Gary Saarenvirta on the Cloud N Clear Podcast

Listen to the Cloud N Clear Podcast

Daisy Founder & CEO Gary Saarenvirta joins the Cloud N Clear podcast with SADA CEO and host Tony Safoian to discuss Daisy's AI technology and transition to the Google Cloud.

Original Source: Cloud N Clear on Spotify

The Cloud N Clear Podcast Transcription

Tony: Please welcome a customer from, you know, one of our favorite territories. We love Canada, we love Canadian customers, we love our office and our people there. I would venture to say that Daisy Intelligence was one of our first big customers in Canada. It's a great pleasure to have Gary Saarenvirta here today, Founder and CEO of Daisy Intelligence. Welcome to Cloud N Clear.

Gary: Great. Thanks very much, Tony. Look forward to chatting with you.

Tony: Likewise. No, legitimately, this is one of the first meaningful relationships we had up there. When we became your partner, I think, literally, we might have had two people, maybe three, in all of Canada and now we're like 45 or so. It turns out that it's not only a great local market, but it's a great pool of talent to serve all of North America. So, we have very heavy investments in engineering and in other functions and we're super excited. So, thanks for being on. I don't get to talk to many of my Canadian customers like this. So, thanks for being on Cloud N Clear.

Gary: Happy to be here and excited to be part of this, and share our story, and have a nice chat today.

How Did Gary Enter the AI Space?

Tony: You're, you know, an 18-year founder and CEO, pretty rare these days. And it seems like, based on what we're talking about, you're still very active, hands on, meeting customers, all the stuff that I think builders like you and I love to do. But tell us about, even before Daisy Intelligence, how did you get into the space and what got you to launch the company? How the company's evolved? Because you've been in the game for a long time.

Gary: I have a technical background. I have a master's degree in Aerospace Engineering from the University of Toronto, which is a great school, which is why there's a lot of great tech talent here, one of the reasons. Yeah so, big background in math and science, and not much of an Aerospace industry in Canada, and I wasn't ready to leave Canada at the time. And so, I kind of backed into working with large corporations and I was really shocked at how little math and science big companies use to make operating decisions. And so, I created this accidental career of bringing more math and science to business, always with a bent to deliver financial results and outcomes. And so, after grad school, I worked with a company called Loyalty One that runs the Air Miles reward program, a large coalition loyalty program, and they have a lot of retail customers and grocers and they got exposed to retail data and became one of the first worldwide users of IBM’s technology. Data Mining was a buzzword back in the 90s. For AI, I think the terminology has changed more than the technology has. I really got into playing with huge volumes of data using machine learning technology back in the day and I ran IBM Canada's data mining practice, and their data warehousing practice, and was one of their global go-to people for high end analytics work. And, along the way, I realized that machine learning in its current form, that everyone thinks about today, doesn't quite work for complex problems. And I think the world is living through the experience that I lived through 25 years ago, just by accident. And I was super excited, predictive Analytics can solve everything, but I realized that it's not quite enough. And so, I thought that IBM is a great business, great company, I'd say they're good at everything, masters of nothing, maybe, if that’s not too unfair of a commentary. And I thought I could take my analytic skills and my math skills, and my goal was to help companies operate smarter. And I thought this AI should be autonomous, that's a key defining feature. And so, we built some autonomous technology.

What is Autonomous?

Tony: What do you mean by autonomous? Let’s educate the listeners here.

Gary: Yeah, autonomous means a system that works with no human in the loop, does the thinking for itself, makes decisions. It can be very narrow scope, our technology doesn't make you a cup of tea, but it can solve its specific problem. I would say, like the ingenuity helicopter that just flew, it flew by itself. It's got a set of software and logic and there's no human in the loop. And I think I brought some of my Aerospace thinking, when predictive modeling or supervised learning doesn't really do the problem on its own, I went back and looked at what Aerospace has been doing for 50 years, which is military fighter jets have an element of fly-by-wire, the pilot doesn't actually fly the plane. And certainly, NASA rovers that are out there are autonomous systems, right? And so, the idea of taking autonomous logic with the science first. All those systems have the laws of physics in it and it's not data learned laws. It's like a theory, a human led theory. And so, we took that approach and we created theories of business first, and then get the data second. And I think that's been a unique differentiator in our approach. And as far as I know, we're one of the very few, if not the only one outside of the engineering and science domains, that's taken that kind of scientific method part of the business. Our vision is the autonomous enterprise, we believe in a future where computing machines improve our lives. That means making companies more profitable. That means it will lower the cost of living for you and me, because smart companies reinvest in price and innovation, which ultimately impacts the consumer. And then, ultimately, we want to make human beings’ jobs easier. I read a survey that said the fifth most important reason you take a job is because you love the job. That's pretty sad. It should be the number one reason. So, if we can take away some of the drudgery, let computers do the gory details, and computing, and repetitive, high-volume activity, and let people do what people are good at. And there's a place for both, I don't advocate replacing humanity, but that's our vision.

Tony: Totally, I think the human factor will always be there in terms of the creative arts, things of that nature that I think humans excel at, at least for now, a singularity may change that sometime in the future, but I agree with you. So much of business process and decision making, with all these tool sets being available, I think, in a more democratized fashion than ever before, more affordable than ever before, it's still just very early stages of how much of decision making gets automated. So, Daisy Intelligence was formed, it seems, out of your IBM experience, but also, maybe, the limitations you felt of what you could do there.

Gary: Yeah. And I thought I could do it better, faster. When I founded the company, I thought fraud detection was the killer app. I thought who was committing fraud against large corporations? It wasn't just individuals, it was, I think, organized activity, organized crime and terrorism. And I thought we can help lower the cost of insurance, but we could also stem the funding for some of these nefarious activities. And I thought that fraud was the killer app. And we started building a fraud detection solution. So, autonomously identifying insurance fraud, and then quickly go to bank fraud. And the world doesn't move as quickly as you want it to. And so, we're still, I think, in the early 2000s when we were doing this, we did a lot of pilots and work with big insurance companies, but they weren't ready to take it and really deploy it. So, I think we've been ahead of the curve all along the way. And that's a challenge when you're trying to grow a business, to be disruptive and different. So, around 2008, 2009, the big financial crisis, and the interest in fraud ended for a while there. And then we picked up in retail after doing a lot of work in retail around merchandise planning. So, helping retailers decide what to promote, what prices to charge, how much inventory to allocate. And again, our systems deliver the answer, even on the insurance side, there's no human required, there's no data scientist on the client side, we could just plug into the back-end system and say pay this claim, don't pay this claim, or you should charge this price, put 100 units of coke in that store, put some in the distribution center. It's that vision of autonomous. But that doesn't mean you replace a person still, because the human is still the boss. Humans set the strategy and the objectives, the AI takes care of the details, right?

What are Some Memorable Pivots in the Story of Daisy Intelligence?

Tony: Yeah, for sure. I mean, that pivot that you described, probably one of many, I think anybody who’s been around in tech services or software for 20 years or so, it has probably pivoted a bunch of times, and at least the successful ones, right? Like you have to read the market, you have to read the tea leaves. Your customers often take you in certain directions if you're receptive and paying attention. So, that's one sort of big pivot you described. What are some of the other memorable pivots? Going from fraud detection to retail, but other pivots, whether they were strategic or geographical, what sticks out in the story of Daisy Intelligence?

Gary: Yeah, I think, we started out doing professional services. And so, IBM had this great mentality, they call it first of a kind development. So, when you do development, you need to develop with a customer. The company is not going to fund product development unless there's a customer in the mix. And so, I thought, okay, we're going to do professional services with a view to build some software. And so, we did that with customers. We worked with a few customers at a time and kind of self-funded the development of the technology. So, at that time we had no outside funding, we ran about 30 million in revenue through the business over a decade, and we took 100% of the profits of that and I poured it into software development. And so, then the big pivot was when we started to get investors, throw away I’ll do work for anybody for anything that's in our domain of capability, and then say, okay, I'm going to focus on this product, I'm going to just do that, I'm going to say no to all these requests for services. And that was a massive pivot.

Tony: Isn't that pivot really hard? Just saying no to anything that you used to do and now don't want to do?

Gary: It’s brutal. As an entrepreneur, I still struggle with that, which is, yeah, I could do that, we could do that, we could do that, you know?

Tony: The other pressure is this FOMO about crypto and other automation. Like, it's all this stuff. We have this role as leaders to have our finger on the pulse of what's going on so we don't miss the transformational relevant things. But you have to really deliberately try to ignore everything else. It's really difficult.

Gary: Absolutely, it's very difficult. So, that was a big pivot point for international services and going from 100% non-recurring revenue to 100% recurring revenue. And we did that in 2015, 2016, I felt our technology was ready for primetime. We had built it with customers over a decade, we had tested it, proved that it worked. And so, not being savvy in the investment game, I think I kind of de-risked some technology and undervalued what that was worth initially. And were treated like a brand new startup even though I had written like 10s of millions of lines of code that we had spent $10 million of company profits building. And so, that was just not knowing. And my family didn't come from big business, my dad was an auto mechanic and he owned his own business, and my mother was a hairdresser and owned her own business. They were business people, but they weren't from this kind of savvy investment world.

Tony: So, what happened after that? You raised capital?

Gary: Yeah, we raised capital. And then we really started to focus on growth and we went from one customer to 15 customers over the last five years. We tripled, tripled, doubled in revenue, then the pandemic happened. And we've raised like $20 million, we did some seed capital rounds, like 5 million, and then raised like 15 million in a series A extension. And we've grown from customers in Canada to kind of 50% of our businesses is in the US retail focused on high frequency retail, grocery, drugstore, hypermarkets. And we’ve got one grocery customer in Europe, and insurance customers in Canada, some international brands, some multinational brands, and channel partners now in about seven geographies from the Middle East, to Europe, to Latin America, Mexico, Brazil, US. So, we’re selling all around. I’m thinking, because we're an enterprise software solution, if I'm looking at how many grocers can I sell to in North America, okay, there's a big market there, probably 100, probably wouldn't meet our clip. And so, if I want to get to 50 customers, I'm not going to get 50% of the North American market, so I need to go, okay, there’s 10 in the UK, there's 10 in France, there's 5 in Spain. And so, okay, I'm gonna go get one or two customers in all these geographies, that was a thought. And certainly, still focus on the US as the biggest market and we spend a lot of time there. But so, we've been trying to grow internationally. And I think we're right now struggling with the fact that we're so disruptive, that's the tough part of the sell. So, getting the product market fit. I think we rode the AI shiny bauble wave; everyone was excited about 3 or 4 years ago about AI. And so, we got a lot of growth from that. But now we're running into, okay, this retailer has 500 people who do the job that our software could do. And that resistance to change –

Tony: It’s very difficult right now.

Gary: Yeah. So, how do you sell a disruptive technology? That’s what we’re trying to figure out. It's hard to sell to those people who you may be altering their job description, it’s hard to do that.

Has the COVID-19 Pandemic Created a Demand for AI Software?

Tony: Definitely. Look, and there's gonna be some new jobs created by virtue of what your software does, but probably not 500 of them. At the same time, I think there's such sufficient pressure, especially in retail, to completely reinvent, completely have a better understanding of the customer experience, customers’ needs. Distribution models are being challenged, online versus delivery, home delivery, retail, curbside. So, I think, in some ways, I think it sounds like your software has a bigger role to play in a retailer's desired operation than ever before, because they actually have to transform now, don't you also feel that demand?

Gary: I feel that demand, totally. I think, I wouldn't wish the pandemic would have happened, but I think it's accelerated a lot of technology change, brought it forward a decade. And so, I see automation as a huge requirement in retail. The retailers are overwhelmed with their ecommerce channels, which if you're an omni channel, now that's all of a sudden grown maybe by an order of magnitude, or close to that, and you have to deal with that and your regular business. And on the insurance side, consumers want to be reimbursed in real time. And insurance has been digitizing already for a decade and I think this is just accelerating it. And consumers are more and more savvy, we love our smartphone apps and how slick and easy they are, we want everything in our lives to be like that. And so, I think there's pressure in all industries. And so, we see that opportunity for sure. And we're still a relatively small company, 50 people here, 15 customers. So, we have this story to tell. And how do you get out there? We're in this cusp of disruption that's happening and we're not the only company that's running into this kind of disruptive change requirement, where people step sideways, don't leave the building, but they step aside and let machines come in and take over some of the tasks. And I think that challenge is being faced in many industries and many different types of technologies. And I think once we humanity that out, and we're comfortable with that, and realize that the machines aren't taking over, they're not destroying us, it's good, I think then a lot of us will really ride forward. And hopefully, we're trying to help make that happen.

Tony: The efficiency potential is still great in many industries. A good friend of ours came from the Google ecosystem, you know, launched upstart many years ago, because they thought the way credit rating worked was highly inefficient and inaccurate. In the insurance industry, there's all this new insurance tech, as you would, disruptors coming up, some being acquired by the traditionals. But like, the way premiums were set was maybe not very smart, could be optimized. So, I think, all these traditional distribution modeling challenges, I think are still very ripe for disruption. And it feels, again, I agree with you, we've seen the same thing in terms of our company trajectories, don't wish a pandemic on anybody, but if there was ever a compelling event that would accelerate traditional conservative organizations in their digital journey, who thought they had 5 to 10 years to do this, they no longer do. And we think that there's a lot of positive, sort of, GDP impact potential, EBITDA impact potential, the fortune 2000, if some of these things are implemented better.

Gary: Absolutely. And we've seen in our customers, on average in retail, we've seen a 3 to 5% total company sales lift. And so, if you're in a 1% net margin industry, we could double net profit. Now, getting the people to buy into that, that's the hardest thing. But I see that in every single retailer we've gone to, from our smallest clients, on the order of 100 million in revenue, and our largest clients, 30 billion in revenue, we've seen the same types of metrics everywhere. Now, I'm not going to have my customer come to me and say, hey, Gary, you made us a billion dollars, I'm not going to hear those words, but we know we've contributed, through continual contract renewals, that there's value there.

Tony: Even in our space, where we're generally providing like the infrastructure and the plumbing and the services around that, I'm sure you're getting pressed to deliver exactly the kind of data that warrants a renewed agreement, just like we are. And I think that makes sense. Maybe you had a few years where people are doing AI because AI was cool, but now it's actually show me the lift to demonstrate the savings or the increase in revenue. And for those not familiar with retail, 1 to 3% is a crazy lift, like it is a remarkable amount of lift.

Gary: Yeah. When your net margin is 1%, or close to 0, or you’re flat, we can double the net income of a retailer. And similarly, on the insurance side, I think that there's 10 to 25% of fraud, waste, and abuse in typical insurance businesses. And so, we've been able to identify millions of dollars of fraud savings. And the barrier there is an insurance company has a human being to look at every complex claim or fraud case to adjudicate it or decide if it's fraud. So, human beings are a bottleneck, there's not enough of them. So, most claims just get paid. And so, that affects the cost of insurance for you and me.

Tony: Then to adjudicate it, you're right. It needs to be automated.

Gary: Exactly. You know, yeah, you're getting a $500 windshield claim, are you gonna spend time on that? But, at the end of the day, if you look at what's been the largest factor driving the cost of insurance in the last 25 years, it's fraud. There used to be something like, I think, 5% of whiplash injuries were a grade 5 injury, which is the worst, most severe. 80% were grade 1 or 2. I think now 80% are grade 5. That's because people have realized that this is an easy system to gain. That's why I started the company. I thought that fraud thing was such a home run, but it's just a willingness to go chase that. Again, it's a tough change. You know, it’s tough, if you're the CEO of an insurance company, to admit that it's happening on your watch, and that's a tough change.

Tony: I think there's going to be enough pressure in terms of digitization automation that even the most, I think, conservative and resistant leaders have to go down that path, because their boards and investors are going to push them to do that. Because we want to pay the legitimate claims quickly, because ESAT is critical, right? And we definitely don't want to pass on the cost of huge degrees of fraud on to our great customers and we want their premiums to be low and competitive. But in retail, I think more than any other industry, they've seen, obviously travel hospitality has been hit really hard, and maybe there's things they couldn't control in the process, for sure there's just nothing you can do when literally you're locked down, but retail has transformed the most for those who had the foresight and the ability to execute. And by the way, over the last year and a half, like we're in the business of delivering these types of solutions to customers. We're in the business of selling technology, like video conferencing technology with workspace and other things, communication, collaboration tools, we were just actually, and I'm sure you're in the same boat of like, holy cow, what if this pandemic hit 10 years ago? If we thought the economic impact was tough now, imagine if you couldn't order stuff, or get stuff delivered, or collaborate, or video conference, or a lot more parts of the economy.

Gary: When you look at it, so what's the value of automation and building more stable business? I get it done today with my people, but then the pandemic hits, and yeah, I totally agree with your point. Sometimes technology's overlooked and some industries more than others are a little more resistant to investing, you know?

Tony: Here’s the challenge, because human memory timeframes are like so short, here's the challenge for you and I as leaders, and the community of people who provide these services and technologies, like to make sure that once things get better, and open up, people don't naturally default to all the old habits and the ways of thinking, because I think, more than anything, what 2020 showed us is that we must be prepared for everything, every possibility, every outcome, every disaster, this is not going to be the last pandemic. So, I think we're trying to encourage our customers, okay, the muscle memory you're developing now, don't go back. Like, curbside pickup will probably now always be an option, and it should be, right? Don't go back to, like, we're done with curbside, everything's in store now. I think that's what we're trying to encourage in our customer base.

Gary: Yeah, I agree with that. I think there’s been a lot of great services, and ecommerce has risen like crazy, we're not going to stop that, consumers don't want to stop that, and the business should continue to support that. It's all about servicing your customers. And ultimately, our technology helps our clients service their customers. If I elevate your employee to free up their time, they're going to spend time servicing their customers, delivering on their mission, giving value to shareholders. So, when our software enables a 3 to 5% sales growth, it's not our software, it's that we freed up our clients’ employees, they delivered the 3 to 5%, we just supported them to do that. Technology is an enabler, right? It's a support vehicle. But the people who live there, they're the ones who get it done.

Daisy’s Tech Stack Journey and Transition to the Google Cloud

Tony: Let's pivot a little bit about the technology itself, because again, you weren't born in the Cloud, there was no Cloud 20 years ago, and probably 10 years ago, that was not like as mainstream, even though Amazon had a huge head start with the stuff that they had. Tell me about your tech stack journey, maybe how it was when you were just doing services, how it evolved when you were becoming a SaaS solution. That's really an interesting story I want the audience to hear as well.

Gary: Some funny statistics, when I worked at Air Miles, like the loyalty program, 170 retailers in Canada, so 80% of the population of Canada, we're collecting their transaction data. So, it wasn't the detailed T log data, but it was like one record per transaction. We had a refrigerator sized server, IBM P series. It wasn’t called the P series back then, I forget what it was, it was an AIX server, and it had 40 gigabytes of disk space. And we thought, wow. So, that was when I started my career. So, we’ve evolved into parallel computing. And so, we're doing this machine learning parallel computing on hundreds of gigabytes of data at the time, and that became terabytes quickly when we got into grocery. And so, I took my experience working with parallel DB2 and parallel hardware. And so, we had our own equipment, when we did professional services, it was always taking the clients’ detailed transactions. And so, our tech stack was built on parallel computing, DB2, then we got into some open source, because when you're getting into like 5,10 terabytes of data, the DB2 licenses were a little too expensive for a small company. So, we moved from there to start to look at Hadoop and these open source kind of databases. And so, we moved to Hadoop, even though running DB2 and Hadoop on the same query, same hardware, DB2’s got way more investment dollars. But at the end of the day, it was free. And we worked to overcome some of the challenges, but we built a lot of capability there. And then eventually, and we managed our own GPUs, so we did MPI parallelism first using multi core processors and doing parallelism on that. And then GPU came out and we started playing with Nvidia GPUs and putting our software, parallelizing that. In one hour, one of our smart young guys, graduates of the University of Toronto Engineering Science program, I went to that program and we hire a lot of people from EngSci, and I said, okay, here's the book on GPU, here's my crappy code that I wrote that parallelized it on MPI, and I think it's really easy to do this, 4 hours later, the code is running 100 times faster, because it went from running on 4 10 core processors to running on a GPU with 3000 processors. Then we said, okay, how do I scale from here? I'm getting more and more customers, I now have 100 terabytes of data, we're managing our own off lease hardware servers, break, fix, popping in drives. And so, I'm starting to build all this core competency and hardware management, and we're trying to write parallel computing software, and I'm going, okay, this is not making too much sense anymore. So, I thought we should look at this thing called the Cloud. It was that pivot. It was that pain of managing our own infrastructure and how long it took to deploy another server when we got another customer, and then we said it doesn't make any sense to be building anymore.

Tony: What year was that?

Gary: That was really like a couple years ago.

Tony: Oh wow, okay.

Gary: We were doing this until when we started to really move to the Cloud. We started exploring, we started playing with Azure like maybe 3 years ago, and then a couple years ago decided let's go to the Cloud. And we, you know, explored Azure versus Google. Those are the two choices. Being in retail, AWS was a bit of a persona non grata for our retail customers.

Tony: Yeah, we hear that all the time. For our other retail SaaS customers, there's just no go, no go on AWS.

Gary: Yeah. And then we had an RFP and a Bake-Off, and kind of moved to Google, right? A year ago with your guys’ help. Now, it's taken about 11 months to migrate, we're migrating our last customer. So, we expect to be done the end of next month. And then really focusing on becoming Cloud native. I'd say we forklifted, upgraded our infrastructure in step one and step two is to get Cloud native. And my goal that I challenged my technical team is I want to run my parallel software, that we all developed together, I want to run that on a million cores. We could do that and I want to run that test. I want to real time optimize for my clients, because, you know, for retailers, when we deliver here's the products you should promote this week, and my client says what if I swap that one product out? What happens? I want to push the button and give them an answer. I want to spin up a million cores and give them the answer and charge them 20 bucks for that enter key. That's where we want to get to, right?

Tony: When it costs you 50 cents and you want to charge 20 bucks, like that's the right model, right?

Gary: Yeah. That's where we want to get to. And we want to take away the headache of infrastructure management, which has been a barrier to our developers. Oh, my God, this hardware thing is broken, we got to get someone to fix it. And so, it's been a huge headache.

Tony: You’re gonna have a really exciting re-platforming journey ahead. Is BigQuery a part of that?

Gary: Yeah, we see ourselves moving to BigQuery and Kubernetes. And the interesting thing, my son works for Google and he's our Cloud engineer on our account, and we just had a QBR today, just before this call, with the SADA and Google team. And so, watching my son present is a weird feeling. Wow, my son's got his shit together, he's smart. I'm also his customer. So, I took a picture of the screen.

Tony: He must feel, oh my God, this is my dad, he’s judging me right now, and he’s also my customer.

Gary: I took a picture of the screen and I texted it to my wife. I go, check that out, right?

Tony: Pat yourself on the back a little bit, like mission accomplished, good dad vibes right there. And what you're describing is the stuff that we really get excited about. Because the reason that we went all in with Google is that we saw the Google path as being innately more transformational than the other Cloud options by virtue of being number 3, by virtue of being very good and heavily invested in the serverless ecosystem. We're seeing more Hadoop to BigQuery migrations than ever before, and I'm talking about massive scale, because most of these Hadoop customers, like Twitter used to run on Hadoop. And so, they moved to, by and large part, to the Google infrastructure and platform services. We’re doing many other projects like that, like huge companies, because they went through the same thing you went through. Like, when the data set got so big, the licensing fees on any other commercial database made no sense, they had to go open source. But then, literally, at this DC based customer, like we're testing query response times and Hadoop versus BigQuery, 80 times faster, 100 times faster. It's just unbelievable. But those are the kinds of things you can utilize to build new capabilities for which your customers will be happy to pay for, but also, in general, allows your solution to keep up with customers expectation of what they see in the consumer software world every day.

Gary: Yeah. We're super excited to getting 100% Cloud native and really taking advantage of these technologies, getting serverless, being 100% ephemeral, that excites me. And the capabilities that brings, the ability that we can scale to any workload, I’m so excited. And that helps us push our capabilities and deliver these new features to customers, exactly as you pointed out.

Tony: You just worry about the code, and of course the UX UI, all that beautiful stuff, but you're just not even worrying about the infrastructure. You can bring on as many customers as you need, they can get as big as they need to get, you can charge them whatever you need to charge. And also, in your global ambitions, again, I think picking Google points a presence and the network speed and latency in the Google Cloud globally is literally unparalleled. There's also data location, data residency needs, they just opened up a Poland data center, but I think they're in Montreal now, which has it’s different things in Quebec, different states in the United States, all around Europe, the Middle East, Asia, Africa, etcetera, and that's going to get really exciting.

Gary: That's one of the reasons we picked GCP, was just the global presence. And I think, also, you guys are a big part of the reason we picked them. When we did the Bake-Off, you guys really impressed us in terms of the migration team, the technical talent you guys brought to the table was like, yeah, we want to work with these guys, you know?

Tony: I appreciate that. I feel like you’ve gotten better in the last 2 years as well, and I'm so honored that we were part of that first big leap of faith that you had to take from on premise to a Cloud. And we were part of that decision to help you choose Google Cloud, that's what we're really proud of. Because what makes me proud of what I do is we truly believe we're making an impact to customers and to decision makers who are choosing this path. And just like you, you're passionate about your work, it's really meaningful work and I feel it's still very early stages of applied AI in retail and these other industries, I feel like it's the early days of Cloud. I mean, the Daisy Intelligence story is very common. Our number one vertical, and now we define it as a vertical where 2 years ago we might not have, is literally SaaS companies are like the number one customer for Google Cloud, and in a lot of ways, our biggest vertical in SADA. And I was on with PacketFabric on the last episode, and I was like the thing about Google is they don't want to be in your business, they want to provide the Lego pieces for you to build your business. And I think that is just a different sort of orientation towards partnership then some of the other Cloud providers have.

Gary: We're super excited, our team is excited to use the modern technology, and tools, and learn. The young technical people want to learn and have a career path. And I think this is the right move for, not only the business, it's for the people in the business as well. So, it's exciting, it's fun, fun for them. And I think we just see great upside continuing forward on this path.

Tony: Totally. We hope to be your partner for many years to come. Thank you for spending some time together here, Gary, and educating the market, and US, Canada, worldwide on your career path, the Daisy Intelligence story, your customers, and how Google Cloud has helped you transform the way you go to market. That's awesome. These are exactly the types of stories that people want to hear directly from founders like you. So, I really appreciate you being my guest.

Gary: Appreciate that. And thanks for helping us get to the Cloud. We really appreciate the support your team's given us. It's been a great year so far, and we look forward to many more.

Tony: Awesome. Talk to you soon.

Gary: Okay, take care.