SaaStr 2025 Partner/Workshop Stage

that this is the first session of the day. It's a lot of pressure to begin with. The fact that everyone here is so attentive and so kind with your time, that just adds a lot to it. But really grateful you're here. I am Gaurav. I am the Chief Operating Officer at ClickUp. I'm responsible for all things money, how money comes in and money stays with us and money grows. So I run all all the sales, marketing, growth, post-sales, pre-sales, anything that touches money in ClickUp's business, it goes through me. So welcome to my talk. Today I'll talk about how to build a billion-dollar business. In fact, it's a multi-billion-dollar business and well into 9-figure ARR. So the secret sauce is There's no secret sauce. The truth is, if I'm assuming a lot of you are in tech yourself, a lot of you are in SaaS, some are founders, some are executives. Every 6 months to 1 year, you have to be willing to reinvent yourself because what gets you to that first million in ARR, then that next 10 million in ARR, 50, 100, $250,000, $500,000, everything is going to be different. Nothing scales infinitely. That's just the reality of business. And you'll have to find a playbook that's relevant to you at each stage in your journey. So today I don't have hacks for you, but I'd like to share some thinking paradigms that has allowed me to scale ClickUp well into 9 figures, and we continue to grow. Really fast for private standards. So the first thing that a lot of companies don't get right— know where you win. People try to adapt hacks and strategies from everyone to their business, only to realize that fundamentally that does not work. So how do I think about it? Let's plot a matrix. You need to know which game to win, and this would allow you to understand where can you truly win. So take your business and plot it on a 2x2 matrix. On one side you have LTV. How much can you make from your customers? That's an easy way to think about it. On the other side you have the volume of customers. How big is your TAM? So you have two things that you're looking at: how much money can you make from your customers, and how many customers are there out there. If you happen to be on the side where you have very few customers and they don't pay you anything, you should not be in that business. You should just leave that business. And I know it sucks, and I'm sure a lot of people in this room, in this conference, are perhaps in that quadrant because they have a fancy idea but you should just run away from it. So the next one, this is interesting. You have very high LTV and you have a very small TAM. You could call this whale hunting. For example, you have a software that you could only sell to Fortune 500 or Fortune 100. So by default, you can't cast a wide net. You have to spearfish. And the channels that work here are very, very high-touch. Field marketing, trade shows, conferences, business development efforts. These are all the channels that work, and you really have to spend that capital upfront because the size of your customers are also going to be big. You have the other side now where you have a huge volume but very low LTV. What do you do now? You can't afford to spend in some of these very expensive channels. So you have to learn how to cast a wide net for as cheap as possible. Content marketing, SEO, search engine optimization, social strategies, building a community. You can't afford to pay a ton of money in CAC. So you have to grind and build these organic channels that are going to be LTV to CAC positive. And then finally, my favorite quadrant, the first one, which is where ClickUp operates, because the beauty of operating in this quadrant is you can pretty much do anything under the sun. You can go build an enterprise sales team, go after enterprise. You can also buy billboards. You can also run TV ads. This is where there's a lot of joy in building business, especially if you are a distribution person or a go-to-market person. Because you can make pretty much any channel work. So knowing where you win is extremely important, and that comes down to you honestly mapping out how much LTV can you make and how big is your market. The next thing, regardless of your space, your industry, B2B versus B2C, study the best in class. If you are trying to build an SEO motion, See what HubSpot did, but also see what NerdWallet did. Also see what Canva did. See what Zapier did. If you're trying to build a brand and use content as a way to build the brand, don't be stuck that you are a B2B business. Look at what Liquid Death has done. Look at what Beats by Dr. Dre has done. You have to be willing to learn from the best in that field regardless of your industry. At ClickUp, our teams obsess about each of these companies and we try to adapt the learning from these companies to our business. So don't be stuck with B2B, B2C, healthcare, whatever. Be open-minded and learn from the best. Which brings me to the next point. Don't be zero-sum. Don't be limiting in your beliefs. Oftentimes I hear this a lot that we are a B2B company. We don't do the stuff that B2C does. When I was in healthcare, I used to hear this that you have to target doctors differently, but doctors are also consumers. So don't be zero-sum in your belief. And I see a lot of that going on in SaaS right now. Product-led growth versus sales-led growth. When the truth is they should be playing together. I hear a lot about brand building versus demand gen. When the truth is they should be playing together. The beauty of being in a revenue role or a GTM role is that everything that you do, you should be able to tie it to impact. And you don't have to build zero-sum narratives when these things are actually compounding. So at ClickUp, we do both really well. We have a PLG motion that is user-focused. That's the stuff on the left. We go acquire users, we activate them, we monetize them. They continue to grow with ClickUp. In fact, some of you in the customers came up to me and you are ClickUp customers, and that's, that's great. Thank you for that. And very likely we found you through our PLG machine. Simultaneously, we have a B2B machine which goes after accounts where we have already landed, and the whole goal is to reach out to those customers. So product-led growth gives us the distribution at scale, and sales-led growth gives us the LTV. So much that we constantly see this behavior when a customer moves from self-serve or product-led growth to sales assist, we see an 11x lift in LTV. Now a lot of other companies here would say do one, not the other. I know so many companies that kill one motion in favor of the other, when the truth is they feed into each other. Product-led growth allows you to cast that wide net, catch a lot of fish, and then sales-led growth allows you to go hunt for whales within the fish that you have captured. The third piece, and I think this is the most important thing for a scaling company, you have to be able to own a growth portfolio. As you grow, your portfolio needs to grow. You need to add more and more diversity to it and more and more resilience to it. So I have a background in banking. I learned some of these tips, especially around financial instruments. But interestingly, they also apply really well to business strategies. So how do you build a good portfolio? It comes down to first looking at your channel mix or looking at your revenue opportunities as key bets. These bets can be channels, go-to-market motions, segments, products, business line, geographies. Each of these is a bet and you need to understand, is that a big bet or is that a small bet? And then you need to, and you need to balance your investment or all the time. And this is the most important skill that a scaling company needs to get right. How do you place more and more bets? How do you protect your downside? And how do you see them come to life? So rules for a growth portfolio. Understand that there is a risk and reward. In one way to phrase that is understand you have revenue and payback. You can take a lot of risk. And now your payback can jump from 20 months to 50 months, or you can decide to optimize for payback and now your top line will suffer. So you have to find the efficient line at which you can deliver top line while taking the right risk. Understand that each channel is not as repeatable as— like channels are not equal. Equally repeatable. What do I mean by that? So at ClickUp, we have a social machine that goes viral. And what I mean by that is we have clips that have 50 million, 60 million views. The thing is, I can't predict which clip is going to go viral. On the other hand, I know which SDR is going to get to his or her number. I know exactly on Google branded search what my ROI is going to be. I know exactly which AE is going to be, which account executive will hit his number or not. So you have to understand that in your distribution strategy, not everything is as repeatable, and that's okay. I use an example here for SEO. We have a massive blog machine. We write hundreds of articles a month, and these are not AI-generated articles. These these are human-written articles, and we get millions of views on a blog. The only thing is Google changes your algorithm, and now your views fluctuate. So that's why each channel is not as repeatable as you might think, and you have to truly understand that. Examples. So Search, as I said, the beauty is for you to place that bet, you need a very small team. So very small team, they can go write articles and the probability of it working is medium. On the other hand, I said sales development, you know how many sales dev to bring in, they will generate a certain amount of pipeline, the probability of that working is extremely high. So you truly have to understand which of your channels are working to what degree. As a fast-growing company, it's important to tell yourself every day that you haven't made it. That is because when you are at a $10 million ARR, getting to $15 million is relatively easy. But when you are at $500 million in ARR, growing 50% to $750 million ARR is extremely hard. Extremely, extremely hard. The only way you can beat that is by telling yourself every day that there's a lot of better opportunity out there that you don't understand. So what I mean by that is you have to assume that there are better distribution tactics, there are better motions, and you have to constantly apply that pressure on yourself and on your team so that you keep finding new opportunities to grow. Oftentimes, a lot of teams get stuck at local maximum because they can't envison what a better version of the distribution strategy could look like. So always be uncovering new ground. At ClickUp, this is how we allocate that resource. So 70% of our resources go towards tried and tested proven channels that we know have very high probability to succeed. After that, 20% of our resources go towards small bets. These small bets are going to give us small, small incremental gains. And then a very small amount of our resources go towards things like going viral, building content that people would share on social. We call them big bets because if you land one or two of those, you're going to land a really big win. But it's going to be extremely rare that you can land them and land them predictably. So a good growth machine, a good revenue machine always prioritizes stability, but then keeps some resource allocated towards finding new ground. That's what I meant. The line that you see in light blue, that is all about what is proven and what are small, small gains that we can add. That is what we sign up for as a financial plan. But then we place a few big bets and we work incredibly hard to land one of these. And if we could land a couple of these, we have an inflection in our growth. And that is what predictable growth with asymmetric upside looks like. I began with this— understand that your channels cannot grow forever. There's a law of diminishing return. After you've hired 100 account representatives, if you're going to hire more 100, you will see productivity decline. So understand that your channels don't have infinite capacity unless you keep working on them and keep building that growth and capacity. And then this is interesting. A lot of companies when they hit scale, they start abandoning it. I spoke about diminishing impact. There are some channels that actually have a compounding impact. For example, if you have a community, it takes time to get it going, but at scale, that ends up scaling itself without you having to manage it. If you have a domain that has very high domain authority with a lot of content on it, as you add more content, the domain gets better. If you have a product that has extremely high virality, think Slack, think Dropbox back in the day. As more people use it, more people share about it, and it actually compounds. So you need to also understand that while some paid channels may not scale forever, there are some organic channels that can scale forever. You just have to build them right. Case in point, Google SEO and Google paid search. You cannot scale Google paid search beyond a certain level. Google organic content can scale infinitely. And then finally, compounding your machine through AI. If you truly see revenue as a very complicated machine, then you will identify there are opportunities to add efficiency at many, many steps. Depending on what type of business you are, you are a PLG business or you are an SLG business, sales-led growth business, or you're a hybrid, these are all the boxes you can optimize. And the beauty with AI is you can optimize these boxes really, really fast compared to what you were able to do before. So at ClickUp, what do we do? We are attacking 80% of these boxes, and our teams are taking an AI-first approach to automating and increasing the throughput. So we have an AI SDR. If you fill out a form on our website, you'll most likely get an AI SDR reach out to you. We just started piloting that last month. She generated about $1 million in pipeline. We produce 40,000 pages every month using AI, and we localize these pages in 15 different languages. That has allowed us to 10x our search impression volume. We have an AI chat agent on our support ticket that allows us to resolve customer issues faster, frees up capacity for our support team to then go proactively work on on key accounts. This is just the beginning. As I said, across each of these blocks, we have a team that obsesses about how to use AI. So I guess it— there are no hacks to building a sustainable, ever-growing revenue machine, but I'll leave you with a few tips. There's not a single playbook, but you can learn from the best. So for that, you need to truly understand what is your LTV to TAM, which quadrant do you sit in. Then go learn from the top 1%. Don't play zero-sum games. You don't have to place things as either-or. Oftentimes you'll find that you can do many more things than you have anticipated. Own a portfolio. And keep increasing that portfolio over time and increase your throughput. The competition is going to be extremely fierce. If you're not using AI to 10x your velocity, then you'll be left behind. So there you go. I'm more than happy to hang out and answer very specific questions in your businesses after this, but really no specific hack, but think through first principles learn from the best, and build your machine ground up. Thank you.

And they'll introduce you if you want to go up towards the front stairs on the front of the stage. Okay, okay, cool. Let me just check one last thing. You ready? Yeah, I'm ready. Please welcome to the stage, CTO, Neo4j, Philip Ratley. This revelation by Klarna made quite the stir, that they unplugged an estimated 12,000— 1,200 SaaS apps through building their own AI. Now, if I look at this from the perspective of the SaaS buyer, this maybe presents a great opportunity for seeing all the dollars that I could potentially save in licensing fees. It also might make me reevaluate how I look at buy versus build. If I look at it from the perspective of the SaaS builder, man, like, this is potentially existential and pretty scary. What do I do about this? What lessons can I learn to survive and thrive? Something significant about this too is There's been all this talk about enterprises that have difficulty getting value from AI, and this is probably the biggest example to date of large-scale success in AI in the enterprise. So today I want to take two perspectives, and I'd like to share Klarna's story from the standpoint of what is it that it can teach you in this room, both as a buyer. So how, when, and to what extent to replicate this playbook. As well as as a builder, how to distinguish, distinguish yourself in a new roll-your-own AI era. So let's look at what Klarna said publicly about their AI journey first. So starting in February 2024, they announced that they had this breakthrough new AI application in customer service. So really focusing on bottom line savings. And saying that Klarna Assistant, powered by OpenAI, has in its first 4 weeks handled 2.3 million customer service chats, and that it performs the equivalent job of 700 full-time agents. Fast forward a few months, they then added to this and said, look, there's been a lot of reporting on the bottom line savings, but what's actually creating the most value for us is this internal employee employee AI assistant called Kiki, which is driving an even bigger change, helping to solve for organizational memory. Um, and it's like being able to ping anyone in the entire company to get an answer immediately. Fast forward a few months, Klarna CEO Sebastian was on, um, the Training Data podcast with Sequoia and asked, well, how did you do this? And this quote is pretty great. He said, well, in general, I'm so freaking transparent, and this is the one time in my life that I'm feeling cagey about sharing the secret sauce because I actually think this is a fairly important strategic advantage. And we'll get to the secret sauce in a moment. Let's drill diver— drill deeper into the journey first. So if I distill the journey, what did they do? They first built an AI app. For customer service targeting bottom line savings and saw huge benefits. And then they said, well, probably there's an opportunity for a lot more benefit. So we're going to double down on AI and on a culture of AI and AI fluency. They then found that in the course of building, they were limited by data silos. And what this concretely meant for them is they had a lot of data, in a lot of places, both internally and inside of SaaS apps where they were only able to use that data for whatever that SaaS app handled or whatever that application handled, not more broadly in order to serve an AI with enterprise purview. So they built a data foundation with tooling and process above it, which then set the stage for them to roll their own. So in a different time and context, things might have gone differently. And I think it's important to, to understand too, something else Sebastian said in this podcast is, if you see something that is just better than anything else you can build yourselves now, then I would buy it. So he didn't go in with an attitude of categorically, you know, anti-SaaS. Um, he actually looked at it from a perspective of What is the top-line value that's available? What am I creating for my organization? So he's purely motivated not by the short-term, but by what kind of long-term strategic advantage does he have to create for his company. So what does this mean for you today and you coming up with the right play for yourselves? So here, let's go back to like business basics and start by asking, where will AI provide the greatest dividends? And understanding that there's a trade-off in focusing on one thing versus the other, let's look specifically at bottom-line savings contrasted with top-line value. Where can you save on the bottom line? Well, there's direct savings, which is subscription spend. There's indirect, which is all the cost of management complexity, um, tracking licenses, overhead, um, and then You know, not, not to be forgotten is third-party security risk. And then also, you know, ways that bottom line savings often come into being are through productivity gains, finding faster and cheaper, better ways of doing things that you're already doing. Let's look at the other side. For value creation, you all know this, it comes down to really one of two things. I'm either going to use AI to make my existing products better so that I can get better retention, um, better, faster customer acquisition, be more competitive. Or I'm going to use it to create new products and services which expand my addressable market. Now with that in mind, let's look at the economics of roll-your-own SaaS. First, you have the contribution in terms of dollar savings in direct spending. And I got this estimate from Zylo SaaS Management Index that the average company spends just under $5,000 per employee on SaaS. Then you look at the indirect side. This is always harder to measure, but the one thing I did— I was able to have, um, to track down was depending on where I look, somewhere between 70-90% of organizations have had one or more breach due to, um, a third party, costing an average of $7.5 million. I also saw like a $4 to $5 million figure to remediate. And of course that's the direct cost of remediation. And as you and I know, there's oftentimes a much more significant cost in terms of brand and reputation, customer, uh, trust and so on. Now this is offset by the opportunity cost of managing, securing, building the thing in the first place. It's not like you're going to get these apps for free. You need to build them. And something I know, you know, from years in product is it's actually a lot easier to build a product as a standalone than it is to build something that then has all the continuous integration and security, security, security upgrade capability and so on, um, to be a durable product. And then there's an opportunity cost. That opportunity cost is what you could have built if you'd focused your efforts on the top line. So those are maybe arguments against, but then let's also look at some maybe more subtle but important strategic impacts. These are intangible benefits of rolling your own. One is AI culture and skills. So internal applications that are maybe less high stakes are a safer playground for building some AI muscle, where AI muscle means skills and people and organization. At the same time, it also means infrastructure that you can build on. And so I think this last point is, pretty important. And so I went back and tried to find some citations that came from maybe outside the technology space to take an outside-in view. And I found this paper in the Journal of Applied Behavioral Science from the December issue of last year. And there are a few quotes I want to, I want to quote from here. One is that realizing the full potential of AI-based innovation requires a fundamental rethinking of strategy strategy, organizational setup, skills, and collaboration frameworks. And then the second is that innovation must be managed across departments and functional boundaries, and increasingly across organizational borders, integrating a network of external players—users, customers, suppliers, partners, et cetera—through the appropriate tools and methods. So, what's this really saying? It's saying that The purview of AI can be greater than that of a single department or a single individual in, in the company. Our brains are only able to encompass so much. So subject to data security and privacy and regulatory firewalls, you should be able to use any data inside of the organization to come up with an AI-based decision and get a better result, potentially better than even humans, um, with a more narrow purview. Can provide. And so this is what Klarna was solving for. Back to the secret sauce. That their secret sauce was that they're transforming the way they collaborate with their GenAI chatbot powered by Neo4j Knowledge Graph and it's having a huge impact on productivity in ways that were not possible before without about graph. This was part of a much larger thread on X, and I'll just read the, the first summary: Karna's decision to shut down Salesforce and other SaaS providers stems from a strategic shift to consolidate fragmented corporate data using AI and graph databases to improve knowledge management and productivity. So let's, let's walk through this journey a little bit. So first step, connect their data. And what Sebastian said here is feeding the LLM a fractioned, fragmented, and dispersed world of corporate data will result in a very confused LLM. And many of you have probably experienced this yourselves. So they took inspiration from Wikipedia as being effectively the largest, at least curated knowledge graph, open, available on the internet, managed by 20,000 people, and is, you could say, a shining example of high-quality, accessible, and accurate data. What could they learn from this? And so they then took this back and said, well, AI can't operate on inconsistent, fragmented data spread out across the enterprise. Let's repatriate it. And then if you can connect your data, the AI can make better sense of it and do so holistically and give you a strong data foundation for AI. And what this looks like is something like this, where they're unifying all the various silos. And you can imagine internal or SaaS applications that each do a particular thing in the realm of customer support or HR or products and services, billing, finance, etc. But bringing that all together and connecting it. So that's effectively what they did. And part of the beauty of this approach is there's a lot of value in just one, and then more value in two, and more value in three. I've never been a fan of the, you know, boil the ocean, let's bring all the data in and then let's figure out what we're gonna do with it. Like here you can be very surgical targeted and experience data network effects as well as use case network effects, meaning that once you bring multiple kinds of data together, you can solve different kinds of problems, um, sort of as a bonus that maybe you hadn't initially planned to. Step 2 was, well, let's build some AI application layers and tooling on top of that. Um, so they used Cursor, and obviously there's a bunch of other stuff, uh, v0, Replit, Bolt, Lovable, and so on. So data foundation and then AI application building foundation on top of it. So let's, let's summarize this and roll this up. As a SaaS buyer, here are some principles whereby you can operate that at least I take away from Klarna's experience. First of all is a long-term view and an understanding that top-line value, which generally correlates with customer-facing applications, is far greater than bottom-line savings. So that, you know, $5,000 give or take per employee, like think of that contrasted with the cost of a fully loaded employee, like best case that's a 5%, 5% of your cost, and then you lever— you you compare that to the top-line value that you can get by significantly improving your product or creating new product lines. Number 2, well, in, in the short term, lower-stakes internal applications present a good opportunity for building this AI muscle. But I would also think in terms of undifferentiated heavy lifting, which is a term Amazon likes to use a lot. Where if you're investing effort into an application that saves you money, but that doesn't really advance your core business, then that's probably not the best use of time. It's probably best to pay for something and then use your unique skills and expertise to build something that only you uniquely can build so that you accrue top-line value. Now, if a SaaS app can be easily replicated, like, you know, if it's a wrapper on top of a database or on top of an LLM, then sure, go for it. Another key principle is embracing AI, of course, means survival. And so working out what building blocks you need to be an AI-first leading company and build a data and process foundation as well as a culture around that and prioritize according to those long-term Long-term, the right data foundation probably means breaking through silos. And then if you add this up, net-net, SaaS will continue to have its place because you're not going about this to replace SaaS. You're going about this to differentiate and build for the long-term. All right, that's the buyer's perspective. What does this mean to you as a SaaS builder? All right, let's first round out the threats, 'cause we're not done here. The other threat is that Salesforce or Workday turns your product into their feature. So you need to defend against that too. So let's look at the spectrum of SaaS features, and the ones on the two ends are pretty bad, so let's all try to avoid these. One is your product gets buried built by your customers using AI, that's death by roll your own. On the other side, your product becomes someone else's feature. That's bad too. But in the middle, your product can change the game by delivering outsized and unique value through AI, through your expertise. And this is where success will lie. So the next generation of new AI-driven SaaS And we've all seen these companies that go from like zero to significant amounts of revenue in record, record time. That's the middle. That's where you want to be. So how do we get there? How can we take these lessons and build this, you know, next generation SaaS future? So let's, let's, let's talk about what not to do maybe first. Is, um, this comes from, uh, the BGE Squared podcast, uh, where Satya Nadella was interviewed and said the notion for that SaaS business applications exist, well, that's probably where they're all collapsed because if you think about it, they're essentially CRUD, um, databases with a bunch of business logic. So basically like a CRUD layer on top of a database. So yeah, Don't do that. Like don't be a SaaS, don't be a layer, thin layer on top of either a database or on top of an LLM. So what can you do? I think there are a few key principles you can follow here. Number one is develop world-class domain understanding. If you understand the domain better than anyone, including your customer, 'cause they're their expert in their own domain, they might, may not be expert in finance or HR or whatever it is your SaaS application is targeting. Um, big opportunity for differentiation here. Number 2 is if you've got some complex process that people hate going through and it's a pain in the neck, create an experience that people love. Next is maybe you have access to some valuable and proprietary data that could create value-add when merged with the customer's own data. Obviously, we just said this, don't focus on something that you can build in an afternoon with AI. And then last but not least is network effects. There are all kinds of opportunities here, starting with people as well as data and others. Also understand the new customer expectations that are emerging. In this new AI era. A key lesson that I draw from Klarna is you need to be— customers, your customers are going to be building more and more of their AI app, more and more of their applications because that's now so easy to do. So you're going to need to be able to interoperate and live alongside these environments. And so if you do that explicitly, proactively rather than reactively, there's maybe an opportunity there. This always goes without saying, like there's all kinds of stuff you just need to invest in that's table stakes, but data security is an increasingly important one for reasons we mentioned earlier. Explainability, so this isn't true of everything, isn't so much true of a meeting summarization kind of AI or, a sales rep, a virtual sales rep reaching out to customers. But for high-value use cases, each question tends to have higher stakes. And if the stakes are higher, then your impact of a good answer is higher, and the impact of a bad answer is commensurately high as well. And this could translate into high dollar value, brand and reputational impact. We've all seen examples of like embarrassing AI responses. And the way you gain trust from your own customers, users, regulators is through explainability. Um, this I see as maybe the sister, um, requirement of the one on the side is data APIs, um, and interoperability at the— so the left side refers to maybe the agent and application layer and the right side at the data layer. And then key from Klarna's journey journey is their data is their AI fuel. You should make it easier for them to get access to it, and you don't want them to feel like their data is locked away. Oh, and then value always goes without saying. It can be hard to me— measure value though, so I'd encourage you to go out and find ways to measure the values that you can then demonstrate this. And then on the tech front, a lesson is remember that AI is more than just gen AI. Like we all get, you know, kind of sucked into the latest thing and the pendulum swings. And I think it actually makes sense to try to use new amazing things to solve 100% of problems. And then when they don't solve them well enough, then you dial back and see how do I combine that with other things. There's a lot in AI and there's a lot beyond AI that can be combined together. And we're seeing this in some more complex agentic workflows. And their lesson here is that the— Klarna's secret sauce, specifically on the graph side, can also be used to your own advantage. So this is a whole separate talk and I'm not gonna give it here, but one of the kinds of graph AI, and there are, there are many different uses, is something called GraphRAG, which is Retrieval Augmented Generation with a Knowledge Graph, which is what Klarna does in addition to and alongside vectors. And the advantages that they and others get out of this is number one, better answers. So that not only means more accurate, but it means richer answers that are more useful. Number two, explainability. You can get gray box or even white box explainability depending on the degree to which you're augmenting your prompt, uh, from the knowledge graph or actually delegating part of a decision to the knowledge graph for say like a supply chain risk kind of problem. Um, and then last but not least, you can actually use your knowledge graph to coerce security rules and access controls. Um, so it's a nice hook for that. And then for you, you're not necessarily gonna have all the connections and all the data, right? Like s— maybe it's true of some kinds of AI applications, I say gre— Lean is probably in that category. But for a lot of you, you might only have 1 or 2, and that's totally fine. Inside of these, many of these core datasets are either networks or hierarchies and can benefit from being stored and managed as a graph because of the flexibility, accuracy, and other kinds of benefits. So let's, let's wrap this up. So the conclusion for buyers is that companies will reap savings by rolling their own SaaS, but I would say the smart companies will do it by focusing on simple SaaS, where the tactical project leads to a longer-term strategic advantage. Um, and obviously applications where you've got like crud on top of a database are the first ones to go. And then focus the rest of your efforts on top-line value. That's the North Star. And then for builders, I believe SaaS will continue to thrive and will be able to do so by providing elegant solutions to intricate problems and push the envelope of possibility by using your own domain mastery and through technology. Thanks. We have like 2 minutes for questions. I don't know if anyone wants to come up. But, uh, please do if you have a question. Otherwise, thanks. Enjoy the rest of SaaStr.