OneStream, Inc. (OS) Earnings Call Transcript & Summary

Great. Thanks, everyone, so much for being here to close out day 3 of Communacopia. My name is Adam Hotchkiss, and I cover the emerging software space here at Goldman Sachs. We're really privileged to have Tom Shea, CEO of OneStream, here with us today. Thanks so much for being here, Tom.

Great. Thanks. Happy to be here, Adam.

So your tenure at OneStream obviously lasts a long time. Talk about the broad strokes of how we got from the inception of the company to where we are today.

Sure. So just to give everybody a quick background, if you've heard of Hyperion in the past or you think of the CPM space, I started my career in corporate finance and worked my way as a user of that product. I thought I wanted to be the CFO of a big national conglomerate. But eventually, I have been writing software all along that journey as well and kind of got into the space, which really informed the founding of OneStream. And so what I mean by that is for those of you that understand what Hyperion is, it's really -- it was the product that was dominant in financial close, planning, consolidation for the biggest companies in the world. And some of the products that are still used today, it's owned by Oracle were written by myself, my co-founder, Bob. So that was fundamental to us why we could sit here today because this is a pretty sensitive area. It's hard to get a chance to earn the right to do statutory reporting for large complex businesses. So that was the founding kind of -- the founding thesis then was why did we do this? Why did we create OneStream? It was because those businesses were built on roll-up strategy, which created a lot of tech debt over the years. So good products, but they formed into these suites and customers kept complaining that, hey, we need to be able to rationalize. We want to get value out of these products, planning, consolidation, reporting these core capabilities. However, when you think about that, it was -- it came with all of this technical debt. When you own it, you had to crosswire all that together. So the one in OneStream, the name is that we were -- we wanted to rationalize that. So Bob and I left after those acquisitions and said we need to go back to the drawing board and rationalize this all together. So that kind of brings us to the founding of the company, which was we need to be able to solve for these 15 different products. Oracle would show up to a sales call of 15 products, SAP would and IBM. They bought all the companies in those spaces. But at the end of the day, there's a little bit of logic behind why you would have 15 products. Financial reporting and statutory is a little bit different than planning. It's harder math. It's got to be right. It's got to be audited. Planning is a little more agile. So there's a reason why you have all those different products. And so as you kind of go on and go through that process, we looked at and said, we need to make this uniquely unified. We need some intellectual property that's very unique. And that's just to solve that core, be able to do all that modeling in one system. But we know there's going to be the need for 17 products, 18. It needs to be extensible like your iPhone. You have to be able to download apps and get new feature functionality out of it. So just right there is the founding premise of the company.

That's great. And maybe talk a little bit about the extensibility of the platform and how you -- how that's actually evolved over time? Where did that start? And what are the most natural adjacencies OneStream has moved into? And maybe following up on that, how has that shown up in what you're doing for customers and maybe some of your numbers?

Definitely. So again, going back to this notion of a platform, the best analogy that I can give you to help you understand the OneStream platform is to think of your phone. Your phone has a camera, has the ability to do e-mail. It has the ability to do location services, right? So these foundational pieces of technology that the phone can deliver those capabilities without an app. You can send in -- they're apps in general, but you expect those -- you think of those as part of the phone, not necessarily an app, right? So that's our core platform. Extensibility in OneStream, you can develop new products inside of OneStream without third-party technology, meaning you can code, you can build applications and you can extend the functionality. So think of it. We have these core engines, consolidation, planning, workflow, data integration, presentation, just like your phone has all those cores. And then we write apps on top of those to deliver different form factors. So how is that showing up for customers? Out of the box within the core, we -- the first significant app that we created was called Account Reconciliation. So you've heard of companies like Trintech, BlackLine. That gives our companies leverage just like you use your contacts in your phone to send e-mails, to send text and do multiple things. Our customers are able to download an application from our store, and we've already loaded trial balances in to do your financial reporting. So we can use that also as the starting point for what has to be reconciled and not reload the data, just like you're reusing contacts. So think -- things like that, we get to get leverage on existing data and existing engines and build new solutions. So that one is in the financial close area that supplements what we do there -- what we do in the core part of the product. But we've since gone way beyond that. We have a partner that develops sales performance management, and that's in its early days of going, meaning there's a lot of data in the OneStream platform that helps do territory quota planning management that you would want to do for running that part of your business. All of our AI products are apps in our App Store. So as we introduce new AI products, we actually have a full platform that it uses to do the AI computations, but those show up. So we basically -- when we see an adjacency, we see a use case, we can develop and then continuously deliver value. So when I'm in front of a customer, the idea is, look, we've got you, we can help you do the core data, but we can also help you expand and solve new problems. One of the most -- probably obvious ones right now that you would see is in Europe, ESG is raging. It's not that big of -- it's a bigger issue here in California and sort of other parts of the United States, no one -- we're not all over ESG. But in Europe, it's very important. So are we going to force a customer to go buy a point solution for ESG and then try to integrate it into OneStream? No, we built an ESG application that can solve and help them report and do carbon calculations right tied in with their financial and their financial reporting. So again, just showing how once you own the OneStream platform, you can get leverage on it.

Okay. Very helpful. How then do you think about pace of adoption into this market? Because I think one of the things we've observed over the last number of decades is that the ERP is as siloed and as low Net Promoter Scores as they have, you've rarely seen a company or a multitude of companies get beyond that $500 million, $1 billion of revenue mark with meaningful revenue growth. And, I guess, my question to you is, why is OneStream different from some of those cases?

So to start off by answering that, we originally targeted the large -- the very large, very sophisticated businesses that are part of large transformation projects. So again, out of the box, when we created the software, we created the platform, we had in our mind, we have to be able to solve these problems for Fortune 1. That is a big -- I don't want to call it a lethargic project. They're large transformational projects that take commitment from the customer, and they have their own pace of going through. That's how we built our -- you can't start at the bottom and make -- and solve a really easy problem and then go knock on the door of Fortune 1 and say, trust me, I'll do -- I want to do your public reporting. So we had that history with Hyperion, we did that. Now over the years, we're at 1,700 customers of some of the most storied brands, Costco, Fortune 1, General Motors, giant manufacturer. We've proven that we can do -- take out 2 to 6 systems and become that rational core financial processing engine. So what is our opportunity to grow, as you said, beyond 500 to go -- to accelerate that pace? It's productization. So thinking about taking that extensibility discussion a bit further, we've developed a plug-and-play architecture that we're able to build packaged implementation. So as we've done these really big implementations, we now are showing up and instead of saying, how do you want to implement OneStream, we're showing up and saying, this is how you should implement the best practice CPM use case, and we call it CPM Express. That's just the first implementation of our packaging strategy, if you will. And so what it is it ties together all of those core capabilities of financial external reporting, management reporting, closing consolidation such as account reconciliations, planning, cash flow, all out of the box, all configure, you can be up and live in 12 weeks with CPM Express. So that is a huge acceleration now, but keep in mind if I go to -- let's pick a giant company. If I go to Disney, Disney is probably not going to want to use CPM Express. They're going to say, I want to do it my way. And we can do that, and we're great at it. But when we go to a company that's emerging, that's a few billion that's now become global [ multi-entity ] consolidation, they haven't owned these types of systems. They want to start with and see if they're going to get a great head start and a great system because we're showing them what we've done with all these others. So as we go on, that's step one, and we're starting to see -- we're in the phase of adoption of that right now, and we're seeing the uptick, and you're going to see different flavors of it. So when I say CPM Express, that's GAAP right now. We're doing IFRS Express for Europe. We have a couple of different flavors of Express. We are working at a [ higher ed ] Express because we sell technologies in university. So this packaging faster time to value is a motion that we are building into the company right now. So it's not really a segmentation thing as much as it's a vertical and time to value strategy.

Very helpful. We'll get into the CPM Express discussion in a bit, but I wanted to address upfront something we've been hearing about software and the potential competitive risks from AI natives and from the LLMs themselves. I think we had OpenAI CFO, Sarah Friar, on stage just yesterday, and she mentioned financial planning offhand a couple of times. And so maybe just talk about OneStream's moat against companies that are coming after what seems to be a part of the market where there are rules-based engines, workflows can be augmented by large language models. What is your value proposition and competitive advantage?

I'm really glad that you asked that. I'm going to try to answer that in a couple of parts. I'll get to our competitive advantage. But I first want to start with an example, and I think I'll use this as a prop for you. So if you -- does anybody in here have access to ChatGPT, do you use it on a regular basis on your phone? What would happen right now then? So you hear Sarah say that we can -- you can do financial plan, type in. Give me a 2026 financial plan for OneStream and see what it comes up with. It will come up. It will go look at the public data. So think of that brain. This is ChatGPT. It's got some access to the Internet. It goes and looks at our public filing, and it will actually go through and it will give you -- it will find that data. It's debatable whether it's right, how current it was, but it will come up -- it will give you something, and it will be pretty assertive, like we think this is the growth rate. This is -- these are some of the assumptions that were used, go all the way through and you could -- you'll be able to read through and get an output. That's useless to me to run OneStream, absolutely useless. Why? All that had access to as that LLM, even though it was as powerful as it is, doesn't have access to any of our new product strategy, any of our planning, any of our plans on expansion geographically. There's no way it could meaningfully contribute to financial planning in the sense of the way we think of financial planning as a moat when I talk about core and being a book of record for OneStream. So let's flip that a little bit, and I'll explore the moat a little bit more, but let's flip that a little bit and see what -- if you're talking about modeling, you're talking about helping somebody on your team that sits in front of a spreadsheet and needs to create a couple of charts and analyze something specifically, you could probably -- you can get some mileage out of that because you individually are going to be interrogating some data, you're going to be charting it and say, yes, that looks pretty good. But that's not something that I'm going to be distributing corporate-wide or out to an external reporting agency that I can't trust. So you really need to think about the possibilities that you see with these engines. It's so important. Our own agent, that's why we've been spending all of our time is because our agent gets connected to the OneStream backbone, and it understands how to go and find that data that's been highly curated, all the calculations, all the things that represent our moat. So now let's talk about our moat and how AI is really an advantage for OneStream, but only if we connect it successfully to the backbone of our platform that has this information density. And so you hear data density and there's so much excitement and interest for Databricks, Snowflakes and all those products because they're capturing all this operational data, which is of interest to AI where we're taking financial information, financial data, turning it into information, which is of extreme interest to AI. I know what the total consolidated number is of OneStream at whatever legal entity by whatever geography, what its calculation state is, the last time it was calculated, the t time a number changed. And my agent sends it, when you connect my agent through an LLM into our backbone, it can give a deterministic answer to Bill if he asked the question, and our team is using this and can then actually prove and use the information because they can get comfortable with what the validity of that information. So what we're going to see over time is that these deterministic AI problems are the real opportunity for enterprise software. And every enterprise software company that owns significant workflows, that's a platform has an opportunity to be that gateway to connect the agents to their specific data and workflows that are -- that represent the moat and the reason we exist, whether that's people data from a Workday or whether that's financial data from a OneStream, that is where the real value is. And I think you're going to see that we're the tools that are going to actually unlock the value of AI within the enterprise.

Okay. That's helpful. What stops someone in the data layer who maybe has access to the data themselves from replicating what you're doing with more sophisticated AI models?

That's a really good question. So now we'll kind of get to the engineering side of it because we use -- we -- it's interesting. You hear the term AI native. We run our AI team as a completely separate organization. I really have an AI native team inside of OneStream. So our entire AI platform is written by AI ground-up engineers. That's all they have ever done from the time they were born is AI. And they're actually all under 30. So it's not that far off. And that team, we use AI every day, and we're seeing speed ups, and we're using AI that team, every single sprint team that's running and building features for our AI platform and inside of our core platform engineering teams as well. They have an AI agent that they use as part of their team. But let's kind of talk generally about what that means in terms of value for OneStream. It means it's not creating a new feature. It's actually coming through and doing refactoring, speed ups, writing particular -- a certain type of routine and helping that team get productivity, measurable productivity, there's real value. But -- so what does that mean? Is that -- is there any chance that I can tell that agent to say, I need you to go and rewrite our in-memory multidimensional analytic engine that is scaled out across multiple servers. And I want to make sure -- we have 1 separate cash synchronization, I want to reduce that to half a second, go write that. Zero chance of it ever doing that. Now I could scope it in as an engineer and say, the cash synchronization routine at this point is not performing that well. Can you go and focus on this exact subroutine and help me optimize it? But that requires me as the software engineer that wrote that algorithm to control that AI to do it. And so we're so far away from somebody -- you still have to have the knowledge of the domain and the space to amplify, but it means one person that really does that can be a force multiplier for businesses. So there is -- that is going to happen. But to think that AI is just going to all of a sudden turn an IT person at a big company into, we don't need to buy software anymore, that's not going to happen.

Okay. LLMs. I know you were built on Microsoft. So you have -- and maybe explain to us how that informs your access to Microsoft's LLMs and the real-time nature of that. And what does the real-time nature of changing and LLMs getting better that you have access to help OneStream provide incremental value for customers?

So when we think about LLMs, and I think most of you are probably evaluating software companies, you kind of kind of have to bucket your AI participation into 2 pieces. Are you a research AI company or are you an applied AI company? We're an applied AI company, which means I'm not going to go out and try to develop a competing LLM frontier model to any of -- it's just a capital barrier, it's not going to happen. Now we use LLMs targeted on our own -- many LLMs on our teams, but we're not going to -- you're never going to see OneStream [ GPT ] out there. So what that means is we cheer every time a model gets better because it means all the intellectual property that we're building. We have a RAG system, though. So how -- when I talked about the backbone, substitute any of the large language models that -- in our agents, in our generative AI, in our SensibleAI Forecast where we're using LLMs, we have to connect that in a way so it actually understands OneStream's structured data, the financial data. And we've built our own -- everybody know what a RAG system is. It's basically the idea that we can give you a secure way within our platform to take unstructured data and break it up in a way that you can reliably use it with an LLM. So not only can I give you the sales number or an accounting value, but I can tell you why you should use that particular account because all of our accounting manuals exist in our RAG and are part of our deep analysis agent as well. So when our individuals interacting, they can not only get the structured data, they can also get the unstructured data. So we have to create a significant amount of IP to apply the LLM reliably into the model, into our ecosystem. So as the LLM gets better, whether it's Microsoft and we access through, we're happy. We just get a better, cheaper result that race to 0 there just benefits us because that reasoning engine is getting stronger.

And maybe talk a little bit about the difference between utilizing just core finance data to do forecasting versus operational data, how that opportunity has evolved for you with -- as technology has evolved.

So sure. So this is an emerging theme that I see and there's a popularity because you see Snowflake, Databricks, the popularity Mongo area where these SaaS-based data warehouses, and Microsoft has to be included in that with their data lake and with the -- there's this new kind of polymorphic data structure where we have documents and relational data altogether in these giant kind of data lake concepts. So substitute any of those names in. We, at OneStream aren't competing in that realm. But that information is of interest to it. For years, we've been pulling data from Snowflake, Databricks because customers that have been adopting those different warehousing strategies, they have their ERP plus they -- maybe they have a lot of different operational systems that they want to co-locate that data into a warehouse. That can become interesting to OneStream. So why is that? I just want to step you back through the processing just to anchor everybody. When you do core in OneStream, every business has to do what we call core. It's what you have to be able to report your financials. And think of that, that's what you're doing monthly, quarterly, yearly, multiyearly. Every business will go through and have to produce financial statements on a monthly basis, get ready for quarterly reporting in your earnings calls. They have to do an annual operating plan. Why do you do that? Because you need to have lead times on your resources as you get a point of view on where you're going to go in another year. Once you have that, that's hard, though. If you're General Motors and you own businesses all over the world or you're a giant conglomerate, that is a difficult process. We help you do that better. Once you do that, though, you want to get into this fast-changing operational data. That's why I'm taking you back now. So you do the core, you become efficient at it. All of our customers -- so why was OneStream reading this operational data? It's because once you're efficient at the core, you want to start trying to think of read and realize what's happening and enrich the daily and weekly data so that you can navigate your business towards those quarterly -- monthly, quarterly, yearly outcomes that you're trying to achieve. And so immediately, if you were to come to OneStream Splash conference, you'll see every customer working towards pulling in operational data. So where -- what we've done in our platform and our operational tier is, we've created a series of technology that let us go with a no-ETL analytic. This is really, really important. And I've been doing this for a long time, coding -- we have this analytic engine and think about the -- when I talk about core, it's heavily audited. The data has to exist for 10 years. An auditor has to come in and be able to prove where every number came from because it's going to the SEC potentially. Operationally, that data is changing. I need to just get a signal out of it. I need a trend. But the math that you want to do on that fast-changing data can be difficult to perform as well. And so if I try to -- our customers would say, well, I've got all this data in Snowflake, but I need to derive this computation like I do over in the finance engine. Let me try to pull all this operational data into OneStream. That's bad because you would be -- you're breaking -- you're putting it into a place where you're auditing it. It's got a lot of weight to it because you don't need to audit it. So we built this analytic engine, that -- the no ETL that we can provide the same mathematical capabilities that we offer to our core finance engine directly on that fast-changing operational data. And that's what we call our AFA services, and you heard -- I mentioned that on the last earnings call. So that's been released. And what that does is I was on the phone with a customer that was taking 12 hours to load a traditional OLAP cube. We could give them the same exact information in 30 minutes. Same analytics in 30 minutes that it took them 12 hours to get the old way. So over time, for us, our moat extends to that operational data because why -- what's different about OneStream to go and leverage that analytic data in Snowflake or in a warehouse is that, when we pull that data in, we understand the difference between an asset, a liability, income and expense, how to do that math, how to do a translation. If we need to do intercompany, if we need to aggregate numbers together in a way that makes financial sense, that's really difficult for generic analytic tools. So even if Snowflake or any of the analytic -- any of the warehouse tools, there's lots of analytics, Power BI. Just take any off-the-shelf analytic tool, they don't understand financial models. They can go look directly at those warehouses. As soon as you need to make financially intelligent information, it breaks down. And that's why so many customers have always pointed us at those operational warehouses to pull the data into OneStream to produce a financial view of the information. That is the essence of our moat. It's why OneStream has existed and been successful, and it's underappreciated. If you survey the market, you'll see how many planning products, they pop up all the time, a new planning product, and that's why it's a target. It's a little bit -- the math is a little bit easier. When you do book of record reporting, intercompany, partial ownership, think about joint ventures. We can add up -- imagine all that data in an operational warehouse and you want to try to make sense of it, but you only own 30% of the company where the transactions are. And A owns B, B owns C and C owns part of D. OneStream can figure that out and can add it up and give you the right net number of what those transactions should be. You try to explain that to the operational warehouse query guy, and it's going to be like -- I don't even know where to start to do that math.

Great. Right. Well, this has been a phenomenal AI discussion, and I appreciate it. I want to switch gears to the business. This has been a consistent 20% plus grower on the subscription side, 30% plus for a long time. Guidance for Q3 is a little bit light of that. And so I want just to talk about your confidence, sort of what's driving that? And then what is your confidence to reaccelerate the business in Q4 and beyond?

Yes. So let's kind of start off with the guidance in the way that you heard Bill and myself talk about that the leading indicators are all positive. Funnel is in the right direction. Business is performing across all segments, what we call large named accounts, the sophisticated big businesses that I described. We're seeing the growth in commercial on a global basis as well. You're seeing in the United States -- you do see a difference. You see the strength happening in Europe, where their monetary easing has put them in -- everybody is looking up in the United States. We're not seeing the crazy kind of deal pushes or like I'm not saying there's this economic instability situation. But I do think if we get into an easing cycle, it will be a little bit more animal spirits and people will be approaching these big projects with more bigger is what I'm -- we'll see more of these projects come up. So I would say you're going to see some of that. So that's just -- I'm setting the backdrop. Why we're confident in our durable growth strategy is our AI strategy. We've got customers paying us real money for SensibleAI Forecast and for AI Studio that we just announced at our last [indiscernible] our agents are in previous. So we're in the AI game in the right place, and we've proven that the interest in the way that we're delivering AI to the finance market is what they want from us. It's got the transparency. It's got the auditability. It's got the usefulness for the office of the CFO. So the growth that we're seeing on that SensibleAI Forecast and the funnel that we see there is in the direction. And I'd like to say this, I've -- I wrote the first line of code of OneStream. I wrote the first line of code of my last company. I've seen what it takes for products to be adopted, meaning you go through this rapid product market fit, you take feedback from a customer, you iterate on it, you get into market and then you get to a certain point where I consider it a distribution problem. And so when I look at our AI, it's becoming a distribution problem. It does have one additional -- I don't want to call it handicap, but one additional -- a little bit of a headwind, which is the education. Everything that you and I just talked about today, there's a generalized lack of understanding of what AI is still in the market. It's getting more clear, but we have an education part of that sale, even though the product market fit is there, you don't know what you're walking into in terms of the enablement of AI awareness at the company that you're selling to. So as that becomes broadly accepted, there's going to be an acceleration there, and we feel really well positioned that our products are -- have the right product market fit. So all in all, the economic backdrop is good. Our product and competitiveness of our product across the geographies is solid. I'm hopeful that as we enter into easing, I'll see more uplift here. The guidance for Q3, though, as I mentioned, we have the headwinds in the government sort of with DOGE. As Bill said on the call, we haven't lost any government contracts. But we -- I want to explain a little just to make sure everybody understands how those contracts come up, they were not -- SaaS wasn't welcomed pre-DOGE for a lot of businesses that we sold to in the government. That is now an item. So what you see in our government forecast is we are predicting that those contracts that were term -- 1-year terms are going to transfer to SaaS potentially some of them. And we wanted to reflect that in the business because that has -- there is a meaningful change from DOGE. So that's kind of what's being contemplated. And that's kind of a onetime event that we were guiding towards there. So as far as reinflecting the business back up, we just needed to acknowledge there's just uncertainty in the buying demand in the government in this short term. But at the same time, our outlook for the government, we're FedRAMP high. We're one of the only vendors in this space that are FedRAMP high, which means that we can do Department of Defense level work in this space. We have a really great list of customers, and we have a decent funnel there. You just don't know what the buying environment is going to be. So anyway, that's why the guidance reflects what it reflects. So all things taken together, the long-term strategy that we have around the business with AI, with government, with our core and advancing our core with CPM Express being able to bring the product to more companies more quickly, they're all in line, and I feel confident that we have a durable growth strategy that we can continue to execute on.

That's great. I wanted to talk a little bit more about Express specifically in the deliberate packaging of your products. What -- was there -- was that in response to hearing from customers that, hey, we just value time to value a little bit more than you can do for us in this initial way that you're implementing the product? Or is this more of a deliberate move for you to say, we had to build our [ CHOPS ] and referenceability upmarket, and now we're making a more deliberate move to companies that maybe just prioritize this a little more. Maybe just break that down for us a little bit.

You answered it perfectly. It's exactly the way that we thought about it. It wasn't a response to a market pressure as much as it was as we plan the business and we start thinking what will it mean for OneStream to be $1 billion or a multibillion-dollar company, what will it mean to go from 1,700 customers to 5,000 customers? What does that mean in terms of putting those customers through the machine, meaning getting somebody onboarded, getting them successful and getting them expanded. We really looked at it and said, well, if every single customer that you sell is, I always call them a PhD project, right? We're going to come in and reverse engineer, you're a giant company, reverse engineer 4 to 5 systems and implement it. That's an amazing moat for us. That's where we've been so successful. That's what we built the company on. But how can I scale faster to that 5,000 mark? What we've learned so much at that high end of the market or at that sophisticated end of the market that we said, we have the right now to tell people how to implement CPM. And that's the foundational element or the foundational reason for creating CPM Express. It just so happens that it highly aligns to our commercial segment of our business. So that's where we started. When we had a commercially aligned sales team, and they tend to be talking to companies -- it's not -- it's more like upper SMB, more upper market. It's companies that are getting into multi-entity consolidation, getting sophisticated. We looked at that and said, hey, this is a great fit for those customers and those deals, and they typically want to get up and running faster. Let's go there and kind of build that motion, but then use that. So Express, again, is a foundational notion for us. It supports this verticalization, which if you see most enterprise software companies as a platform, they move into a verticalization strategy. Once they get their core, they start looking at, well, how can I deliver more value to more and show up to an actual industry or to an actual segment with something that's productized for them. It's exactly what we're doing. So you're just going to see -- you're going to see and hear more and more about Express over the coming years. And right now, the sales team, the partner community, when we talk about where that is in its adoption phase, we're excited because it's inflecting now. It's still early days. But I'm excited about what I'm seeing and the feedback that we're getting from customers. We're going to have more customers on stage at Splash in London. Partners are adopting it, and we're starting to see it -- we're working it through different regions. It's the right strategy.

Okay. That's great. How should we think about the typical customer profile and what they're using you for? I think there's sometimes confusion because you do offer a consolidated platform, but you also have the financial consolidation and close and you have FP&A and you have things like ESG reporting. And I presume not every customer uses every piece of that today. And so maybe talk to us about what the average customer looks like? What does that breakdown look like to the extent you disclosed it? And what is the future you are hoping to espouse?

The majority of customers start with some shape of core. So when I say core, again, remember, that might be -- they might start with planning or they might start with consolidation, but they intend to do both of those at some point. And that -- because it's just -- that's what you have to do. And it's so illogical to do one of those in OneStream and not do both because they share the same structures, the same -- if you're going to do your plan, you need to variance your plan to your actual. So they start with some shape of core. However, AI, if we didn't have the AI story that we have as a business right now, that's becoming increasingly important about every net new sale. And we are seeing some customers not only start with core, but move in parallel to do AI at the same time because of the operational impact that it can have. So -- and we've actually had a couple of customers start AI first. But again, they still need to have the core -- like an Express-based version of the platform up and running because remember, AI outputs are not really meaningful to the business. That's the other key differentiator for OneStream. We own both sides of the equation. If we put -- if we develop a better forecast with our SensibleAI and the output of that, when it comes out, how does the business use it? It goes back into the core models that represent the logical way you're running the company. So immediately, when the data comes out of our AI engines, it cycles into the core and is actually evaluated by our AI as well and can be turned into a meaningful decision-making piece of information for the business.

That's great. I know we're just about at time, but last question for you. What is your #1 priority over the next 6 to 12 months as CEO? And then when you look out 3 to 5 years, I can guess your answer is probably AI, but you'll have to tell me if it's different, what are you most excited about for OneStream?

So most -- the thing that I'm most excited about is the productization, the Express motion just because having done this for so many years, it's transformational for the company for us to have to be able to take a stance and productize. There's efficiencies everywhere in the business by doing that from the way we support the product, from the customer, the quality that the customer gets. So I'm really, really excited about that. But that's -- I almost can't decouple that from AI, though, because they're so executing and delivering on our full AI suite. The interesting use cases we're delivering around AI Studio and our agents. I just think it's all coming together for us, and I'm so excited about it.

Tom Shea, thank you so much for being here.

Thank you. Appreciate it. Thanks.