work on building in my organization are going to create meaningful efficiencies for the business over time, and we're already starting to see a little bit of it, but there's things that we -- I mean, how much easier it is to build some of these things. There's something that we had an idea for. And six months ago, we were talking about buying a solution, and we actually built our own pretty efficiently for something that we're going to be using that we need to get our auditors comfortable, but we're pretty excited about it.

And then in the R&D side, we're in your finance hat. Maybe talk a little bit about the cost of token cost, which is everyone sort of grappling with versus the the productivity and the value you're getting out of these things. .

I track our spend on a daily basis.

I'm worried about it, right? .

I have it daily, right? I'm very not worried about it because I know it, but I know it because it's important. So I don't know what you saying that cyclical statement there is, but it's something that we think about that we track very actively. We want to make sure that engineers are using AI because it's going to give them -- you can't create a tax engineer without AI and we want to make sure that they're using it efficiently that they're using the right models that spend isn't getting out of control, but also that we are getting real spend into our AI usage. So a tracking on a daily basis, I know how it's being used, and we want to make sure it's being used efficiently and then it's creating value in our organization. But it's very much a cost that we are aware of Cognizant and planning around. And we're excited about, because I don't want to -- it's -- I don't -- the last thing I want to do, you have to balance it. Like this is going to change the way that people work. This is -- I don't want people thinking about, "Oh my god, I can't spend this $10 on when they can be so much more efficient with AI is out." It's an engineer who is really strong with AI, if they we're told, hey, you can only use this many dollars of AI, they would effectively stop working once they ran up to that limit. Because it's like -- if we drive from San Francisco to L.A. and we went out of gas halfway through, we wouldn't get out and walk. We figure out a way to get more gas in the car. And I think that that's what AI can be in terms of force multiplier is the same way a car can move you from here from an 500 miles so much faster than Forsen. I think that AI versus manual coding in the same way that no 1 is going to a short amount of time, it's going to be pretty rare that you see people manually coding things is my belief. And that's not an elastic believe that's an Eric belief based on conversations I had with our team here and some of the things that I've seen that some of the engineers are able to do that is pretty remarkable.

I've been hearing similar comments this week from several different people at the conferences I'm attending. So yes, there's -- we're heading towards a point of where coders may not be coding all that much at all. .

I think 1 of our developers have said something to me that really resonated. If we were in a car factory, -- and you said, "Oh, this isn't working fast enough. Like you have all these machines and putting the cars. You wouldn't go down there with a hammer on nails and like work alongside the machines. That would be really inefficient. And his belief is that at some point, pretty soon, we're going to get to a world where that's where coding looks like. Where you're setting up the factory for the agents to develop code, but you're not there writing code next to them.

Right. You're setting this strategy and the direction and the future descriptions, if you will, and the processes, but then let it run and let the the Q&A.

And the testing and the code review, you're having all of those being done by different agents who have a set of parameters that they're working under and are able to utilize in order to really drive the best results. .

Has Elastic looked -- sorry, has -- have these coating tools changed your hiring practices or views at all within the engineering group. .

I think that across Elastic, we need to be thoughtful around what our organization is going to look like going forward. I think in FY '27, you'll still see us add head count on a net basis, but it might be in different organizations, than we would have without the benefit of some of these AI innovations. And it might -- the shape of the organization, the places where it grows is going to be different than it would have otherwise because there will be some new efficiencies that will be introduced because of AI, and we're going to use AI as many places as we can, but -- for example, the selling function, you still need people talking to people in order to sell software. And I don't think that's going to change in the immediate term. And so we'll see head count continue to flow into the field work, obviously, but there will be changes in terms of how Elastic growth looks in this fiscal year.

I guess in your SMB side of your business and your monthly business, are you utilizing or applying AI there to help drive up consumption? .

I think it's something that there are more opportunities for us to do more with AI in that part of -- in the low touch part of the business, and that's something that we're looking into and exploring and there are different ways that we can get benefit from AI for sure, in that sort of the self-service business. .

Okay. Great. Maybe just as we're getting close to the end of our time here, just to step back a bit and just sort of remind us in terms of M&A technologies, how are you looking at the M&A world right now, just given the fact that things are changing so rapidly, is it -- are you more focused on internal development? Or are you still looking for -- because you bought a couple of interesting pieces in the last year. So maybe just to frame that up for us. .

Yes, I don't think there's going to be a step-wise change in our M&A strategy in the immediate term. I think that we've had this perspective that if things can pull forward our road map if they can bring innovation to the business that we wouldn't have otherwise or that would take us much longer. We're open to doing something with M&A -- and that doesn't preclude us from doing a certain size or scale or what have you, if it's accretive to the business in terms of the growth in the tractor our business and the direct to moving in, which is getting more and more AI set more and more identic -- like if there's something in the -- in that space that's going to support that, I think we're very open to adding that technology to our business. And so because I know we're short on time. So .

Yes. Appreciate it, Eric. And it's great. It's great to see the business performing so well in the last couple of quarters and it seems like the opportunity window is -- has been getting bigger and continues to get bigger for Elastic. So thanks for sharing your thoughts and your time with us today. .

Thanks for having me, Blair and I really appreciate the conversation.

Okay. Great.