The Infra Pod - Infrastructure cost will change how company runs in the future! Chat with Damon from Millwork Analytics

Episode Date: April 29, 2024

Ian and Tim sat down with Damon Fletcher (CEO of Millwork Analytics), after running as a CFO at DataRobot and Tableau for many years, decided to start a startup as he sees how infrastructure cost is no longer a feature but a product that he sees organizations need to run their business successfully in the future.

Welcome to yet another Infra Deep Dive podcast. This is Tim for Essence VC and Ian, let's go. This is Ian Livingston, helping Sneak Turn into a platform. Couldn't be more excited for our guest, former CFO from Tableau and currently the CEO of Millworks, Damon Fletcher. Tell us about yourself and what you're working on. Yeah, so well, thank you both for having me on the show today. Again, my name is Damon Fletcher, tell us about yourself and what you're working on. Yeah, so well, thank you both for having me on the show today. Again, my name is Damon Fletcher. I spent a long time as the CFO of Tableau Software and then recently launched a new product here in Seattle called Caliper,

and we're helping organizations with gaining better visibility into their cloud and other usage-based software spends. So really helping solve one of the critical problems for many of those in the CFO's office of uncontrolled spend and decentralized spending. So we're happy to do it. Amazing. I'm curious, what got you started? You have sort of this storied career of Tableau, data robot, and you said, okay, I'm going to go start a company in infrastructure cost modeling. What was the thought process? Why? Yeah, sure. So, you know, this was one of the

biggest problems that we had when I was at Tableau and then later DataRobot. You know, as we, you know, expanded back in probably 2016, we moved from being on a data center model at Tableau to moving our internal workloads to the public cloud. And then later also from our customer standpoint, we moved our customers from perpetual on-premises subscriptions to, you know, leading first with our sales team, with our Tableau online product. And, you know, what I learned from that experience was when you're in a traditional data center world, you know, CapEx goes through kind of purchase requisitions. It's a highly considered purchase. You model out the center world, you know, CapEx goes through kind of purchase requisitions.

It's a highly considered purchase. You model out the capacity needs. You go through a requisition process. That all changes dramatically when you move to the public cloud. You know, developers, you know, need to turn on test environments. They need to run different tests. They need to have, you know, developer machines. They need to onboard new customers. And all that can happen very fluidly. And what we learned is that it can have a highly unpredictable cost structure. And the existing tooling that was out in the market that I experienced, at least, really didn't allow you to ask questions and drill in and understand which of your customers are costing you the most or which of your engineers are forgetting to turn off the lights in the evening when they're done with their tests. And I thought there's got to be a better way. And so my co-founder and I kind of worked on this product

and it's really adding visibility to those who are trying to kind of solve those critical business problems. Interesting. As an engineer and having run engineering teams, this is a problem near and dear to me. I think of it as the costing in generally just giant balls of spaghetti. It's almost impossible to work backwards to figure out where is the money going, why, and how much of that is going towards product usage, customer usage, critical infrastructure load, what's the baseload cost, how much of it's actually variable, you know, how much of it's a variable as the company scales up or as the workload scales up and scales down, but also like, where's my opportunity for savings? Like it's a hard problem and as an engineer, it's hard to do and it's actually like, it's an operational burden. So I'm kind of curious, like,

have you given a lot of thought to how to like, how do you bring the developer into the loop? I kind of see what you're describing as a collaboration problem kind of between like the CFO and the finance department and sort of like engineering. It's like, we both want the same thing, which is like a scalable product that works. And it's also like cheap to operate. So we can hire more engineers or do whatever it is that we want to do. Right. Like I'm curious to hear how you think through that, that collaboration. Yeah. So that so I mean, that is the value proposition that we're trying to build with Caliper.

So one of my observations was, you really have to get the data into the hands of the people who are actually making the day-to-day decisions on when to turn machines on. So I think the best run companies in my experience from a long time working at Tableau are those companies that unlock the data

for their employees to be able to make better decisions. And so our product is really designed to create a language of data around multiple teams interacting together with one version of the truth, right? So, you know, a lot of the platforms that have been in this space before, you know, have canned dashboards where, you know, it's one size fit all, every single company gets the same experience. And I know from a fact, just from my experience, just with Tableau Data, very, very different usage of the cloud. And so then you throw in other tools like Snowflake and other things that we help companies with, and it's very different business problem than the public cloud. So what we want to do is unlock all those employees, both in engineering, in the FinOps role,

in the DevOps role, in finance, all to have access to the data, be able to see the trends and be able to drill in and ask questions of that data and be able to do that in one platform across multiple different public cloud, as well as other usage-based software and have one version of the truth. And so that's really the value proposition that we're bringing. So I'm very curious because this is definitely not a new problem. The cloud has been there for a while, and there has been a suite of products for the last decade, I guess, of giving you a visibility and optimizations for a cloud cost. But I feel like for the most part, what I've seen, and correct me if I'm wrong, all the

products tend to be more engineering specific, where like engineers start to use it. And then we export some data for a CFO to see. It's not really meant for CFOs directly using these products for the most part is my gut feeling, looking at all these products. And coming from your background, right, you're jumping into existing space. And assuming you want other CFOs to have a pretty predominant major user persona here, what does that product feel or look different that is meant for CFOs of the world to use that current product just doesn't have? I would say, when I look at the landscape of products that have been in the market for a number of years, you can tell many of them were developed by DevOps. So they're very, very focused on just a DevOps persona. What we're trying to do is have a rich analytics experience to allow question and answers of the data.

You know, and many of those tools that exist in the market are really focused on the public cloud only. We want to help with any decentralized spend category. So whether it be your Snowflake costs, your Datadog costs, your New Relic or Databricks, we want to help you with all those different data sets. And so it takes a very different mindset of engineer to build a product that allows you to drill in and ask questions of all these disparate systems, but doing it in a way that's one single platform. And so what we do for customers is we really build a lot of the data engineering that they would have to do on their own to get that data ready for analysis and be able to do that across hundreds of customers is a really complex problem. And so many of know, many of the other providers in

the space are very narrowly focused on the three major cloud providers and really focused on what I'll consider to be output-based measures like, should we have the customer commit to a reserved instance or not? That's kind of a common use case of these existing platforms. What I want to help them with is the inputs. What engineers are driving the highest usage? What customers are driving the highest usage so that the inputs can be managed in a better way? Because I think that's ultimately going to lead to more persistent cost savings over the long run. So it sounds like your model is to deal with the input. I'm curious to understand in your experience, does he Tableau and in other roles, has that generally been from your observations?

Is this something you learned through those experiences? Hey, actually, I realized at my time that this is the right model. What brought you to this conclusion? Absolutely. I'll give you an example. Several years ago, Tableau released our data management product, which was a data preparation solution that was going to compete with Alteryx, an existing player in the market for a number of years. And Tableau is historically priced on a user basis.

And so when we looked at the data for how our beta customers who are using the product were interacting with it, we saw, well, some customers are really using much more capacity in our public cloud than other customers. And so we had to change the pricing to accommodate for that different usage pattern of customers. And using the existing, and I'm not going to say the competitor name, but platform that was in this space, it didn't give us any of those answers about what customers were costing the most and the different behaviors. We had to do a bunch of rich analysis that was having to do very manually. It took weeks or months to prepare. And the good thing about our product is we do all that work on behalf of our customers so they can, you know, within 15 minutes of meeting with me for a demo, we can get them set up and looking at their own data and being able to see, like, which customers are costing them the most and things like that and not have to go through that kind of entire data

preparation process. So absolutely, it was kind of a learned experience that these problems within the public cloud and really managing these decentralized spend categories require a lot of careful analysis. And you can't just assume that a high-level kind of monthly dashboard and a reserved instance kind of modeler is going to solve your problem. That's just going to have you commit to a higher contractual price or committed price with one of the cloud providers. It's not actually going to lower your usage. It's just super interesting, and I'm thinking through my own experiences, Salesforce, Snyk, having cost being a huge part of both companies as they scale the revenue and ensuring that there is solid margins, which is what we all want. We want to get the free cash flow positive. We want great margins. I'm curious, how do you think about that feedback loop to the engineer and the decision point. There's some philosophy. A good example is Snyk.

Snyk sells this idea of shift-left security. If you move the security stakeholder's ability to influence developers' decision at the point of change, they will have a better capacity to anticipate whether or not the work they're doing is going to introduce a security vulnerability and maybe we can stop it before it starts. That's the whole thesis for security. I'm curious, in your mindset, on the cost side,

does that thesis fit hold for you? Do you do you think shift left cost modeling is like a real thing or not? I'm kind of curious to consider how you think about like helping develop the engineer better know at the time and if that is a feasibility or not. Yeah, I absolutely do. I mean, you know, when I joined DataRobot, one of the big push at the company at that time was to transition our business from on-premises to the cloud. And what we saw as a company is just kind of sharp accelerating cloud costs over that next several months. But we were also hiring significantly in our engineering department. And so some of that was driven by the kind of growth in engineers and the number of different tests we would do. It's a, you know, it's a machine learning, you know, AI company. So there's, you know, really rich data sets that require a lot of compute.

And so what we really needed to do was get the engineers involved in those cost savings. And so, you know, one of the things that we unlocked was having the data available for the engineers around leaderboards of who is saving the most money. And then we rewarded those people with things like spot bonuses or, you know, executive recognition. We would send emails to thank them for their contributions to helping, you know, drive savings that could be reinvested in innovation. And so without giving access to all that data to the engineers for them to help with that business problem, you're not going to get very far.

And so I'm a strong believer of providing data to the people who are going to drive the day-to-day decisions. So absolutely, shift left is the right strategy. We're working with a number of customers that are providing our platform, not only to the central FinOps and DevOps teams, but starting to unlock the managers and individual contributors in the engineering department to be able to see and understand the data. And I'm curious, like, how does, you know, there's this word that I started to see popping up.

It's not a new word. It's actually like a job role that I've seen pop up recently in the last couple of years, just like this idea of like this FinOps person. Like, how does this FinOps role fit into the view of like the way that you see an org? Like basically that, you know, we talked about like this collaboration between engineering and the finance that you want to come up with the best outcome for everybody. Where does FinOps fit in this sort of vocabulary between the developer, you know, yourself and your finance team? And like, how does that role work? Where does that fit in the equation? Yeah. So it's, I think you're right. It's a new role that's probably, you know, origins probably in the last five or so years where you're seeing individuals who, you know, sometimes have kind of a finance background, sometimes it's DevOps, but, you know, have an interest in driving savings at their organization and really want to enable kind of data-driven decisions. And so I believe that the best-run organizations have these individuals kind of paired or probably

a key stakeholder within the CTO's office, as opposed to sitting in procurement or sitting in finance and having kind of a person outside of your department. Make them a key business partner within your development organization to be able to unlock the data for all of your engineers to make those decisions. And I'd say that the persona of who are most of our kind of customers are right now is it is that FinOps person, whether that's in title or, you know, sometimes that person still sits in FP&A, but they're dedicated to the engineering organization. So I definitely think it's a new trend and it's for the best because I think many companies, if you're spending $10 million in cloud spend, they may not yet have a

person really focusing on this. And I think in my mind, if you have that right talent in that role, your ROI and hiring that person and having them put in programs and procedures to drive savings is more than compensates for itself. So maybe like as a macro view of thinking about how infrastructures spend or just infrastructure in general, I think we've been learning not just how the cloud works. Now we've learned how much the cloud costs.

From like organizational point of view, it seems like we're also learning how to adapt into figuring out how to incorporate this into our life. Like you talked about like personas and i guess what gamifications and i feel like that's a way to incentivize but doesn't seem to be a root cause sort of like fixing do you see companies run infrastructure costs differently in the near future because i think one of their first posts talking about why you started a

company you're saying like, we've been traditionally trusting DevOps, making the optimizations and the budgeting spending decisions. And a CFO seems to be more in the backseat. That seems to be what I'm reading and between the lines. I'm not entirely sure what does a new world look like?

Like, do we just need better drill downs and better precision of reports? And that's sort of a way we should run. What do you see teams running in a more efficient way with costs in mind? Pretty broad question, but just curious your thoughts. I think there's a couple of things that I'm seeing with customers. I think more and more customers are choosing to be multi-cloud. I think they feel like having the ability to choose different services in different clouds is best for their organization.

You're also seeing more and more companies go to technologies like Kubernetes within their cloud. So they're getting to a level of granularity that's different than just kind of pure renting machines or leasing machines from one of the major cloud providers. And so you're seeing that is a big trend. And then I think because of the way that, you know, the last couple of years have unfolded with, you know, higher interest rate environment, more constraints on capital, you're seeing more active involvement of the C suite and the CFO's office in kind of holding the team accountable to be more efficient within the engineering organization. I think if you go back five years ago when, you know, it was growth at all costs and, you know, there were so many kind of demands on the engineering team to release new products,

new features, you know, there wasn't as much focus on kind of managing your business more efficiently. And, you know, there was a lot of waste that kind of built up within your public cloud environment. Even today, I meet with customers and we find that they have machines turned on that were turned on for tests and have been turned on for six months and haven't been used. And it's just surprising to me that, you know, in a period of time where there's not clear path to capital raising for a lot of companies,

that, you know, there can be this level of waste that, you know, permeates throughout organizations that with small amount of cost and a little bit of tooling, you can make a huge difference. You know, so I think those are some of the bigger trends that I'm seeing is just, you know, more active involvement and then a shift in some of the strategy around how they move to the public cloud. You know, one of the things that I think about a lot, both as a builder, you know, as a startup guy, like I think about a lot, both as a builder and as a startup guy,

I've found some companies. Help me understand, if you were an infant provider today, you're an upstart, you're a startup, and you're trying to sell to the CFO, what are some of the anti-patterns you've seen startups come in and trying to sell to? Engineering teams are excited about some SaaS service or something, and they come to you, and you're experiencing a robot, or even in your experience trying to enable these CFOs. At the end of the day, what you're really doing is you're trying to enable the CFO,

and as a stakeholder of the engineering team, what are some of the entry patterns you've seen not work? Like, don't do this. And what are some of the things that you've seen say, hey, actually, if you come in with this angle, these things tend to work. I'm very curious to sort of get your experience because at the end of the day, like the premise of why you started this company and the premise of why this is really important right now is cost matters. Money isn't free. And we have to like, make sure that we build stuff that ultimately makes, you know, a profit at the end of the day. Like that's the goal. That's what the exercise we're all playing. Yeah, absolutely. Well, you know, as the CFO of Tableau and Dead Robot, at both organizations, I ran both IT

and finance and other back office functions. And so even at Tableau, we were early on in my tenure as CFO, we were still running the corporate cloud infrastructure within my group. And so I think as a CFO of a technology company in particular, you're always trying to try, and I always embrace new technologies. I mean, we did a robot within a six-month tenure, I think, after I was named a CFO. We implemented about 20 technologies to get ready for an IPO. And things like Workday and Coupa and other great technologies that are core to running a successful business. And so as a CFO, you're really trying to understand what is the ROI of those tools and how can

we make our teams more efficient? I love consumption price services. I think it aligns the value that you're receiving with the cost of the provider in a better way than traditional software subscription sales. But inherent in that model is a lot of risk. And as a CFO, the biggest thing that you're trying to manage is risk. And so I learned a lot of the behaviors from these vendors that were selling consumption price services to me at Tableau and then later DataRobot about their business model, about how they unlock our different groups to be able to very quickly add on more usage with little constraints

and sometimes little tooling to control that or little visibility, and then go through some form of a true-up process with our procurement team. And once you learn the behaviors of those vendors, you know, and what is driving, you know, their sales compensation for their sellers, and, you know, what's important to them, because they're also wanting to show consistent growth and predictable growth. And so once you better understand what they value as a vendor and supplier, you can better align what you value, which is more predictable and lower ultimate cost and more efficiency in your organization. And then both companies can win with a structured relationship.

And so I negotiated against some of the largest technology companies in the world to kind of get a good outcome for both my company and their organization around these particular business models. And I think it's, you know, it just really helps you as a CFO to really understand what are the incentive compensation, what are the needs of their investment, you know, leadership team at those companies, and then try to align and structure the deal. That's a win-win for both parties. So I maybe want to drill down a little bit into what questions do you think CFO actually cares about when it comes to like, is it just what team are spending on, what costs? It seems like there's actually a lot of different kind of questions that are very important for CFOs today. If you're talking about public cloud and usage-based spend, what kind of questions are important to better manage those costs? I'll use a couple examples. So, you know, maybe for Snowflake, what's important is understanding, you know, which queries are costing more than others, right? And that's not a readily available kind of metric that, you know,

you get in your standard invoice with that vendor. You have to kind of look at the data available kind of metric that you get in your standard invoice with that vendor. You have to kind of look at the data and kind of triangulate it with the warehouse sizes and the warehouse that were selected and the time they've run. So I think there's questions like that. It's understanding which users in the organization are spending the most on your snowflake costs because maybe they're in less critical functions and less critical analysis, or maybe they're storing too much data that's not necessary. With tools like Splunk or Datadog, are you kind of retrieving too much data, right? And so I think it's important to understand all these different inputs and try to align the business value that having telemetry for certain

features and how much value is that versus, you know, what is the cost to, you know, the sales operations team, you know, doing kind of a query of Snowflake every 15 minutes to see how the pipelines progression, is that necessary? Could it be every hour? It could be every day. And so without having that data at your fingertips, I think it's very hard to make that informed business judgment about what is the frequency, you know, what size of the data should be, you know, retained, stored, analyzed. And what we're trying to do is be able to help with all those different questions in a way that's very simple and easy to use the product as opposed to a customer having to send someone off for two weeks to go find an answer to one of those questions. We want to be

able to enable them to do that kind of instantaneously. Are there foibles where you've seen deals fall apart, like specific with startups? I think that a lot of our audience also tend to be entrepreneurs or investors in early stage companies do you have advice for like early stage investors or early stage companies and how to like talk to the cfo like you know maybe they're bringing a consumption-based model at that point they haven't they don't have you know the full suite of spend analysis and all the things like how do they get in right how do they land a deal and get your checkbox and are there things you're looking for in that to say, yeah, you know what, this is a reasonable bet and we have controlled the risk

on our side. Or, you know, I'm kind of curious to like, you know, what advice would you have for these people to help them land a deal? You know, in some cases, like they just don't have the offer, but there's definitely things that fall apart, like at the CFO stage. I'd be curious to sort of get your advice on. Yeah, absolutely. I think, you know, what I said earlier about de-risking the transaction. So having some form of cap in the relationship, especially in the kind of earlier periods of if it's consumption priced of what it can cost you, it can be one that, you know, has some level of, of true up associated with it, but it needs to, but there needs to be some give and get there so that you obviously don't want to make a vendor go upside down because for both companies,

it's important that there's healthy margins, but you also want to not have costs that can spiral out of control. And so I was very, very interested in making sure we had protections and contracts. If it's an emerging technology like a startup, there's a lot of risk for a company embracing that. So sometimes you want things like termination for convenience if the technology if it's a critical infrastructure item that if that company were to go out of business, you would need to maintain that for a period of time. So I think those are some of the common things that I've seen that we had to get from very, very small companies that we wanted to embrace the ecosystem and help them out. But it was such a critical process that we wanted to make sure that it ran well, even if the company, you know, that we were partnering with kind of faltered. I think as you go up the chain with larger technologies, you know, when you're doing, you know, 20, $25 million vendor relationships

with companies like Salesforce or other things, it's very important that you have some level of growth built into the contract, you know, your team, that you're getting other potential products in their suite unlocked to be able to use without any additional cost. So I favored trying to lock in longer-term ELAs with more of the strategic vendors that we knew we wanted to partner with for a number of years. And we trusted that they would continue to innovate. And so it's very important for me to lock in those deals early and have a more predictable cost structure over a multi-year horizon. And that was something we tried to do

with a lot of our critical vendors. All right, so we'll jump into the section, our favorite section called a spicy future. Spicy futures. It is exactly what it sounds like. What is your spicy hot take about infrastructure costs in general? What do you believe everybody will be doing or not believing yet in the next three to five years when it comes to costs? You know, I think this is probably just the moment we're in in history, but I see a tremendous amount of investment going into, you know, I'll call it vertical AI solutions, AI features in

existing platforms. And that level of investment in my mind is being highly subsidized by VCs at the moment and the public cloud providers. And I worry what happens as these technologies get further and further past the exploratory phase into actual implementation, how quickly these companies who've received a significant amount of cash are going to burn through that and whether or not there's sufficient capital in the kind of landscape to support these companies continuing their mission. Because I think it's a little wild, wild west from what I can see. And this is not only both off kind of my interactions with other, but actually seeing in customers and talking about real business problems where people are trying to solve with AI technologies. And so when these subsidies start wearing out and the cloud providers can no longer kind of infuse capital into some of these companies or no longer are willing to, and they try to extract more traditional revenue streams, what is that going to do

to this landscape? How is that going to impact the broader fundraising environment? Because there was a lot of capital burned a few years ago when it was growth at all costs and every SaaS company was trying to compete to steal the sellers, every other company's sellers. But there's almost just as much waste happening now in a different way, which is in pure infrastructure. And unless we get a very rapid decline in the cost of serving AI, this is not going to end well for a lot of VCs and a lot of smaller vertical AI startups, because they're not going to be able to get funding and they're not going to be able to

pass those infrastructure charges on to their customers. And so, you know, my bet as a kind of a leader is how can I help as many of those AI companies be more efficient? You know, some of the challenges I face is they're so singularly focused on getting to market and growth, and they feel like they've got a lot of capital that they are not always as interested in managing it efficiently. I find that very troublesome, but that's not for me. But they'll face the accountability that'll come in a few years when they run out of capital and they don't have a clear path to grow. So that's my hot, spicy take on some of the challenges that we'll face as an industry in one to two years at the current pace we're going. So I think there's two sides to the ledger sheet or the balance sheet. You know, on one side, you have the productivity gain.

At the end of the day, like the way I think, you know, with the AI, this is true for all technology at the end of the day. It's like I'm buying something because I expect some like productivity increase, which means the productivity increase will reduce my cost basis. So I'm going to make more money somehow. Like I think that's like an AI fits into that equation very well. That's one side of the equation. That's like, I just bought an AI CRM tool and now I need like a 10th of the sales agents or something like that's the proposition. So on one side, it's like, that's, there's a proposition there that that could be

true or false. And the other side, there's like the cost that that would actually create for the CRM provider to actually run that thing. And is your concern the left side of the equation in terms of like, is it actually giving productivity improvements or is it on the cost side? I guess maybe it's both. I'm kind of curious to see in your thought process, where are you most worried about? Yeah, well, I think there's absolutely productivity gains going to happen. I'm skeptical that customers are going to want to, traditional companies are going to pay more for those productivity gains. And so it comes down to the revenue model that the AI companies are going to have right now, because it's exploratory

and it's early stage conversations. And there are a lot of times not charging for that. I mean, I can go into open AI right now and have it rewrite a press release for me and don't pay it a penny, right? And so when are those costs going to be covered by revenues for those technology companies that are providing that AI? We use some AI-enabled products. We use one in our go-to-market motion where it saves you dramatic time in writing emails and things like that to potential prospects. And, you know, I've used some of the tools from like LinkedIn and

others that help, you know, kind of rewrite your posts and all of them provide incremental value, but it comes down to, am I willing to pay for that incremental value? And I'm not quite sure that payment on the revenue side of the ledger in your example is going to happen for most of these technologies. But the cost is certainly there. And that's all right now being subsidized because there's a big bet that there will be some winners. Where will those winners accrete to? And I think they're going to accrete to the major providers. I think it's going to be the companies at scale that can charge a very small fee for some of these technologies are going to dramatically win against the smaller companies who would need to charge a larger fee, but are not going to be able to get across the finish line

in a sales motion for that larger fee. But you have companies like Microsoft who can add it to your ELA for $20 or whatever it is. That's where all the value is going to accrete to. Interesting. So your basic bet is saying that AI is actually a win to the existing companies. So cloud was interesting, right? Because cloud created this, you have on-prem and then there's cloud. And it was this architectural shift, change and go-to-market motion. It was just like a fundamental reinvention of the way that we bought, sold, and used and built software. And I guess what I'm hearing from you is AI isn't really a fundamental shift in what we're doing. And so you have these AI upstarts, you know, like a pseudo AI CRM or whatever that are trying to reinvent CRM. But what they have to backfill

in order for a company to be successful with is so big. Plus, they're going to add some AI features. You feel like it's the, you know, sales forces and the codes of the world that actually will probably end up gaining the most value, I think is what I'm the, you know, Salesforce's and the code of the world that actually will probably end up gaining the most value. I think is what I'm hearing from you saying is because they actually have already built the feature set and then they can add the layer on top. And that layer can be quite cheap. And they already have the existing relationship. Yeah, that's exactly right. AI to be added to my existing Microsoft account or Salesforce account or ServiceNow account, as opposed to I am an AI first company that comes up with the new modern version of these.

I feel like those companies are going to have a tough road in an environment that it's still soft out there for selling technology. It is not a robust buying environment. I'm helping people save money and they're still sometimes just worried about, and I'm a low cost product and sometimes worried about the cost of the service. So I think they're going to have headwinds. Now, for investors, the reality is you'll make your bets and there might be a lot of these companies that get acquired, but there's also regulatory risk right now. As you saw some deals recently recently kind of couldn't get through the UK. And so if you have an environment where acquisition is not a clear path, it's hard to generate

sufficient revenues to cover your cost structure, there's going to be a lot of these AI companies that are point solutions to go out of business. And then all the value will accrete to the bigger platforms. I have no issue with this. Those are all great companies. They drive the US economy and they're all going to do quite well. But I think it'll be, there'll be only a handful of winners when it's all said and done. It's such an interesting conversation. I'm curious also, you know, when you think of these, I don't know if you saw Cognition AI or if the Devon, like these semi-autonomous sort of agents,

like if you were you know you're sitting in and i mean maybe even like your product is like well we're going to build a semi-autonomous like cost saver that goes around and like find the only cost saving does it for you and like maybe there's a human in the loop maybe there's not like putting your cfo hat on for a second how would you think about even being approached with that as like a proposition like a productivity proposition how would you think of like something like that? I think the underlying question I'm really trying to ask is, with your CFO, how would you approach an idea where someone comes and says,

hey, we're going to reduce the number of heads you need because now your developers are going to be three times more productive because they're using this semi-autonomous AI thing. Do you think people believe that value prop today? And if someone came to you with that value prop, with your running a Tableau data robot, even your own company, like, how do you think about like, just making the numbers and the bet and the risk work? Well, you know, if you take the word AI out of that business problem, I've dealt with that for, you know, 20 years now, where someone comes

to you and they make a justification to make a big technology purchase. And they say, it's going to reduce headcount. And the first thing I say is, which heads exactly and when are they going to be put into different roles? And so then it becomes real for that individual. So as long as there's executives willing to kind of stand behind some of those statements and the vendor has to help with that, educating the company on where those heads will come. And I think there's going to be some customer service roles, technical support. I think AI is going to dramatically cut the number of knowledge workers in those particular roles based off how it can interact in a written form with individuals. And so that's going to be pretty dramatic. In my industry, I think I've seen some of the automated recommendations that come out of some of the tools. I actually met with a CEO last about a week or two ago that got burned by using some automated preserved instance modeler and then

overcommitted himself and now wants out and it's really frustrating with him. So sometimes it's these technologies are not as quite as accurate or they your business and facts change and suddenly you're you've taken a recommendation that was automated that didn't inform human judgment and it can come to the wrong answer. So I think in my business, it probably will evolve. We've looked at some AI technology to help us summarize the visualizations in written form. And we're not happy with some of the way that that content gets created and how useful it would be for customers. And so we haven't yet embraced that technology. We don't think it's quite ready yet, but certainly down the road, I think it could be an exciting

way for us to interact with our customers, to provide more written summaries and key insights of the visualizations that we provide our customers. Awesome. That's definitely a lot of stuff to digest, but it's really cool to see your perspective, especially given this infra landscape has been mostly DevOps, as you mentioned, from our side, at least. So where do people learn more about Caliper, the product and Millwork Analytics? And is there a way for folks to check out what your product does? Some information of learning more will be great. Yeah. So calipersoftware.ai is our website. You're happy to go there.

You can request a free trial. One of my sales team members are going to jump on and give you a demo. Set up the product takes about 15 minutes, and you're up and running with your data. And I'd say every customer that we've onboarded on a trial has within a week or two, we'll have our first check-in. And they've generally 10x the cost of the license in savings. And so it's very easy to use. There's no training. There's no enablement.

You're up and running within a few minutes and you're immediately driving savings. I'd like to quote, don't spend a lot of money to find out where you're spending a lot of money. We've really tried to make our product as inexpensive as possible for our customers. We want to reach 20,000 or so FinOps professionals in the world and have every single one of them use Caliper. So, you know, contact us at calipersoftware.ai

if you're interested in learning more. Amazing. I mean, thank you so much. It was such a pleasure. I hope you have a great rest of your day. All right. Thank you both.