Snowflake Inc. (SNOW) Earnings Call Transcript & Summary

So I have been to Melbourne a few times and somebody from Seattle, I've always been impressed Seattle has a strong coffee culture. Melbourne absolutely does as well.

Well, I say the best coffee in Seattle is the coffee that comes from Melbourne. So I've spent a lot of time there over the years also. So I'm sure I'm going to get a lot of timing comments as a result of that statement. Jeff, thanks for taking the time. You are deep into product and what we're doing from a product perspective, day in and day out. Very, very topical for this audience, and I appreciate you taking the time to come and speak to the APJ partner ecosystem. So I'll pass over to you, and I'll see you at the end of the session.

Perfect. Thank you so much. So yes, I want to take just a few minutes here. It's a great segue with the panel you all just heard with Sridhar, talking about the state of the business and the direction of where we're headed. I just want to spend a few minutes here and talk some about some of the pulse of the product pieces and specifically, I want to spend some time thinking about what we're doing around some of the investments around AI. Now Sridhar already did a phenomenal job from a high-level overview of what we're trying to accomplish. So that's going to actually save me a bunch of time. But I want to just focus on 3 big areas of investment that are happening in AI right now. The first one being agents and intelligence, Sridhar was able to spend a good amount of time talking through some of that. The other one is around AI SQL. The last one is around ML platform. Now all of these are just 1 part of a slice of our investments. If I think about Snowflake all up, we're making investments in data analytics, data engineering, apps and collaboration and AI. So I'm just kind of focused on 1 slice of it. But you'll see here in the next few minutes how really Snowflake both directly to customers and through partners, through both partner solutions and partner-assisted deployments is trying to help integrate AI everywhere from ingesting data all the way to getting insights on the data. So again, Sridhar already set up a little bit around what we're hoping to do with agents and intelligence but I'll just double click and I'll even show you this in action here in just a second. So for us, what it really comes about is how we can bring AI on top of all of that proprietary unique data within an organization, within an industry within an ecosystem and really start to accelerate the business transformation and the business insights that happen as a result. You likely are using generative AI in your day-to-day life. But once it comes to your work job once it comes to your enterprise context, oftentimes, the AI without the data becomes very, very less useful, like useless almost. And so bringing that data with the AI is where we see a bunch of potential and a bunch of things coming together. So one of the pieces here is we want to make it very easy to create agents, specifically agents that work on top of your enterprise data. So agents is a very exciting term in the industry right now. You can almost think of it as just how you can start to use these leading industry leading LLMs and models to actually go perform more complex tasks, sometimes entirely autonomously all on top of your data. So I want to give one very critical example when it comes to data agents. And that's with how easy it can be to get access to the data that you need at the right time. If you look at any enterprise and an employee within that enterprise, they're making dozens of decisions every single day. Now sometimes those decisions have huge consequences, potentially millions of dollars are on the line with a different decision and how frequently are those decisions made without access to the right data. And it's very understandable why that happens. Sometimes finding the right data can be very time-consuming, you're navigating through a bunch of dashboards, you're trying to remember what was that report that had the right data. Maybe you're going to find the right report but then you have to slice and dice it to the right scope. There's so much work. Very often, what I have found in the past is that I would just end up e-mailing my data team and being like, hey, I need this specific slice to the data, and I'm kind of waiting for some human manual effort to go and sort that out. So what we want to provide is a better way with bringing agents on top of your data to provide the solution of Snowflake intelligence. So let me just show you quickly what that looks like in just a few minutes. Sridhar already mentioned how he uses this and some of his day-to-day interactions because inside of Snowflake, we have about a dozen of these agents that we're using, while we work at Snowflake to help, I do everything from managing our road map and managing our backlog. Sridhar mentions that he can look at customers and what customers are doing. But here in this demo, I just have a simple agent here, and this is called my Product Insights Agent. So this is an agent that I've connected to a few data sources. This knows a bunch of information about sales that are happening in my organization and also has access to a bunch of data that is unstructured. You could think of like slack conversations, e-mails, calender invites, customer support tickets. Now this is fairly limited, just for the sake of the demo but this list can get as long as you want it to, which says, hey, agent, you're now an expert in all these pieces of the data, this is data that is securely governed and running in Snowflake. So what this means now is if I have a question, how are sales doing over the last 2 weeks by region compared to forecast. I'm just going to ask Snowflake Intelligence. See what happens right away is this agent is immediately looking at my question, and it's figuring out what's the best way to answer it. Now a few important call-outs here. The first one is all of the interactions that are happening here are all happening on top of my secure data. It knows who I am. It understands the data that I have access to. All of that is being enforced. Even the models themselves, the reasoning, the LLMs that are powering this, these are all running out of Snowflake. So my data is never leaving. It's all in my control with role-based policies enforced automatically. Now you'll see here in just a few seconds, I get my answer. It's given me some information here. It's even been smart enough to render this as a table in my case or is a chart, apologies. You can see here one more thing I'll call out even too. You can get nice little things like this green shield. Now I love this green shield any time I work with an agent because this sets me not only did this come up with the right query behind the scenes to answer this question but it pulled from a query that had been certified by my data team. This is a verified query. As Sridhar mentioned, you have a specific definition of things like revenue or a customer. Well, this is pulling from it to help me get the insight to the answer. Now I'll show one more quick thing here before we jump into some of the other product updates. To me, this is great, like this already saved me time from having to dive into or dig through different charts manually. But what I really like and where the power of AI starts to shine is what often happens is I'd look at a chart like this, and I'd quickly say like, hey, what's going on here west, right? All my sales seem to be going about to forecast but something is going on here with the West. And this is where often traditional data exploration tools really struggle. But where Snowflake Intelligence and because these agents are deeply connected to all of the data in Snowflake, instead of sending a slack to my data team, I'm just going to ask this agent tell me why the West is underperforming. That's a very abstract question. There's a bunch of things that could lead to that. And you can actually even see here, I'll go ahead and expand this because this might take a little bit to run. The agents now thinking, okay, why I have inventory data, I have marketing data, I have my trend data, what's the variance like over time. The agent is now thinking, okay, why I have inventory data, I have marketing data, I have my trend data, what's the variance like over time. The agent can actually now go through and go through a bunch of things. You're saying like here, oh, it's looked at marketing scores. It's looking at inventory levels. There's a number of queries and steps that now my agent is exploring the data in the same way an analyst would in my organization. I have now my own personal virtual analyst right here providing insight on my data, coming up with the right context. You can even see here, I'll quickly show here. This has even pulled in some snippets from Slack. Like it looks like I have some Slack conversations that might be relevant to sales in the West. This is giving the agents some clues, some additional context across that business data so that now after a few seconds, I can have some contributing factors, not just on what happened but why it happened and what I should do about it. This is incredibly helpful. I get recommendations, I get root cause analysis, all of this happening empowered through my data inside of Snowflake. So this is one huge area of investment that we're making, which is how we can bring these AI agents on top of your data in Snowflake seamlessly, connected to the business semantics, connected to the role-based access control, all running securely with your data. So agents and intelligence is a big piece of investment. The other one I want to quickly shine a light on in the last few minutes here is AI SQL. Now Snowflake Intelligence is a great way to access insights for anyone in the business, whether you're the CEO, a support engineer, a salesperson, you name it, Snowflake Intelligence is for you. AI SQL is great for the builder. And to me, this is a powerful set of AI tools that you can bring directly into your data workload. I'll just show a very quick example here. So the other week, I was working with a large data set of Snowflake data. It was actually survey results. And as part of the survey that we ran with some of our customers, we had a bunch of verbatim responses. What would you like to see Snowflake doing better? We had thousands of these responses. And I wanted to understand, well, what are the big trends across of all these thousands of responses? Like what are people asking Snowflake to do more of? Now if I wanted to, I could have tried to copy and paste all of those thousands of responses into an LLM and asked it to summarize but it would blow up. It would not work. LLMs can only deal with a certain amount of context at one time. But what I could do and what I did do is I wrote a single line of SQL code that looks very similar to the code here, which said, "Hey, Snowflake, you know how to query huge sets of data, right?" Almost infinite sizes of data Snowflake knows how to query. It now is also powered with these AI functions like AI summarize or AI filter or AI classify. And I said, "Hey, Snowflake, for all of these rows of feedback, I want you to use an LLM to summarize all of the findings and give me a 2-page report of what everybody said for this question." and Then I clicked Run, a single line of SQL that had some AI summarized just like what I've showed. It spanned for about 30 seconds. Snowflake behind the scenes was scaling out my query the same way it normally does. But as it scaled it out, it was weaving in these LLM calls to generate and summarize the components. So after about 30 seconds, I got back a simple 2-page report that said, "Hey, Jeff, for the survey, the thousands of responses, these are the big trends." It was incredible. It was mind-blowing. This truthfully just happened about a month ago. The power of being able to weave in AI to do more complex type of operations is a big area of focus. And we want to make sure that as you're doing it, you're doing it in an easy, efficient way. The last one I'll just mention here quickly before I wrap up is around just general ML platform as well. So you might be using generative AI either to do data processing, as I mentioned, or data insights with intelligence. But sometimes you might want to build more custom models, custom forecasting models, custom next best action models, where we have a full-fledged platform to do this inside of Snowflake or alongside our amazing set of partners. In general, from a product standpoint, how we focus on things is we want to make sure we have a general platform that can take care of the basis to make it easy, connected and secure. But we want to make sure that we integrate phenomenally with partner solutions because we know partners are incredibly good at finding industry-specific, business-specific differentiated solutions. So from ML to AI, you will see us always working to make sure that we're providing a path so that we can continue to work with partners to provide those specialized solutions. So to kind of wrap up on this pulse of the product, I know this is just a small taste, a small pulse. Across the board, you heard Sridhar say the same thing. We want to make sure that this is easy, fully managed infrastructure. You don't have to worry about spinning up the complex pieces needed to run either agents, generative AI or models. All of this is connected across the data in your organization, across the data in your ecosystem, across the data in the marketplace and trusted at its core. Now if you're interested on how you can get started and learning more about these AI features, just a few next steps I'll share is my last slide, which is start today, think about how you can bring everything from unstructured data, slot conversations, e-mail, support tickets, survey results to that structured data. We have use case evaluation workshops where we can help pinpoint some of the high ROA use cases, things that we're seeing across the industry are moving the needle. Then you can even join some of our prototyping workshops where we can go not just talking, actually building, like what would it take to get your very first agent similar to the sales agent that Sridhar uses or the product agent that I just showed. We're excited to work with you, continue to part with you and figure out how we can deliver this goodness and this exciting potential of AI into the hands of our shared customers. So thank you so much for letting me spend some time with you this afternoon or this evening, and that's what I want to share with you all today.

Excellent. Thanks so much, Jeff. So much amazing work going on. I love those demos around Snowflake Intelligence and Sema4 agents. But I also love the fact that ETC, easy trusted connected, or easy connected trusted depending on which way you want to position it. I had no idea you had that slide in there. So pleased to see that, particularly following on from the discussion with Sridhar. Thanks for jumping on. Always a really valuable session around the pulse of the product. So I appreciate you doing that. Just a quick reminder to all of our participants, if you do have any questions, please use the Q&A function at the bottom of the screen. We have a bunch of people sitting here answering those questions. So keep them busy. Okay. For our next session, I am going to be speaking to one of our fantastic customers from Indonesia. But before I do, Hwee Bee, I think there's a short video that you would like to play.

Yes. [Presentation]

Great. Thanks, very much, Hwee Bee. And Sami, welcome to SPN Pulse.

Thank you, Ash. Happy to be here.

Now I was wondering where we had met before, and I just figured it out. Every time I walk into the office in Singapore, I see that video playing. And that's why your face was so recognizable when we connected the other day. So thanks for joining us, and thanks for being an amazing customer of Snowflake.

Happy to be here, Ash, and happy to be a customer for Snowflake as well.

And I should also say congratulations. You were just recognized as one of our Data Driver of the Year awards, Data Executive, I believe.

Thank you so much for that.

So I joke with you the other day that, that means that we use you for all of these presentations moving forward. So you've got to commit to multiple hours every week to do these things.

100%. And like I said, the learning is mutual, right? So I always learn a lot from these sessions. So looking forward to that.

Okay. Good stuff. So Sami, just to get started, I think there's a couple of things that we should clear up. So in that video, we just saw a company called XL Axiata but the title of this session is all about XLSmart. What's going on there?

Yes, yes. So basically, last year, we were at XL Axiata, and earlier this year, just very, very recently in April, we actually completed our merger with the fourth operator here. So that's a premerger used to be called SmartFren. So we merged XL Axiata and SmartFren and produced a new company, which is called XLSmart. So we are now much bigger, much well positioned to fight in the market.

So we need to get Mr. James Butler from our marketing team to come and film a new video with you.

Correct. Correct. Exactly.

Good stuff. So the other thing that I just noticed on there is it said XL Axiata had 57 million subscribers but the notes I have here are a much larger number.

Exactly. Because now we have completed the merger process in this Q2, we had our first quarterly report out as well as a merged entity. So we have now a lot more customers.

82.6 million mobile subscribers if the data I've been given is accurate.

Exactly. That's the correct number.

So considering you serve so many customers and so many people in Indonesia really trust your company, what role does data play both in terms of your growth strategy but just running the business on a day-to-day basis?

Exactly. So basically, data is at the center point of everything we do. And I'm not just saying it as a clichéd word but practically, every day, every moment, whenever we are making a decision from a very strategic decisions all the way to very tactical day-to-day stuff, data is all across -- is used widely all across the organization. I will give you some examples. So Indonesia is a complex geography. So it's a lot of islands and the access to those islands is also difficult as well as we really need to be sure where are we deploying our CapEx because, I mean, telecom itself is a very CapEx-intensive business. And it's a $3 ARPU market. So we need to be very sure that wherever we are deploying our CapEx, the market demand is there. So we use a lot of internal and external data together to build -- and we have built our models that helps us in identifying the geographies where we should expand our network. But not just from a coverage perspective but also a customer experience perspective. So what is the profile of the customers, what kind of devices are they using? And what is the right experience to build because it's very easy to build, for example, 100 Mbps network all across the country but then the key is the profitability as well. And when the ARPUs are like very small, it's the volume gains, right? So what is the best network dimensioning to serve these millions of customers serving their mobile needs. So we do a lot of those decisioning on using data. Then the other very big part is the personalization. So how do we reach out to our customers in a meaningful way where they feel connected, where they feel that we know we understand their needs and what we are presenting to them is something which best suits their needs, not only just a product perspective but in the journeys, in the channels we are reaching out to them, we have more than 50% of our revenues coming from the digital channels as well. So digital plays a huge role, and that also opens up a huge opportunity for us to know our customers better. But all of that is only possible when you have the data in the right format, in the right context in a trusted way, which then can be activated and used for these variety of use cases, Ash.

That's a great overview. And I heard you say customer a lot but it's both in terms of using data to provide better customer experience. But also you sort of take the circular approach there where it informs business decisions and investment decisions that ultimately results in better customer experience as well.

Exactly.

So large company, we hear time and time again that many large organizations or many organizations, large and small, really struggle to scale data initiatives. They look good on a whiteboard but getting them into practice can be pretty hard. Maybe you could share a little bit of background in terms of why you decided to modernize your data platform and some of the things that you were really looking to solve as part of that project.

Yes, yes. No, 100%. And I would say that we were one of those organizations as well where we were literally struggling to get the analytics out of the of large data that we had before the migration on-prem. So much so than whenever -- most of the times, whenever we are doing a complex analysis, I had to wait for weeks for my analysts because usually the response I get that, we have to schedule it over the weekend because during the weeks, oh, we don't have enough window to run this analysis. And some of the examples that I said, you can't wait for weeks because you have, let's say, the Board meetings upcoming, you have to get the approvals from your shareholders of the business and you have to prepare a business case. And just this -- this is just one example. So it was really big pain for all of us to be able to get those numbers out. And as I mentioned, I mean, XL has always been a very data-driven company. We try to -- and we aim to gain -- get all our decisions based on the data and facts. So the scalability was significantly hurting our speed of execution and speed of decisioning. Then the other thing was the reliability of the numbers of the data itself because just like many other companies, our legacy platforms had built over the years, and there were silos. There were different numbers of different versions of the numbers. Security was also another thing of concern. So ease of use, scalability, reliability, all of those channels -- all of those were the reasons that actually pushed us to actually making this transformation as a must-to-have, must-win battle rather than have something, which is like we are doing it just for the sake because everyone is talking about AI and transformation. Let's do that. No. We had very clear reasons why we want to do it, and we pushed all the way on the execution.

Okay. So that's a really good overview in terms of sort of the what or the why. If we dig into the how a little bit, like how are we partnering together to really drive this transformation?

Yes. I think that's a great question because that how becomes extremely important when you have to have an excellent execution of an idea. So all the ingredients were there. We were very much convinced that we need a modernized analytics platform. So we ran a whole process to select the best technology that we -- that can potentially serve our needs but then how do we execute? And that's where I think the word there is a trusted partner. We are not talking about the cheap best. We are not talking about somebody who just says yes to all of our needs and desires and then fail later in the project but a trusted partner where we can actually have a dialogue, who can understand what is our business challenges, what are we trying to achieve and also help us in shaping the program in the best way possible, right? So that's what we started to look out for that who can help us, and why I'm saying trusted because one of the challenges we are facing is because of the legacy, almost nothing was documented, and I will be open on that. So we wanted somebody who actually gets their hands dirty with us, go into the scripts, look at the logic of how the data is being processed over the last 10, 15 -- I mean, XL was 29 years old, 28 years old organization. So how those things have been developed and work backwards. So it's a lot of reverse engineering that was involved. So we wanted somebody who can actually work very closely as one team with us. And that's what exactly happened that helped us in delivering this project on time.

Okay. So I know that I love the way that you said that. It's not necessarily about the cheapest but it's really about driving the outcome. And I know that you guys leverage partners heavily as part of your project. But one that I want to call out, in particular, is the partnership between Snowflake and AWS. How did that collaboration kind of fuel the success of this project? And you've spoken about this publicly on quite a few occasions as well.

Yes, yes. No, exactly. Because, I mean, when we were starting this journey, and I think moving from on-prem to cloud was already giving us a bit of anxiety in terms of build shocks and how all of this work and then working with 2 separate partners at that time and even before starting the project, we were actually mindful of how all of this tri-party kind of a thing will work because AWS is there hosting the platform and then Snowflake is the AI data platform that will be there. We will need expertise from AWS in specific areas but we'll also need expertise from Snowflake's team in the product specifically. So how that will work. But just from the beginning, when we were actually evaluating all the way through the tender process and when we were scoping the project, what worked very well is AWS and Snowflake actually, those teams come together as one team. And when our teams joined, when I was going on the floors, when the actual action was happening, you couldn't actually tell that who is coming from Snowflake, who is from AWS, who is our own team and then there were other partners involved. All of that blended very well as one team. And I think that was the key success of the project that it wasn't, no, no, no, AWS -- there's no problem with the AWS, everything is working. and it's the Snowflake, which is not working. Many times, they both supported the cost and make sure -- I mean, in the initial beginning, we had issues. There were performance issues. There were quality issues. But when everyone worked together on this, they were able to solve those problems. That's the heart of the success -- the key reason why we were able to be successful.

Sami, that's -- I couldn't have asked for a better way to finish this conversation. When you said that you couldn't tell who was from AWS, who was from Snowflake and who was from XL. For me, that's the testament of an amazing partnership. So we spoke offline just around some things that are kind of really important and some key messages for partners. So we had lead with data value, show customers how to put data at the center of decision-making, prove quick wins. So faster analytics and double-digit cost savings resonate across every industry, position security and governance. So if we look at the T in ECT or ETC, depending on which way you want to describe it, trust is a massive thing for customers as well. And then this concept of ecosystem enabler. And I think this is where for our data cloud services, SI partners on this call, helping to bring the whole ecosystem together plays a really critical role. So Sami, thank you very much for taking the time. I truly appreciate you coming and sharing your journey with our partner ecosystem. Thanks for being an amazing customer of Snowflake, and congratulations again on being recognized as a -- with a Data Driver of the Year Award. I really appreciate it.

Thank you, Ash. Thanks for having me here and looking forward to our continued partnership and driving even more value from this.

And I'm looking forward to seeing the new video release. So when I walk into the office, I'm seeing XLSmart.

Let's do that.

Good stuff. Thanks, Okay, Sami.

Okay. Thank you, Ash.

Thank you. Okay, Hwee Bee, am I doing a quick close?

Yes, you have a quick wrap up, and we have a few exciting updates on our partner programs as well, right? A quick...

Yes, absolutely. So firstly, a big thank you to everyone for joining us today to our presenters. And yes, really, really, really appreciate you guys taking the time to jump on. A couple of quick program highlights that I just wanted to cover if we can jump to the next slide, Hwee Bee. We have a lot of work going on in the background at the moment to ensure that we keep pace and we continue to provide the right benefits, the right incentives, the right programs, investments into the right areas. So Sridhar touched briefly. We have a new Head of Worldwide Partner and Alliances, Chris Niederman. Chris is going to be joining us on the next SPN Pulse. So he's committed to come on, do an introduction and really frame up some of the thought process around how we think about partnering. But there are a couple of changes that we are rolling out now. So if we can jump to the next slide, Hwee Bee?

Yes.

We can just build all these out. Okay. There we go. So firstly, you're going to see some new metrics come. So in order to qualify at different levels within our program, we wanted to provide some more flexibility. We wanted to be able to recognize partners that are more focused on customer acquisition versus expanding within our existing customer base. So stay tuned for some announcements that are going to be coming there. Very, very relevant to the APJ market. We've got some new country-specific tiering goals. So not all markets are created equally. Many markets are at a different level of maturity and a different level of market size. So we've been able to tier some of those program requirements. We're also enhancing our services registration process. So what we want is to make it easy for you to tell us Snowflake projects that you're working on. It was a little bit of a convoluted process before. So we've really streamlined that as much as possible. And off the back of that, we're rolling out a new incentive to reward you for registering those services projects with us. So stay tuned. There are going to be some announcements coming around each of these, and we're updating SPN portal as we go as well. Next slide, Hwee Bee Tan. So look out for these enhancements, these e-mails. There are some enrollment steps required, particularly for our new services registration incentive that's going to be coming. Keep getting those in. And as I mentioned, jump on to the portal and you can see all of the program updates there. So with that, thank you very much for joining us. I hope you enjoy this Pulse. We had a lot of fun delivering it.

Yes. And Ash, I think thank you so much for hosting the whole session. And now I'm opening up to the partners in terms of some feedback from you, right? So let us know what's your feedback, any interesting topics you would like us to cover. We are more than happy to accommodate. We have come to the end of today but I'm going to stay on screen for a couple of minutes to wait for your feedback. So let's take 3 minutes to give your feedback so that we want to hear from you. Yes.

Excellent. Good wrap up, Hwee Bee. Really important that we get this feedback folks. We want to make sure that this is a valuable use of your time. Tell us what you want to hear more about moving forward. Thanks, folks.

Thank you, Ash. Team, I'm going to stay on screen 1:10. So in about 3 minutes, and then we'll close this webinar. So 60 more seconds. So any feedback is really, really welcome. Thank you so much. And we will have our next Q4 SPN Pulse coming soon in November, yes. We'll keep all of you informed. So more 30 seconds, I'm going to close out. Thank you all for joining us today. Really, really appreciate it.