2,000 AI Apps in One Year: When Users Build Without Limits

Speaker A: Welcome, everybody. This is Ryan Staley, and I am back. And I have a very special guest today. I have Mike Palmer, who is the CEO of, uh, Sigma Tech. Mike is basically had over 20 years of experience in the tech industry, has had roles as a cmo, the cpo, and now the CEO. So real excited to talk with Mike. They just recently crossed 100 million in revenue. Congrats, Mike, uh, on that. And, uh, welcome. Happy to have you on the show, man.

Speaker B: Glad to be here. And, uh, thank you for the, uh, congratulations. As I tell everybody, it's like it's the third inning, so it doesn't mean anything, but, uh, it feels good to be playing the game and, you know, and making some progress.

Speaker A: Well, yeah, I mean. I mean, it's. I think it's more than that, but I love the, you know, the downplaying of what you've done. So walk us through, man. Like, give us a real quick backdrop on you, your journey, how you kind of got to this point because you've had, uh, I should say some similar path to some that I've seen and completely different than others. So would love your take on kind of how you got here. And then, you know, what. What's it like being the CEO now of $100 million plus company? Wow.

Speaker B: I mean, the path here was definitely not. I tell everybody that, like, asks, like, how do you come as CEO? I tell m. Them, it's like, I have no idea. I, um, think the answer could be found your own company, at a big company, maybe be there a long time. Um, but I don't think there's a path. You know, in my particular case, the thing that I advocate for, someone taught me this expression a long time ago, is that you have your learning years and your earning years. Maximize your learning years is the only best advice I can give people. And in my case, that meant running marketing, running product, running engineering, doing small companies, working at large public companies, working at private equity companies. You see a lot of stuff. And as you and I were chatting about before, before we kicked, uh, off the recording here, pattern matching becomes, like, a pretty useful skill. When everyone thinks that something is brand new, most of the time it's not brand new. You know, you can look and think, well, there's some new to it, but 80 or 90% of it has been done before in some way, shape or form. And then how did we treat it? Then? What worked? What didn't work? That's what I feel like I've probably gotten at least a little bit good at. But on Your second part of your question, being a part of Sigma is amazing. Sigma, you know, has one of the highest growth rates out there, and there's just no better feeling than a high growth rate if you can augment it. It's probably when you have a high growth rate and, and a high retention rate because then you know your customers are getting a lot of value from your product. I love talking to customers who tell us, you know, how much we've changed, whatever it is in the area of the product they're adopting and how much better it is for them. That's why we do this every day. We move pretty quickly. It's just fun to add new. The best part about working, I think actually in a SaaS company in general is how fast you can change the product. For those of us that spent many years in hardware and software and you got like two chances a year to do that. You know, the fact that I wake up every day and there's just something new and different in Sigma is so exciting. But it's, it's good times. Data is a hot market to be in. There's still a ton of transformation out there, whether it's at the warehouse level or products like ours, obviously with AI, uh, then a whole bunch of things that no one ever put in our category before, like app development. There's just so much to work on.

Speaker A: Well, walk us through that, man. Like so, because I didn't do that, like in terms of what Sigma does and who it serves and the outcomes it creates. Could you just give us like a one paragraph summary of, of that? So everybody has context.

Speaker B: Got it. Uh, so many, many, many, many companies, tens of thousands of companies have moved from some sort of premises based data architecture and warehousing to the cloud, right? So this is the advent of the databricks, the snowflakes, the big queries. And that sort of movement is not slowing down actually, it's accelerating. But it's sufficient at this point that everyone kind of knows that market. And the question that Sigma asked was great. It's another infrastructure transformation, so who cares, right? Is it the IT team, you know, okay, they used to have a piece of hardware and now they have cloud, you know, okay, fine, or can we do something more, right? And when you put something in the Internet, it automatically becomes a collaboration opportunity, an access opportunity. So we wanted to change the lives for frankly the 90%, the people who leverage data to do their jobs, but almost exclusively use spreadsheets to do those jobs. So we invest in all kinds of expensive storage and Software and security, by the way, to manage data and then just extract it. So, you know, our marketers, our salespeople, our inventory managers, they just do stuff on their client because they know how to build a pivot table. They have some additional information to add. They want to write a scenario model. That's where the work, the real work gets done. And we wanted to get the real work into the enterprise system. So we took that cloud data warehouse transformation and we built an access layer on top of it that ranges from, you can do your technical stuff, you have a SQL notebook and a Python notebook, but equally, you could build a 6 billion row pivot table, you could do a 50 billion row scenario model. You can do these things in seconds. You can do them yourself because you have the technical skill in the interface. And then we've extended that, uh, we've made that into an embed model where you can build a data product. You can now build an entire workflow on top of this. So we moved from BI to spreadsheets all the way into application development. Uh, so just a continuous opportunity in the cloud to offer new value and new services that are a lot closer to what the user wants to get done. As opposed to, like the tech category that like, Gartner wanted you to be in.

Speaker A: Well, okay, so that's. I think that's a. I love that. Right. What does a user actually want versus the tech category? Design. So what are you seeing the most demand for from the end user side in terms of what you do and the workflows? Because you mentioned the old term workflow. Right. Yep. So what are you seeing bubble up as like the top three right now?

Speaker B: So I think everyone does need some form of BI tool because you do want to be able to chart data. Right. And I think that's fine. And everyone does that with us. And we have thousands of customers that do that with us. Great. But what we've. When I think about the answer to your question, I think, like, what is like the most organic adoption? Like, what's the thing that pulls.

Speaker A: Uh-huh.

Speaker B: And the data apps, part of our business is that example. We didn't even have like the concept of a data app a year ago. And in one year, all by themselves, customers have written over 2,000 applications. That for me is shocking. Right. That when you don't have a go to market motion, when you don't have a sort of a consultant out there.

Speaker A: Yeah.

Speaker B: Uh, this is just customers that take raw features and come up with their own use cases and deploy them and actively Use them. So for me, coming back to what a user really wants, they didn't really want in the end to look at historical data that was, that gets you part of the way down the path. They didn't want just to be able to manipulate that data to come to their own conclusion. And this is where I think about like, I don't just want to see a pie chart, I want to like Matt, I want to slice and dice that data. I want to, I want to understand what's really happening and I don't just want to forecast what's going to change, but I want to take an action when it happens. And this for me is where back to the point of what users want, they want all three and they want it all to work together. So I think that's a super exciting part of what we're doing. And to have so many hand raisers out there take it upon themselves to build without any direction I think is pretty remarkable.

Speaker A: Yeah, I mean so 2000 apps is a lot too. So are you seeing, is it like from the user perspective, is it kind of like a tool like cursor lovable like at that level where non technical users could build or is it more on the tech user side and they're just taking it the extra mile?

Speaker B: I don't think of us like any sort of um, translation layer of trying to take a technical thing and make it non technical or to make a technical thing have like a uh, just like a simple user face interface change. In our particular world it always starts with the data that's in the warehouse. And one of the things that's interesting about our apps is that you get the database for free. All of the data is your, is your schema and that's pretty unique. In most other apps you actually have to put the data in in order to use their predefined workflow. In our world the day is already there. So what you are doing is configuring how that data should show up in terms of what you want someone to do under certain conditions. So they're just building the conditions and then they're building the forms and the user interfaces. There is no code in Sigma, none. So we're not like a uh, natural language way to generate, you know, some sort of code back end m literally. And uh, one of the things that we're going to have so much fun demoing at the upcoming uh, warehouse summits here in San Francisco is the ability to do all this in natural language. You could tell a model, build me a, an Airline gate management application. It'll come back, it'll describe what it thinks that is, and it'll just start building it. It'll leverage the tables in your warehouse, it'll create navigations, it'll create relationships between forms, it'll create forms for you and then you can iterate. I mean, think about how many people would love to create some sort of automation for their own job, but would never have gotten above what I call the hurdle rate of getting access to a developer in order to do it. And that is exactly the sort of thing that we're seeing is that in the iceberg of applications, most of the applications that should exist are below the waterline. And there's still manual processes of things like file shares and spreadsheet updates that we're giving the world an opportunity to actually create an automated workflow with.

Speaker A: Yeah, I mean, it's true. Like, I see it all the time when I'm, um, working with clients. There's tons of use cases that people like. The most common question, like, I can show someone how to take something that took like 45 minutes into like 30 seconds. And you know what the next question is? All the time, why can't you just automate this? Yeah, why, uh, can't it just be automated? So I think that's the natural expectation because Amazon's done such a good job of reducing friction the way we buy personally. Right. Netflix as well, with like watching movies, like everything's become so easy and in the B2B space that hasn't always translated. So it sounds like you're kind of stepping into that path with what you're doing and the app creation side. So I absolutely love that, Matt. So let's go back to something you were talking about as we kicked it off. I wanted basically the ability. So folks had context on Sigma, um, and what you're doing. But you mentioned about pattern matching and pattern recognition. Right. Which is kind of funny. Something. That's what Tony Robbins talks about all the time. Right. The pattern matching, pattern recognition and pattern innovation are the three areas he talks about. So with you, man, obviously with your diversity of roles and back companies and all those different areas, what do you see as the pattern that's most consistent for, um, the best companies you've worked for and the highest growth companies you've worked for and maybe companies in the work, I should say, times of the journey that you were in, it's probably better because every company has ups and downs, you know what I mean? So what would you say are the patterns that You've seen that um, are prevalent in those situations.

Speaker B: I think one of the, one of the things that we're kind of doing in the AI market right now is that we want to take this. So first of all, I always refer to things as like a pyramid in this case where I refer to Maslow's hierarchy of needs all the time. But I can make anything a triangle. But in this particular case you get like tech change at the bottom and then a little bit more slowly you get product change and then a lot more slowly at the top you get people change. And the, it's a useful framing to think about. Like just because there's a bunch of AI change doesn't necessarily mean that people are going to do something radically different in, uh, a short period of time. People will do things in their consumer lives different. And you know, OpenAI is a remarkable company viewed from the consumer lens.

Speaker A: Mhm.

Speaker B: But from the B2B lens, I think that we've been talking for two years now about absolute revolution in, in our jobs. First of all, our jobs are going to go away. And then they didn't go away and then we weren't going to hire any more people. And then we realized oh no, we need people. Uh, and, and everyone's being whiplashed by the messaging. And the messaging is sort of self reinforcing because you get companies that want to tell you everything is about to change and they're the change agent investors. Pile onto this and tell everybody AI is the only thing you should invest in. And then you know, and you get the self reinforcing cycle. I think the reality of the pattern is all these things are really important, but you have to go through the progression. And those that are willing to invest through the progression of creating great products, leveraging changes in technology that recognize people, that recognize the actual jobs people are trying to get done and deliver that value proposition which inevitably takes more time. Right. It's not that you can then say like I'm just going to throw a veneer over the top of something and then all of a sudden revolution is going to happen. If you spend that time, you can build a really amazing and durable company that, that customers love. And in software love is measurable in retention rates.

Speaker A: Right.

Speaker B: They buy it and they buy it again and again and they buy more. If you're willing to invest through the cycle, you can build something amazing when new technology like AI comes around. You know, for me that's the one thing that I've learned and you have to learn how to build Product market fit. You also have to understand how to build for scale and durability. I think if there's something I've tried to learn, you know, through my career, is that how do you make those two things work together?

Speaker A: Yeah, I think that's a good way to describe it, man. Um, because I had kind of a rush of thoughts when you're talking through that, of what I've seen where you're right, there's just the veneer thrown over at the AI. Veneer, the AI, uh, language thrown over it. So when you're talking about that rigor, like, how do you systematically approach it? Right. Maybe this is the, uh, chief Product officer and you're coming out right as you're going through it, but through the CEO lens, like, how do you build it in your product? And then how do you build it in your people?

Speaker B: These are all great questions. One simple framing always is, like the three whys. Like, why do you need something? Why should we be the one to do it? And why should you do it now? Right. If you can't answer the three whys, then you're sort of making it up. If there is a why behind those things, you know, you probably have a good lead on execution. You know, why now when AI is obvious, there's a productivity opportunity. I think everyone sort of recognizes that in any particular area, though, the why someone needs something in their particular area is something that you can't just walk up to them and say, AI, that thing. And I think there are too many companies that are trying to do that. You have to really unpick it. In our particular part of the world, we deal in structured data. So I tell everybody, in AI structured, unstructured data, it's like, that's as close to a revolution as you can get. My metaphor for everyone is like, if I ask Chat to write a paragraph for me on George Washington, it'll do it and it'll probably be right. And if I ask it again and again and again, it'll change the words and they'll still probably be right. If someone asked, what is the revenue for Sigma? There is one answer, and if you ask again, it better be that same answer. Structured data and unstructured, very different. So we have to understand what is our customer trying to do. Are they going to take an 85% accuracy answer into an investment committee meeting? No, 85 and 0 are the same in this particular case.

Speaker A: Right.

Speaker B: It's either 100 or it's not. So we had to think about that job to be done Right. How do we get them an answer that is not just fast and easy, but accurate? So we add things in our product. We don't just give them an answer. We give them like a chain of thought. We show them what data sources we used, we show them what formulas, we allow them to change those things so that they can manipulate the answer and try to pressure test it. We give them some contextual information. You asked about revenue, but we also showed you bookings. Did you really mean revenue or did you mean nacd? Like, because people ask very ambiguous questions at times. So we do a lot of thinking not about what tech can do, we do a lot of thinking about what the customer is trying to do and then apply technology to do it. The second thing that we do is we measure everything. We measure every click in the ui, we measure every what the adoption rates look like, what your activity patterns look like, and then ultimately we take all of that measurement and then we come to talk to you about it. Because we can see a lot from data, but you still cannot hop over the. I'm talking to you. Did you get what you wanted? What were you trying to do? Why did you do it that way? What would have been easier? So that cultural respect for data, but not losing sight of the people is the magic. I think that creates great product.

Speaker A: Yeah, I mean, like, I mean, I think you nailed it, like, with thinking about what the customer's trying to do, measuring everything, and then having that, ah, contextual insight because it's. And then the accuracy. Like, I can't agree with you more, um, for certain roles. Right. Not every situation. Right. You're not, you're not presenting earnings numbers all the time and for every situation. But that's true. That's true. I'll give you an example, man. I was just talking to Executive earlier today, and one of the things he said to me was he's like, hey, like, I've tested a lot of different agents and I am not comfortable with basically something that I created from agents because, like, I don't have the detail behind it to validate that. That's accur it. Right? So whereas he's like in automation, he's like, I could say this is the exact process. This is what happened. This is step by step. He's like, but the black box element I can't deal with. So is that one of the things that, uh, your clients love about what you guys do? That is exactly 100% nailed.

Speaker B: You've nailed it. You know, that is exactly right. This is what we call Trust and transparency. Right, Transparency. And use the term black box, which is obviously opacity, like in trust and transparency. I know how the sausage got made. So I can. If someone questions my answer, I don't just say, well, AI told me, like, it's like it went here. This is where I got that. It got the data from here, you know, like, this is the formula got applied. Like I pressure test. AI shouldn't be the replacing me doing my job. It should be me doing my job more efficiently. And part of efficiency is understanding. Like it's still my job. I still have to justify the answer and understand how it was drive. So absolutely part of the thought process there is we want to reduce the workload by 80, 90%, but we don't want to lose any of the accountability or any of the reliability that you as a human have to take into that job.

Speaker A: Yeah, that's a great. Right, that's a writer downer. Get 89 there, but don't lose the accountability. All right, well, we are going to have you back for part two, but unfortunately we're up on time for part one. This absolutely flew by. Michael, where can people find you? Where can they find out more about Sigma and then we'll wrap this up

Speaker B: for today.com on the net. Uh, uh, maybe best, uh, if you're into the space that we're in, we hope to see you at one of the summits for, uh, both Snowflake and Databricks coming up. And we do, uh, lots of events with them in and around every city in the US throughout the course of the year. So, uh, come up to our booth and talk to us.

Speaker A: Excellent, man. Well, Michael, it was a pleasure having you on the show. Thanks a lot, man.

Speaker B: Thank you for having me. Great talking to you.

Speaker A: It was good. And then we will see you all on the next episode.