All right. Perfect. Why don't we go ahead and get started? Thank you, everyone, for joining us. Amit, thank you so much for being here. Yeah, of course. Sorry. Here we go. For those that don't know, Amit is the President, COO, and Chief Product Officer at ServiceNow. Before we get started, a couple disclosures. My name is Arjun Bhatia. I am the research analyst here at William Blair who covers ServiceNow. I am required to inform you that I personally own shares of ServiceNow, and a complete list of disclosures and conflicts can be found at williamblair.com. Okay, let's go ahead and get started. I am very much looking forward to this discussion. Cool. Obviously, the big debate right now that everyone's going to be focused on is the AI debate in software. There's two sides to this camp, right? One is AI is going to disrupt incumbent software vendors, and the other is it's a huge opportunity. You've, at ServiceNow, launched a lot of new AI capabilities, a lot of innovation on the AI front. For a general investor, which I think there's a lot of them in this room, who doesn't maybe live in this enterprise software world, just explain why workflow orchestration, the position that ServiceNow has, is a tailwind to or it benefits from AI rather than is being disrupted by it. Maybe you can touch on your moats in that answer. No, thank, again. I think it's a definitely important question. I think for people who don't follow enterprise software or understand how enterprises use software, it's a pretty complex environment as whoever has been following it. There are a lot of different disparate systems which needed to be connected, which needs to be implemented and orchestrated on a regular basis. There's a lot of versions management, so backward compatibility, forward-looking compatibility, and very integrated in terms of how those things operate. ServiceNow has been in this business for 20-plus years, really managing that various different environments customers have in the enterprises to make sure the business run efficiently, predictably, as well as give customers the outcome they expect. That software which we've been building for many years, the platform we provide, is very integral to every every enterprise out there, all the Fortune 500, if not all the Fortune 2000. Good thing we have done over many years with ServiceNow is that we keep on innovating on the platform. It remains ahead. It keeps up adopting latest new technologies to make sure customers don't have to worry about doing it themselves. Because what it allows them to do is as they upgrade a software, they're getting the latest IP, latest capabilities, latest innovation in the same platform without having to learn something new themselves and not have to break anything which exists today. It keeps the business running intact. With AI, we're doing the same thing. AI is definitely a great technology. It's very, very helpful for automation as well as be able to do things much more efficiently. We brought AI into our platform today for a couple of years already, and we've been delivering that as integral part of our platform. Customers, when they upgrade or when they take our latest software as part of their normal day-to-day jobs, they're getting the value of AI with the idea that it works in the existing environment, so you're not breaking anything, while future-proofing you because you're getting updated versions of capabilities to make sure you're getting automation, you're getting efficiency gains, you're getting good revenue growth, as well as bottom-line improvements as well. It is something which everybody's used to. Ripping and replacing those things, which is what I think the narrative out there is that you can go and build anything because software building has become easier. Software could be built before as well. People didn't try to build custom software after packaged software became much more better and made customers' life better long term. Similar things are happening with AI now, right? You can always build something, but building part is very, very small, right? It's 15%-20% of your cost. It's the maintenance, the governance, the security, the compliance, which is very, very important for all enterprises. You can't operate something in an environment which you can't predict, you can't be secured, you can't be compliant to all the regulations. What we bring in a software is the value of AI, but also the value of all those different things which people think that it is not needed when you are thinking of consumer software. In the consumer world, you probably never have to worry about it. Enterprise, you do need to care about it a lot. That's the work we do heavy lifting. Building, as I said, is 15%, 20% of your cost, but the maintenance and the upgrade, as well as the compatibility, is huge amount of very hardening work required. Not sexy, not exciting, but you have to do it. That's where the moat for us comes in because we know how enterprises run. We know how they operate, all the systems which needs to be connected, and how do you get the effectiveness out of it, but at a very good value as well. Value creation is happening in our software as well today. We've brought AI to our platform. We've been innovating. If you look at the technology stack we have today and the platform we have is as modern as it gets. Better than pretty much any other vendor out there, while we're preserving the value customers expect from the products without having to rip and replace. I was using an example earlier, like just because you can grow vegetables in your backyard doesn't mean you become a farmer and stop doing your day-to-day job. Same thing's happening here. People can say you can build software, but why would you do it if the software which you are using today can do all these things at the same cost, if not better, and give you the value? That's happening. If you look at our AI businesses are growing very fast, and it continues to accelerate because we're innovating while preserving the investment customers have made. There are a lot of our discussions we can have around it, Arjun, but I think the reality is that these enterprises do require something which gives them peace of mind, gives them control, gives them visibility, as well as the innovation associated with that. We bring all of that together for ServiceNow today in our platform, and that's why we continue to please our customers and keep on growing. In that, you're the domain expert in all the different fields that you serve, in all the departments that you serve inside the enterprise, and that's where you're able to stay a step ahead, essentially, of what a customer might build themselves because they're not an ESM or an ITSM or a customer service. Yeah, good point you make. No doubt, I think the other part associated with that, and I'll address it some. The context, one of the things in enterprises is not like everything is documented. As many of you in your businesses today, a lot of the content out there, the document about standard operating procedure is pretty partial. It's a very small amount. A lot of things happen outside the documents, right? Exceptions, who approved what, why they approved it. Our business processes we are running inside ServiceNow is collecting, we run today 100 billion workflows on ServiceNow platform for our customers, and seven trillion transactions every year. It's growing at 20%+. We're collecting a lot of context about how a business decision was made and why this business decision was made. That data, all this related content is being bring into something called Context Engine, which goes on top of all these AI systems to really enrich it, but also make a decision which is little more guaranteed than any system can do otherwise, right? Our outcome, usually 90%-100% accurate, versus all other systems are 50%-60%, because they're just depending on some documents they read and try to run a workflow. We are doing it with the context we brought in, the data we brought in, and the domain expertise we have. Things like employee onboarding is a very good example, a very common thing, right? If an employee joins a company, you have to go and update maybe 20 different systems in one company. Some other company will have 17 different systems. It needs to be also done based on what department you're joining, what systems you require, what exceptions you require, what you need temporarily, what you need full-time. All that stuff has to come together when you want to get an employee onboarded. We can get an employee onboarded on our system in less than a day and make them productive next day. If you do this from build from a mindset or something which you don't have the domain or context, it might take you two weeks. That means your employee's unproductive, and half of things will not be right, so that means you go redo it again. By the time employee gets going, it's a month or two months wasted for them. That's kind of the example of things we see, not just around employee onboarding. Resetting a VPN access or giving you access to something temporarily if you're going to China, for example, with the right laptop. HR-related PTO requests. How do you resolve all this stuff? The context we bring in makes a big difference in the domain we bring in, and those systems are now AI-enabled, completely agentic, and lets customers really get the efficiency of AI, but with the guardrails and the harness around it, which makes it much more realistic and valuable to our customers. This is sort of the system of record advantage that you have. You've been serving your customers for decades, and you've sort of built this context over that time. One of the questions that I always get from investors on this is how much of the data inside ServiceNow is ServiceNow's versus the customer's? When you're bringing this context in. Yeah is it your own sort of proprietary elements that you're bringing in in addition to the customer's records, or how does that work? I think the customer data is usually not huge, right, this information which is in any kind of system out there which you can easily get access to. It's a lot of the runtime is where the data gets generated. Right? The metadata we create related to a particular process is very unique every time you run a transaction. Every time you run a workflow. That's the Context Engine. It, of course, applies the customer data and the relevance to that particular information, overlays with the metadata, and which is very distributed, by the way. It's not like you got one table. Thousands of parameters constantly being updated and constantly being collected and related to all the different systems you might have. We have a system technology called Workflow Data Fabric, which is also connected to all the different data warehouses. Today, no company has everything in one place. Everything is very distributed. Workflow Data Fabric connects and does federated information collection, overlays that with the metadata we have, which is our IP, and makes decision real time to get the outcome, right? Yeah. The data is ours. The metadata which we create, the context is really dependent on that one, and that's not available to anyone. That's why, as I said, a lot of time people miss this idea that you can do the work, but how effective the work has been is a really differentiator, right? Yeah. As I said, if I finish a task and you never have to reopen the task, I'm 100% effective. If something you'd open every other time, that's really waste of time, and it's not efficient, and it's costing you a lot more money, not just the cost of software, but the business time. Yeah. Okay, you have the context, you have the domain expertise, and you're building the agents. I want to talk about one announcement that you made at Knowledge, your customer user conference, which was basically opening up the platform. Yeah to third-party agents, right? We see a lot of agents out in the marketplace, and you've essentially made the decision that this context that we have in our system of record, we will allow customers to power third-party agents with it. Something that they want to build with Anthropic or OpenAI or anybody else out there. Maybe just talk about the rationale behind that decision, because you're obviously building your own agents as well in a vertically integrated stack that you're trying to provide to customers as well. We've been always an open ecosystem provider. One of the reasons we've been successful for our customers is, one, that we understand that customers have a very disparate and heterogeneous systems, and we cannot say everything needs to be like us, and only thing which everything has to be through ServiceNow. You have to work in thousands of other environments. We've been always thoughtful about that, and openness has always been core principle of the way we build software. Specifically in terms of the idea of that how do you get access to a system? In the traditional days, everybody use UX, right? UI, you log in, and you try to do things. Nowadays, there are going to be also agents calling into our system. It's not only humans interacting through a user interface, but also agents now asking you to do something. Maybe asking for data, but in our case, really asking us to do something, take action. We are really a system of action. The way we think about this is that you can also ask when a request comes in from an employee. It can come from Claude Cowork, it can come from Copilot, it could come from our own user experience or an AI agent asking for it. We need to really provide the value to our customers that you can now take that action and guarantee the outcome. That's the experience layer on top of us through agents or UX we provide. That's the headless or we're going to call it Action Fabric. Yep. The idea is not data access, it's really action. If you want to now onboard an employee, an agent can tell us, "Please onboard this employee for this particular company," and we would take the full work and get that outcome back to the agent. Yeah. We're not giving them the context data. We're giving them the full work. Yeah. I'm doing that work. That's why my value to every enterprise grows considerably. On top of that, it opens up more aperture for us, not just to our UX. Now, any other system can also get access to us. Now suddenly, as you said, there's an AI tailwind. For us, it's definitely because now giving me ability to take the IP I've built for years, understanding of the context and the data and the integration I've built, now open it up to so many more use cases. Yeah. I don't give them the context by itself. I'm not giving them access to like, "Hey, you can ask me." The Context Engine which I built, the data is so difficult for anybody to understand because really our unique IP and our secret sauce, they can't use it themselves. It is what we overlay on top of an agent. Yeah. We do the work for you, and that's what they pay us for. That's how we monetize it. We want to open up the opportunity for us broadly with all the work we've done through UX, our UX, third-party UX, agents, whatever it is, we don't really care. End of the day, our job is to really finish the work for our customers. Yeah. It seems like it's a TAM expansion sort of motion for you that there's generally more agents getting created and used in the enterprise because in either way, you're sort of benefiting. From a financial perspective, even if you are powering these other agents that are outside of your platform, that is a monetization. 100%. I think you're right, we are opening it up, but also the reason Anthropic is working with us is because the Claude Cowork, for example, when they want to have somebody to do something for them, they need someone to do the actioning part of it. Claude Cowork integrating with Action Fabric gives that full end-to-end. Versus they would go and do something in a particular system, the security issues, compliance issues, tracking issues, as well as the employees should not be going and updating things without permissions. Yeah. We put a layer, something we have launched, and I think you saw it at Knowledge as well, AI Control Tower, we launched it last year. Giving customers full visibility and control over every AI system they have. Not just ours, but third party. We understand what Claude is doing, what OpenAI is doing, what Gemini could be doing, what SAP Joule is doing, Salesforce, and we discover all those AI systems in the company and put it into this central control plane. We were doing that for assets inside the company before anyway, for enterprises. Any hardware and software. Now you have full governance layer, cost structure management, in terms of how much you're spending, which department is spending what models you might be using, but also all the security issues you might be running into. Then we bought this company called Veza, which does this access graph. If non-human identity is becoming a big issue, what Veza does is really manages non-human identities and ensures they're doing nothing wrong in real time. That goes into AI Control Tower. We have full visibility across everything now. Customers can see that real time. Then we open up a platform for all these different use cases. We have full ability to now manage the security, the compliance, but also finish the work for them. That is where the monetization becomes much more bigger. We believe integration with third-party systems makes sense. Yeah because we have the visibility and control, but also the actioning part of it. Right. Can we talk about just the pricing model real quick? Because I think this is another sort of narrative that's out in the market of, historically, you've had multiple pricing models, but a lot of it's been seat based, and now there's concerns about whether seats go away or the seat growth algorithm changes. You have all these AI capabilities, including powering third-party agents. What is the pricing model for that, and how do you evolve the business? I don't know, do you see that as a challenge or is it- Yeah, no, I think we've been evolving our pricing model. One thing we have to be always aware of is what are the customers, how they want to use our products. Yeah. What is the best way to kind of show them value and monetize. We have to be balancing on that one. We have changed our pricing over the last couple of years. We introduced something called Pro Plus and Now Assist at a higher end tier, providing AI capabilities in a hybrid pricing structure. It's a combination of seat, but with some idea of something we call Now Assist entitlements. You burn down that assist. It's an entitlement in terms of number of volume of assists you get. Customers are predictably in terms of what the ceiling is, but also flexibility in terms of how they use it and when they use it. We have evolved that pricing structure for our premier higher-end SKU a couple of years ago, and it's been very, very effective business, as we've said, $1.5 billion and a half this year planned ACV and growing very fast. That hybrid pricing structure has really resonated with our customers, and it allows us to really add more and more capabilities. What we've done now going forward is now taken that idea and applied to all our SKUs. We have a whole full set of AI SKUs starting from the base SKU to the higher-end SKU, functionally graded. It's different level of AI functionality depending on the SKU, and allowing customers to now use AI, Now Assist fungibly across all the different tiers as well. That is the structure we're going towards. If you look at our business now, and we shared earlier that net new business, 50% of our revenue is non-seat based now. Just shows you that our change in terms of how we've been monetizing is more reflective of how the world needs to be. We're not completely dependent on seats. There will be seats always, but there also needs to be another consumptive element, but with predictability, not this idea that I have no idea how am I going to pay this month. Yeah. That doesn't work. I mean, this idea that you go away and spend as much as you want and it reward you, that's silly and doesn't make sense long term. We are being very careful and thoughtful about how enterprises work and how customers think about it. This idea of Now Assist burn down with some predictability is what we're doing now. Our pricing structure is very straightforward now. It's across all our SKUs, so our go to market becomes very simple. Customers get AI across all our products. There's no idea of non-AI and AI. Everything needs to have AI as a base building block, but then you surround it with a lot of deterministic and core capabilities around it and give the customer outcome with some prediction. Yeah. Feels like it makes it a lot easier for CFOs to implement AI in that way, as opposed to, I think there's been some reports of individual employees burning through tens, if not hundreds of millions. I think it's- AI credits. Yeah. AI credits. Yeah. It's amazing that people can get away with that. Right. It's illogical. Right. Maybe just thinking about the agentic capabilities that you're building on ServiceNow, on the platform itself, what do you think is the advantage, or how should investors perceive the advantage of you providing a full vertically integrated stack with agents, data, governance, compliance, all in one SKU? Do you think customers and enterprises are more likely to go that route, or are they more likely to put together external agents with your infrastructure? Yeah, I think there'll always be interoperability required. I don't think there's ever going to be any enterprise that uses one product ever. Yeah. It will be multiple products. Everything will have unique needs associated with that. I do believe even the orchestration layer, there will be multiple orchestrators. You will have to integrate between different systems. We do this through agent to agent, but also from the business process level. There will be some unique build you might do. If you are a manufacturer, you build your own supply chain, which is more your IP. Sure, you'll build it in-house. You're building it before, you might build it with AI. Makes sense. You will need to connect it to your core operational systems, which is what ServiceNow is very good at. If you're running your IT department, your HR systems, your finance systems, your customer service, which are more operational, there's some uniqueness, but it's usually little more homogeneous made between companies, which we can provide at a much scale. It will integrate with your unique IP build in-house or third party. That's the future going forward. It's agent to agent, no doubt, but in this idea that you will have different layers combined together. Yeah. When we build this full vertical stack, it's really to make our product much more I would say AI native. The whole stack is very modern, and it has to have all the elements you require in the AI world. You can't just say that I will build pieces of it and then depend on somebody else to complete the story and not provide a solution. Eventually, customers want solution. They don't want piecemeal. They don't want spare parts. Yeah. The spare part world in enterprise software has been done many years, many times, and always has failed. Nobody can keep up, and you take your best people who should be building a business building software which is not needed, where you can have somebody who is much more uniquely qualified to do that for you. That's, I think, going to be the future where people will still buy solution. That's why we introduced something we call AI Specialist. This idea of autonomous workers. Eventually, what people are trying to do is reduce the amount of human labor, get automation, reduce the time to fix or fulfill some issue. That's what we want to provide with autonomous AI agents, and take out the human labor cost, but also do something which used to take two days, do it in 20 minutes in a predictable fashion. Yeah. That's the solution they want. Why would you take AI agents and cobble them together yourself if I can give you a higher level solution on top of it. at a better price and reduces your labor cost, right? With a much more prediction, because everything doesn't have to be AI. Yeah. You can have things where you're updating a database record. You can do that with just normal calls through API. Why do I want to run a token on an LLM? Which pricing might be going up every year. Who knows? You need to be smart about how you build your software stack and be understanding of what part you want to do it through a traditional software mechanism. What do you need AI for? What do you use ML for? Which model version also use? Everything doesn't have to be Claude Opus. Yeah. Right? You have to look at tiering where you need small models, cheaper ones, while you also build IP on top of it, around it. That's how we're building our top software stack. It's not this idea that it's fully vertical, idea that everything is owned by us, but we connect it to everything else. Yeah. In that, how do you view the model layer? Because you mentioned it's not one model for every use case. You have partnerships, I think. Yeah with pretty much all the frontier labs. You're using multiple. Yeah models, I presume, and just talk about that a little bit. Yeah. We are model agnostic, to be clear. We provide customer choice. Just like across the whole stack. We always this idea of you can run on top of any system of record. You can run on top of any cloud, hyperscaler or our co-lo or private cloud, for sure. Any model. Any data layer, as well as any engagement layer now. Any tool you can build with cloud code on top of us. We have a build agent. We've been always this idea of that very open, but any kind of choices available to customers. On the model is the same thing. You can use any model underneath. Model for us, LLMs, the frontier labs are probably 10% or 8% of the full stack. Yeah. 80%, 90% of IP, 90+% IP is we build. That's where the differentiation comes in. A lot of these models, in some cases, are interchangeable. Whenever the pricing changes, we look at which is the best pricing. For some example, customers might choose something. They say, "We standardize on this. We want to use that." That's okay with us. There might be some sovereign requirement some models can't meet. Yeah. There are also use cases where we do a lot of optimization. Through AI Control Tower, we know the cost structure of everything, who's using what. Then we also look at where the cost structures are to understand which model to use for a particular use case. Yeah. We're switching those things underneath the covers. Just like I think nobody cares what chip you use underneath your cloud most of the time. Same thing will happen to the models, right? As long as the results are there. Yeah. Exactly. Yeah. As I said, the autonomous AI Specialist, if it reduces my ticket volume, fixes something in 20 minutes versus two days, why do I care what I use underneath? Yeah. Right. I think that over time. Right now, this excitement and the FOMO going on, that everybody's looking at it every day in terms of features and all this stuff, it's going to even out. Right. You're at the forefront of at least getting AI into the enterprise and top-down processes. I think there's maybe a dispersion of how ready enterprises are actually to adopt AI. Talk about what are you seeing in terms of where are we in the adoption process today? For the customers that have adopted it, how are they thinking about ROI? Because they are at least increasing their tech spending. Where does the return come from? Yeah on AI adoption? Yeah, I would just look at from last year, for example. Early last year, when we were talking to a lot, agentic was becoming what people wanted to do, and the technology was getting better. We had an agentic solution. The customer didn't know. Whenever we used to go talk to them about it, first, they don't know where to start. What use case makes sense? Second was, how do I do it? Third thing worrying for them was: is it secured? Is it compliant? Can I have visibility? Those were the barriers early last year. That was how we thought about AI Control Tower, because what we wanted to give them is, let's take out this governance, security, visibility issue off the table first. When we introduced AI Control Tower, a lot of customers and the CIO started feeling comfortable, like, "Hey, I can implement AI agentic without having agentic go haywire and break systems underneath." You've seen some examples out there, like Pocket OS and all, where the whole database and the production system got wiped out because you had no control. That was one part. We did that middle of last year. We launched AI Control Tower, very successful. Second thing we did see from adoption perspective, as I mentioned, was: where do I start? What we did was we did 100 different use cases, agentic frameworks with a point-and-click kind of mindset. You can get going very fast in a few weeks and give them very prescriptive ways of getting going. You want to start with incident management. Do you want to do resolution planning? Do you want to do triaging? Do you want to do SecOps? We need to find those use cases, say, "You're currently running this system. This is the agentic version of it, and we can get you live shortly with this control around it." That took away the next barrier. Where do I start? Yeah. The third one was: how do I go doing this? We also brought in some FDE mindset, right? This is not like armies of people every day available to customers. It's for a short amount of time, taking the first use case and getting it live. Specialist engineers. Specialist FDEs. Yeah. Yeah, Forward-Deployed Engineers, which are basically black belt, very heavy understanding, deep understanding of AI and our products, and they can get our customers get going in a few weeks. Yeah. That they can take the barrier out, show them the value, show them the ROI. We do calculations and all that kind of stuff. Once we saw these one or two use cases go live, it was just open gates after that. The customer like, "Oh, yeah. I'm going to do this for this one, this one." We started seeing the volume go up. Yeah. That's really the adoption pattern we're seeing. This year, you see it. The amount of customers now doing agentic with us is pretty high. The volume has gone up considerably. That's why we felt confidence to increase our plan for our Now Assist to $1.5 billion, 50% growth over what our plan was. Yeah. Because of the volume of agentic going up and with the AI Control Tower surrounding it and the security. We invested aggressively on the security platform to make sure customers feel comfortable adopting this thing. That's the bigger worry. That's, I think, the trajectory now we're seeing. I think you mentioned volume, which I'm trying to figure out, is the breadth of customers increasing? There were, last year, I want to say, a lot of customers that were in pilots that probably moved into production this year. Yeah. Are there more customers and enterprises coming into the pilot phase? Yeah looking to expand? Yeah. We track the whole pipeline. There's a lot of POCs, and then we do pre-prod and prod. I would say 70% of our customers now who are doing AI are in pre-prod or prod. They are going very fast on the production environment. The usage, that's why it keeps on going up, because now they're starting to see multiple use cases. Usually, the one use case unlocks. You suddenly get four or five. The multiple department wants to get involved. That's where I think the production systems have gone up a lot more than last year. I think they were a lot more experimental. Yeah. This year, I think most of the case discussions we're having with our customers are pre-prod and prod. Pre-production and production so that they can start seeing the value of the investment they're making in this area. Yeah. Maybe last one for us to close out on. You have a revenue target out for $30 billion by 2030. Just talk about now that you are seeing some of this traction on AI, what role does that play in that long-term target, and where are we in getting to the interim milestones to get to 2030? Yeah, I think we are very confident about our 2030 plans. I think this is the base case we laid out. We want to be also very prudent to make sure that we know what we can predict and how well we can do in that. The $30 billion case for 2030, which doubles our business by 2030, is the base case we laid out. We do believe we have all the core products very differentiated, well-liked by customers, and TAM, which keeps on increasing. Yeah. There's also, I think, our expectation, and if you ask Bill, 30 is definitely not even to think about. Yeah opportunity in front of us, and we'll get there. The thing which will drive this, as you said, what gives us the confidence? One, I think we have a lot of unique IP and very modern platform, and it's getting adopted very fast. I think if you look at now you combine the things we have, AI with workflow, with data, and our data business is on fire, and it continues to grow considerably. You combine that now with security, which we have invested aggressively, and we have CIOs and CISOs are two big buying centers. CIO, number 1, and CISOs, Chief Information Security Officers, are number 2 buyers for ServiceNow. A lot of people don't realize we are one of the largest cybersecurity provider in the market today. If you look at the growth engines now we have, security, which is going to be and continues to be one of the leading areas for us. You add data with the Workflow Data Fabric, the RaptorDB, which we launched just last year, $100 million ACV in less than a year. Very fast growth, and we are looking at a $1 billion-plus kind of business growing there for sure. CRM, which has been there for a few years, especially customer service, field service management, CPQ, doing very, very well. You layer our core, very solid IT business and HR business, and then you layer the EmployeeWorks, which we build with Moveworks. The idea of engagement layer for any employee to really Like CVS Health, for example, uses it for 160,000 employees every day to get any kind of help they need inside the company. We have a very solid portfolio, and I think we feel the investments we've made are lining up very well. The AI is definitely a tailwind to let us get into a lot more conversation when you surround that with AI Control Tower and the security products. Yeah. A lot of growth opportunities, and we feel very confident where we are going. It's been reflected in our numbers, and it continues to do better and better. All right. Perfect. We will wrap it up there. Amit, thank you very much. Yep. Thank you, everyone.