The Data Stack Show - 256: The Rise of the Citizen Developer: Solving Business Problems with Alteryx and AI with Andy Macmillan

Episode Date: August 6, 2025

This week on The Data Stack Show, Brooks and John chat with Andy MacMillan, CEO of Alteryx. Andy discusses the evolving landscape of data and AI, focusing on empowering business users to solve complex problems. He explores the concept of "citizen developers" and how tools like Alteryx can bridge the gap between IT and business teams by democratizing data access. The conversation also emphasizes the importance of creating controlled environments where business users can leverage cloud data platforms and AI technologies to reimagine workflows, without bypassing governance. Key takeaways include the need for organizations to enable innovation through accessible data tools, the potential of AI-driven agents to transform business processes, the critical role of employees who understand their business functions in driving technological transformation, and so much more.Highlights from this week’s conversation include:Andy’s Background and Journey in Data (0:54)Early Web Development at General Motors (2:23)AI Challenges in the Enterprise (9:03)What is Alteryx and Its Value Proposition (11:25)The Importance of Empowering Business Users (16:10)Bridging the Gap Between Data Platforms and Business Users (20:04)Evolution from Desktop to Data Cloud (25:28)Access and Governance in the Cloud Era (27:57)The Return of Local Data Work and AI Governance (31:24)AI Data Clearinghouse and Governance (34:11)AI-Enabled Workflows and Business Impact (38:13)The Future: Agents, Data Platforms, and Business Logic (41:05)How to Get Started with Alteryx or Learn More (46:54)Product Management Lessons for Leadership and Parting Thoughts (47:56)The Data Stack Show is a weekly podcast powered by RudderStack, customer data infrastructure that enables you to deliver real-time customer event data everywhere it’s needed to power smarter decisions and better customer experiences. Each week, we’ll talk to data engineers, analysts, and data scientists about their experience around building and maintaining data infrastructure, delivering data and data products, and driving better outcomes across their businesses with data.RudderStack helps businesses make the most out of their customer data while ensuring data privacy and security. To learn more about RudderStack visit rudderstack.com.

Hi, I'm Eric Dots. And I'm John Wessel. Welcome to The Datastack Show. The Datastack Show is a podcast where we talk about the technical, business, and human challenges involved in data work. Join our casual conversations with innovators and data professionals to learn about new data technologies and how data teams are run at top companies. Before we dig into today's data. episode, we want to give a huge thanks to our presenting sponsor, Rutter Sack. They give us the equipment and time to do this show week in, week out, and provide you the valuable

content. Rutter Sack provides customer data infrastructure and is used by the world's most innovative companies to collect, transform, and deliver their event data wherever it's needed all in real time. You can learn more at RutterSack.com. All right, we are here with Andy McMillan from Alterix, Andy, so excited to have you on this show today. How's it going? It's going well. It's going to be a lot of fun today. You guys run a great show. Awesome. Well, thank you for the kind words. And Andy, before we get started here, will you just give us a quick background? Tell us how you got started, what you've done, and what you're doing today. Yeah, I'm the relatively new CEO at Ultricks. I've been here only about six months, but my background's quite technical.

I started off as a developer building websites for General Motors and worked my way up through product management roles at Oracle and Salesforce. And this is now my third CEO job. Awesome. Andy, so really excited to have you here. One of the topics I'm really excited about talking about is this analyst citizen developer, somebody that's embedded in the business, but has some technical skills. And I think Ultricks has been a really good, done a really good job over years now of serving that customer. So I want to talk about how that role might be changing with AI. And then what are some topics that you want to make sure we cover? Yeah, I think that's a great topic. And I'll sort of jump off that topic. I think there is sort of a new chasm emerging between these new big cloud data platforms and empowering people in the business to use all those capabilities, people spending a lot of money to invest in these platforms. And I agree with your point of then how do I use those investments to reimagine my business with AI? And so I'm looking forward to talking about that with you guys. Awesome. Yeah, some topics I am very excited to dig into as well. So let's do it. Awesome. Let's do it.

All right, Andy, welcome to the Datastack show. We are so excited to have you on. You guys have a great show, and it's going to be a lot of fun to chat with you. Cool. Yeah, we are excited. Like I said, you shared a bit about your background in the intro. Let's go just a little bit deeper. Will you just walk us through, again, a little deeper than the kind of super high level from the intro,

the kind of arc of your career, how you got started. And I know I really want to dig into you. were a web developer at General Motors early on in kind of the internet era. So I'm excited to maybe dig into a story or two there. There's a fun tidbits there that we talked about for the show. That's right. Yeah, I've been doing this a long time. So I saw a little bit like one of these old guys now talking about back in my day.

But it wasn't a while ago. It was the initial internet sort of dot-com boom. And I just graduated from Michigan State and had a background and doing some software development. And at the time, a lot of what we were doing was figuring out of how to, I was working at EBS, which was sort of the big IT outsourster that did work for General Motors. And we were doing work with GM to connect into their backend kind of legacy systems, a lot of mainframes to actually put vehicle information on the internet.

So I worked on the early versions of GM.com and Chevrolet.com and things like that. And we actually had to rebuild all this infrastructure. We're using Java to sort of connect to these old mainframes and figure out, like, How do you present this kind of information on the web, which at the time was really new, the idea that you could browse an inventory or shop for a car online. Yeah. I want to go in, you know, I don't know how far back the web archive is going to go, but would love to see screenshots of like, you know, Chevy.com from back to the first now.

It had its challenges. I mean, there were, you know, agencies had one view of the world of like how the things should look. And again, we were on like really old technology, again, not to go into the wayback machine, but like, you know, we're all on really small. all modems. And so, you know, we'd get these giant files from the creative agency that would take, you know, six minutes to download and they want that to be in the background image. And what was possible was also quite challenging and just a lot of really old tech that was, but it's really fun. I mean, I think this is what people get excited about. Like, how do you go

from one generation of technology to another and how do you solve those problems? It's something I found I really enjoyed, even though at the time it was really challenging. It was really fun to try to figure out. Yeah. And you worked on a specific app back then that I mean, speaking of just old technology, and sometimes right, old technology just persists, the window stickers on the cars have the price and all the kind of specs of, you know, here's what this model is, which is funny because, I mean, the way we shop for cars now has totally been transformed. But in the old model, right, you walk around the lot and you just go look at all the cars and read the window stickers. You helped bring that kind of online.

Tell us a little bit about that. Yeah, it was, it's funny. Even that you write, like, the names for things. Like, you don't think about it being on the window anymore. It's on the website. But it used to be. You're right. The website stickers.

Yeah, exactly. And every car, it's still to this day, a car in most states, it's required on the, usually the driver's side rear window, there's this big giant sticker. It's like a full sheet of paper. And it has all these esoteric codes on it from the car dealership telling you what all the options are on the car. And then, you know, all the pricing. And so legally you have to present that.

And so we had to figure out how to build this thing on the website. And all of those codes were in these really old mainframe. systems, and there's all this logic on how the window sticker gets built. And so, yeah, I got to write one of the original versions of that for the Jonah Motors platform. And as I was telling you guys before, the show was funny for years. I would go, you know, just look up a card, click the button and see if the same window sticker code would run. And I think it was until about four years ago, the code that I wrote in the late 90s was

still running on most of the websites. And I think it just goes to show, you know, code can last a really long time sometimes. And I would argue I was such a great developer very early on in my career. People introduced me to product management. So I'm not even sure that was great code. but it did apparently work. It was still running on the Internet for quite a while. Yeah.

I love it. I love it. Well, okay, so tell us moving forward, web development at GM early on. What came next and how did you end up where you are today at Altrux? Yeah, so I had an experience early on in my career. Maybe some of your listeners have had it too where I felt like it was really fun working on the technology, but there was a whole business around the technology that I didn't really understand. And sometimes we built something really great, but we didn't win the business.

And sometimes we built something so-so, but it ended up being a big contract and we'd win. And so I wanted to know the business more. And so after EBS, I went to University of Edinburgh and Scotland and got my MBA. And then after that, went to work at a product company in Minnesota called Stellant, which was maybe back to what I was doing at the time, very web content management. We used to publish a lot of websites for companies. And that's where I got into product management. And I found I just really loved that job, just sort of figuring out what do users need,

and how do I get an engineering team to go build and how do we take it to market? That company, Stellant, ultimately got bought by Oracle. And so I went over to Oracle as part of that acquisition and spent four or five years at Oracle and the middleware division, sort of helping people do big transformational projects using Oracle technologies, again, in the product side of the world. And then Cloud came along and Salesforce seemed like a really interesting place to be.

So I jumped over to Salesforce, again, as a product leader, and worked on something called data.com, which is now a big part of their data cloud strategy, and ended up being the general manager of that product line, ultimately was the product group COO at Salesforce, and then started getting calls to go do CEO jobs. And I thought, that sounds really interesting. And so I was the CEO at Acton Software

in the marketing automation space for a couple of years. Then I went over to user testing, which I was at until about a year ago, which is a platform that helps you get feedback from real users. We grew that a bunch, took it public, on the New York Stock Exchange, ultimately sold at the Toma Bravo for $1.3 billion, part of a merger with user Zoom, which was the number two player in the space. And while there, started playing around a lot with AI. We started building a lot of custom GPTs, partnering with chat GPT.

And it was really where I discovered the need for how enterprises are going to have to rethink, not only their business process, but how their data comes into AI. And I actually, when I left user testing, my plan was to start a company that would. solve that problem. And while I was thinking about that, I got a call from Ultricks saying, hey, we're hiring a CEO. And I knew Ultricks, you know, I was where the company. And it was a little bit of a eureka moment of like, oh, that's exactly the product they should have been using to solve this problem. And so, you know, felt like something I just couldn't miss out on. What any, and I know we'll get into this kind of as, as it applies to Altrix as the show goes on,

but while you are at user testing, you know, it sounds like pretty early adopter, some of these AI tools, what were just some of the specific problems and kind of pain points that you were running into of like, man, I could do this if only. I think when you start playing around with AI, you very quickly start to be amazed by what it can do, but also disappointed with all the untapped potential that you're not able to sort of grab a hold on. And what we found was we were doing a lot of really great things where we would build something creative from a relatively small number of other creative assets.

So you can upload 10 documents into ChatGPT of marketing content and ask it to create a marketing brief on something. And it does a really good job. But it's also really hard to then turn to ChatGPT and say, hey, that was really helpful. What are my sales reps in the Northeast just left? Can you recut all our territories for the Northeast? Like, it doesn't know anything about your business.

It doesn't know how to do that. Or, you know, hey, I'd like pricing this deal. Can you look at all the deals in this region with this partner from the last five quarters and help me put a deal together. Maybe I could go try to manually find all that data, put it into a small enough number of files, upload it myself into chat GPT. I'd also then have a governance problem.

Like, is it okay? Am I allowed to do that? Should I put all this in there? And so I think the idea that AI will be incredibly powerful when it understands your business is clear to people. I don't think anybody understands, how's it going to understand my business? And I don't know any company today that is going,

hey, you know what I'm going to do is I'm just going to let whatever AI tool I buy just crawl my cloud data platform and vote for the best. I mean, that's not happening either. And so I think there's a lot of value of resolving that inherent conflict. And I think that's the opportunity. Yeah.

Give us the kind of quick overview, kind of what is alterics and then maybe part two of that. Could you tell us like, why in the context of, hey, I'm about to go start my own business to solve some of these problems I'm running to

around AI and like how do we get it safe access to data? Tell us what Altrix is and then like why are you so excited and why is this just like the Eureka moment, like why is this the place like that was the right fit for you and that you're excited to go and solve this problem? Yeah. So first up, I'm a big believer that the really magical things that happen with software is when it democratizes the ability to do something. And so part of what I love about Altrix is the company historically has been around more than 20 years. One of the main things Altrix did in the early days was frankly, democratize the ability to do really incredible things with data to people that weren't software developers. So if you are pretty good with Excel, you can do amazing things

with Alter. So you sort of understand data, but you're not necessarily a developer. It used to be a lot of folks had to deal with that at the desktop level. So, you know, you get sent a bunch of reports and a bunch of files and, you know, you're on the supply tank team or the accounting team and you have to, you know, hey, reconcile the end of month financials and you get all these spreadsheets. and Ultricks was a tool that would let people pull that into a single canvas and do really amazing things. Do all the prep and blend and clean up and then automate it. So you can say, well, I did it once.

I don't want to have to do this every month. Like, how do I turn that into a routine that just magically happens? And so I think that's the first thing that Altrix did really well. Now what I'm seeing is sort of get to the second part of your question. How do we do that again now that we're building these cloud data platforms? And so the data is not necessarily spread out all over the place. It's often been aggregated. But I think the way to access that data and understand it and do things with it has not been re-democratized, if that makes sense.

And so I think that's one big area. And then I also think in this idea of AI will transform the way a lot of businesses run, it will transform the way a lot of jobs operate, the way a lot of functions operate. I believe it will be the people that know those functions and know the business, that are going to be the ones who can reimagine that process. and think through that. And so I think we can play a role in helping empower those people to say, you know, if only I had access to this set of data and this amazing AI technology,

I could change the way we do the thing that I do. We can go to those people and say, great. Like, we can give you in a controlled environment a way to do that. And I think that's what's got me really excited about altrucks. Love it. So cool. Well, John, you mentioned in the intro this term citizen developers. You have a, I mean, I would say a very.

technical background, but you've also been like very close to the business and many and I would say probably most of your roles. Let's just talk a little bit about kind of like citizen developer versus like being on a more technical team. Andy, you mentioned, you know, the cloud data platform. That's going to be owned by the technical team, right? Like I said, there's a lot of kind of overlap here and would love to just camp out on alterics users and talk about the kind of technical versus business bit for a second. Because John, I think you have. a lot of, I think, unique insights there. Yeah, let me start off with kind of a fun anecdote.

So I first was exposed to Altrux over a decade ago. I think only a couple of years after, I think, you guys rebranded to Altrux. And I still remember. So I walked into an analyst office. I'm on the IT team, more of a database DBA role. Walked into an analyst. He was a friend, Analyst Office. And on his laptop, he was like, oh, like, you'll never believe.

leave this thing I downloaded and like, you know, and he's like showing me. I was like, I can like connect the spreadsheet and this file and do this. And he was like a client facing analyst. So like he had all sorts of requests coming in from clients. Like we want to see our data this way or that way and we want to do this or that. I was so excited about it. I still remember the like dual like feeling of one like, oh, my IT hat is freaking out a little bit and then like my you know and then but on the other side like you know being really happy like for this analyst who just like bought back you know hours of his week every week so that was kind of one of my first exposure to this like type of tool and since then have i think gone even more

toward the like belief that this like citizen developer this like enablement of the analyst is really important and it's just it's probably possible but it's nearly impossible to replicate it by trying to insert like a product or a, or requirements analyst or something from, you know, from an IT org. Like it just, for whatever reason, it just seems like the analyst or the, the people that are embedded in the business that are in marketing or sales or ops or whatever that do it day in and day out can, they can just get further with really understanding things more deeply and therefore equipping them with tools is the better move than versus

like trying to pull people into like a central ID group and like start projects and put project managers on things like I just haven't seen that work as well I really don't I think the business has context it has scale and it can make its own decisions about prioritization and which things are most important yeah and so I think when they're empowered a lot of good things can happen and to maybe connect to the analogy all the way back to my first example I like to compare I think one of the easiest to awkwardization examples is how we build websites today every company has web infrastructure. There is development that happens around that.

There is an SDLC process to change how the code is deployed. But nobody says, you know, in marketing, every time you want to change a word on the website, file a ticket, right? To the requirement, send it over to IT, and we'll go and encode that change. At some point, we sort of sat down and said, hey, there's a certain domain of the website that our IT team is going to deliver or a technology team, right? And there's a large portion of this that we're going to, in a controlled way, say, hey, marketing, like, you can run this because you're going to make changes all the

time. And I sort of think that next iteration in the data universe means to happen, which is I'm not ever proposing that an Ultrix user goes in and sets up the cloud data platform and determines all the data governance rules. But I also don't think it scales to say every time you need some data, file a ticket and get in line, and we'll figure out when we'll run that for you. And yet there's still a lot of organizations that operate that way. And I think what happens up happening is people go, well, I'm not going to make that mistake again. I'm just going to start downloading spreadsheets again directly from my business applications because I can't wait in line to use the cloud data platform.

And then one day the IT team goes, why does nobody use the thing that we built? And it's like, well, we made it hard to get to. Yeah. Well, and I think there's a real, the good news, the progress I think we have made is there are more companies that have data that's in one place. There's companies like, there's several of them out there, but there's a lot of these ETL type companies, ETL's gotten easier. You've got Databricks, Snowflake, companies like that.

The storage warehouse thing is in the cloud now. There's less firewall rules to deal with, like things of like legacy. Still some, but some of these legacy things that were barriers before. So I think there's a lot of progress there. And of course, like there's reverse ETL tools as well. But I think there's a lot more on that.

Like, all right, we have a plan to land the data and store it. And like, and we have this platform concept. But there's a lot more progress to be. be made, I think, after that. Because the default answer, at least for companies I talk to, is, well, we hooked up our Power BI tablo. We hooked up our BI tool to it. Okay, great. But I think there's a huge, like, gap still of, one, let's get data in tools people already use. So product team, sales team, ops team, you know, there's that gap. But then, you know, we've already talked about this a little bit. Then there's the big gap of like, well, like, what else can we do? Like,

AI agent stuff, the, you know, or other just workflows that sometimes their agentics, sometimes they're deterministic workflows with the data, both of which are really valuable. I agree. I think you have to empower people to work with iterating solve problems with the data, not just look at dashboards that come out of the system. I think that's absolutely right. And I also think there is, you're right. I also completely agree.

It's gotten a lot easier to create a large central platform and fire hose all your data into it, which is great, people should do that. But then you have to turn around to your business analyst community and say, hey, great news for you. Forget why it's good for IT that you've got all this under control. Great news for you. We, for you, have put all this data into one great centralized location that you can now use to solve problems.

And the question I think we should be asking ourselves in the IT side of the house is, have I made it easier for that user to come to the cloud data warehouse or the data lake or wherever to get this, then to download a number, NetSuite report themselves in Excel and work on their desktop. And I don't think most teams are setting that as the bar. And then they have surprised when, hey, my users just keep downloading the, you know, the report out of the business app. And if we change that mindset, I think the next thing that happens that's really interesting

is those companies are building really incredible capabilities in those data platforms. And so we've been talking to these partners of ours to say, hey, when you build this amazing stuff, when you build new, you know, Gemini, capabilities inside BigQuery or you rolling out new AI models in Snowflake, how's the finance team consuming those? If they're only consuming it through tickets to the IT team, they're not going to use it. They're not even going to know what's there. And so we've actually been looking at ways in our product to actually expose features that those vendors are shipping in our product. And so imagine you're on the accounting team and maybe I do account reconciliation every month

in Ultricks. But now I'm doing it out of this cloud data platform that has these really cool new features, imagine me being able to now use those. I would say, hey, I'm really glad the IT team made this big investment in, you know, pick your favorite cloud data platform, because again, now I get value as a user. I think that potential is there. I think it's really something to be unlocked in a lot of organizations. Yeah, it's funny. We just wrapped up a data platform project. And like, this makes a ton of sense. But the question was like, well, how do we get to the data now? And like, in the expectation, this is funny like this particular data platform was like well can i like log in and like right click

and like download the data from the table and the answer is like if you don't sequel no sequel actually no not actually you can't yeah so there's just some funny things like there where like you get it's all organized and it's modeled and you do all this like great stuff and then like of course people expect you to hook up a bi tool but like in this particular case they're between B.I. Tools looking for a different one. They're like, can I just download it? Can I just use it? And that's just a simple example, but I think there's a lot of those that are also more complex where it's just a practical, like, did we make it easy to get to the data? And the answer's no, a lot of the time. And only the amazing new features. I mean, I completely agree

there was one of the partners we were talking to, and I don't know if this is a public use case they're sharing it. So I want to be slightly guarded and how I talk about this. But they were giving us this example of imagine loading a lot of large unstructured content into their data lake. So put, you know, whatever, 100 million bank statements in there. And now you can query those unstructured documents like they're structured and do things like, say, hey, give me all of the balanced totals at only these two banks only in May. And it will go essentially use AI to parse those documents, figure out the totals and

all of those statements and provided to you. That's really cool. That'd be really valuable if you're on the accounting team. Now, do you think the accounting team is going to open up the console and rip out some Python code to go do that, no, right? Probably not, yeah. More importantly, do they even know that's something they could do? No way. Yeah, I'd say definitely

not. So that's like, imagine now, you know, they're in Ultricks. Like, imagine a palette pops up and like, here's the document totaler. You'd go, great, like I'll drag that into a workflow and save me all this time of going through these documents. And so I think there's both a feature exposure

problem and then a feature usage problem if you really think everybody in the business wants to go learn Python and SQL. I just don't think that's going to happen. Yeah. Can we back up just a little bit? We talked about citizen developers. You mentioned finance accounting a lot. Tell us

a little more like who are the kind of bread and butter users of of altruicks. And you've mentioned like some great use cases here. But what are like, you know, if you had to bucket use cases into like a few high level buckets, what would you say those are? Yeah. I think it's the people that are, I would say,

moderately too amazing at things like Excel and Google Sheets. are fantastic users. You can, in your own company, I'm sure people get to go who those folks are, but commonly, it is finance and accounting. It's also supply chain, revops, marketing ops, you know, a lot of operational roles will do that. A lot of functions now in most companies have some kind of ops team that maps to them as well.

And so, you know, that universe, but I also find more and more, I think working with data has become a core skill set for just a lot of business users. I don't love the term business users. It's a very nebulous kind of concept. But I think there are a lot of people who in their job are now perfectly comfortable opening up a Google sheet and doing some work and solving a problem. Right. And if you're able to do that, it's not a very steep learning curve to go to. Well, I could now do that in Ultrix.

And I could do it in Ultrix. Again, part of the value is the scope of the data you can work with. Part of the value is the data assets you can get to. But a big part of it is doing really powerful things in a simple interface and then automating it, right? I mean, how many of us have something where it's like, oh, yeah, I went in and I figured something out in Excel, and you're like, I'm going to have the same problem next month or next week. And so that's really our core user. Anybody who can work with data, understands data and wants to solve problems with data, that's really our core user. Yeah, yeah.

And tell us about, so the data cloud, you are investing a lot. in there, partnering with, you know, some of the giants. But this is a little bit newer. But obviously, I mean, it makes so much sense to give all the folks you just mentioned access to all of the data, that the IT team or data team has spent, you know, so much time and resources around, hey, let's aggregate. Let's centralize this data here. But this is like a newer thing for all tricks. So what was like the old way of getting data? And then now like you're on top of the data cloud. Can you just talk us through the evolution there? Yeah, I think we're solving the same problem again, and I think that's pretty normal. I think a lot of things

in technology can be explained as a pendulum. I think when we first started helping the user solve problems with data, they were really doing it at the desktop level. Again, everybody would get sent reports, sent files, sent, you know, Excel spreadsheets, and they were manually having to pull all this together, and Altrix became a way to automate that. And what we find with a lot of our customers, they go, hey, that was working really well. And then one day the IT team came along and said, hey, great news. Instead of having to go all these bespoke applications and get all these reports, everything's now in our data lake. And so turn off all those connections to those systems. And while that's a good data strategy,

it's actually a really good data strategy, it sort of recreates that same problem again if you're not giving that user access to get to that data in a way that makes sense. And so I sort of think the first version of Ultrix was connecting people with data and letting them solve problems. Now the data's moved. How do we? we just go solve that problem again in a way that makes sense with the cloud data platform. So that doesn't necessarily mean sucking all that data back out onto the desktop, to be clear, like we're not trying to, you know, hey, I now connect to Snowflake and just pull a bunch of data out of Snowflake and Ultricks doesn't make a lot of sense. What I want to do is build a workflow

and, you know, push it down into Snowflake or maybe I need to use, you know, Snowflake, plus I've got another division that uses Databricks, plus I've got a division that you got on the cloud data platform. So how do I, as a business analyst, pull all those things together? So we're, again, solving the same problem, which is not uncommon, but really helping democratize the ability to solve problems with that data asset, but now doing it on top of these cloud data platforms and importantly doing it in a way that leverages the strengths, the capabilities of those platforms versus simply trading them as a generic data source. It's a lot different to connect to one of these amazing platforms than to connect to a spreadsheet sitting on a file system. So we treat that a lot differently. right now one of the things that I think that you've already mentioned once is like thinking back so you've got this evolution of like right everything's on the desktop you know I'm connecting files

and stuff and now you're connecting to these data platforms more like if I guess I'd be curious like what's been your experience with people as far as access because there is a sense where the citizen developer the analyst like probably doesn't have like the same degree of control anymore. Is that kind of, is that a barrier? You think? It certainly is. It certainly is. I think this is where, you know, this is not about going around IT or being shadow IT. I don't think a good answer is, you know, use Ultrix so you don't have to use the cloud data platform because it has good governance and you're trying to avoid that. Like, I don't think that's a really good strategy. I think it's, how do you actually connect to the governed data in a way that's using the credentials

and the access that makes sense. Like your RebOps team should have access to your sales and marketing data. They're going to go either directly in a Salesforce and Marquetto and get it and download it, or you can make sure they have access into the replicated data

sitting in your cloud data platform and do that in a governed and well-martialed way. And I think that, to me, is really important. I don't think we're trying to have people go around those models. I think we're trying to say, if you want those models

to be the governing factor for your data, you have to give people a way to get to that data and work with it. Otherwise, they're going to go around you. Like, I mean, this is the thing. It's sort of like, you know, I can build a perfect walled garden that has, you know, amazing data governance and amazing security. Just nobody's allowed in the walled garden.

Yeah. Doesn't really work. And I don't think that's what anybody I teach trying to do. Right. I just think when we get that dial wrong, you know, users are like water. They'll find all the cracks in the system to get their job done, right? And so, you know, you've got to make your strategic data platform the best way to solve problems, not the hardest way to solve problems. And I think that's a great

goal for teams that. We see teams that have that goal. I think that's a great goal. And if you do that, then I think you get good governance that works for users as well. Yeah. Well, one of my beliefs about software has always been the best software typically sits in gaps between teams and helps teams work better together. Right. So like your gap is like part of his traditional IT and then these like technical business users. Like that's, I think, all tricks is, you know, historically, and I think continues to be in that gap. And that just makes for some of the best software

because you're actually facilitating the business running better, right? And you're facilitating communication in some sense, you know, between the teams. And that rapid iteration of change, right? And I think, you know, we're going through another period. I've been through a couple of these in my

career now. And we're going through another period where I don't think anybody should be getting on a podcast and saying, I know exactly what the world looks like in 18 to 24 months because they don't, right? And no business user knows that either and no IT team knows that either. And so what are we all doing to sort of facilitate a reasonable way for the business to iterate and move quickly that involve IT, that involves that governance because, again, otherwise, they're going to do it without you, right?

And then you're going to go, oh, my gosh, like they're running a custom GPT they built that downloads data from our business applications directly and is sitting on somebody's laptop and runs a batch process every day? Like, I mean, you know, it will happen unless you go, oh, actually, here's a way to build an agent using our data, you know? And so I think that acknowledgement's important during these times of change in particular. Well, I think there's two components. One, I want to drill into something that you said a few minutes ago, but the other component

is we absolutely, in some ways, are going to move back to the early days of Altrux where a lot of people will have local things on their computer. Excuse me, they're downloading, you know, from the, from some system, from the data platform and then doing things on it'll run local on their computer i mean that will happen that is happening yes yes yeah and then the other it's the same one it's the same one i tooled right like either you find a well-governed large language model that you can manage the security around or reality check your users are pasting stuff into whatever tool they're using on the internet like you know exactly these are your choices pick a smart one and sort of empower people

to get their jobs done right yeah yeah but you know, for a really good IT team out there and there's a lot of them out there that wants to, you know, go down the road of enablement and do a good job with that. I think there's this interesting, like, and you've already mentioned this, position for companies like Altrix to do two things. One, enable discovery of like, oh, hey, you know, Gemini has this feature, Snowflake has this feature, Databricks has this feature, and then expose that so users can take advantage of it. And the second, which is just as important, is the same with these data platform teams of like, oh, hey, your team just put out a new model that, you know, models the attribution for sales or marketing

or that, you know, they've created a customer 360 model and they've added a new column to it. There's a lot out there where there's platform features and there's, I mean, most technical teams that I work with struggle to advertise their work, right? Like they're doing this work and they do it for a user and they fill out the ticket and the one user is like, oh, great, thanks for adding that column. But if you had a tool like, you know, in Alltricks, it's like, oh, hey, like literally, you know, any user from any group can now see like, oh, look at this like new, you know, customer lifetime value column. And like, then 10 people might see it and use it. And I think that's a super valuable spot, you know, for you guys to be on too. Yeah, that's right. That's right. And I think there's an element of enabling the business to reimagine what they're doing with AI as well that I think IT teams, I agree with you. I think a lot more. teams these days are on that empowerment strategy. By all means, don't mean to imply this is the standard IT of 10 years ago. I think that has changed.

But I think AI in particular is an area where, you know, I joke all the time, there's these two competing mandates from corporate. There's the CEO who is on stage telling everybody we're going to be an AI first company. And then there's the email. Everybody gets shortly afterwards from legal and complying saying none of our data is allowed to go in any AI tool of anything. Yeah. And so, yeah, totally.

And so, again, back to this, like, how do you, what is actually happening then? Well, what's happening to that is other things are getting your data. A better answer is, you know, imagine turning to your users and saying, hey, we have a well-governed set of data that you can build workflows around that might drive agentic AI. We have an approval process for how those governed models can get approval to go into a specific AI capability. So, again, we're not saying everybody gets to do whatever they want with AI. We're saying, you have the ability. as a user to put together a data pipeline or an agent, and then you have a process to get that approved.

And I think that is simple yet missing in most companies I talk to. There is no vehicle for the company to say, we welcome the innovation in an innovation lab kind of mentality. And then we have a pathway for that to go into a production use case, you know, something that's observable and can be changeable, and includes those key declarative capabilities. That's the other thing we're seeing a lot of is people that are going pure AI first are saying, well, I don't want to ask chat GPT, do you think this is GDPR compliant?

Like, I want to run my GDPR rules, right? And so is there a way to empower these kind of analysts to say, I know how to do that? I know how to grab the right data, do a GDPR check, but I also want to use an LLM to maybe write an email, right? And then I want to be able to go to the corporate types and say, hey, can you approve me doing this? Can I use these six fields from our sales and marketing system? And here's where I do the GDPR check. And here's where I manage PII. And here's how the email gets written.

And here's the audit trail. And give people an avenue to say yes. And I think that's, to me, the next big unlock. I think we're going to see that in the next six months as people running to like, how do I set up something like that? We've been calling that our AI data clearinghouse. But some kind of model where the company can say, we're not going to go, nothing's going to be AI driven. And we're also not going to go to like, we've bought a large language model.

We let it crawl the data lake and we're hoping for the best. Like that's not happening either, right? So what's the middle ground? Yeah. Yeah. That makes a lot of sense to me. And I think that middle ground is going to be the case for a lot of companies. We're going to take a quick break from the episode to talk about our sponsor, Rutterstack.

Now, I could say a bunch of nice things. if I found a fancy new tool. But John has been implementing rudder stack for over half a decade. John, you work with customer event data every day and you know how hard it can be to make sure that data is clean and then to stream it everywhere it needs to go. Yeah, Eric, as you know, customer data can get messy. And if you've ever seen a tag manager, you know how messy it can get. So rudder stack has really been one of my team's secret weapons. We can collect and standardized data from anywhere, web, mobile, even server side, and then send it to our downstream tools. Now, rumor has it that you have implemented the longest running production instance of

Rudder Stack at six years and going. Yes, I can confirm that. And one of the reasons we picked Rudder Stack was that it does not store the data and we can live stream data to our downstream tools. One of the things about the implementation that has been so common over all the years and with so many rudder stack customers is that it wasn't a wholesale replacement of your stack. It fit right into your existing tool set. Yeah, and even with technical tools, Eric, things like Kafka or PubSub, but you don't have to have all that complicated customer data infrastructure. Well, if you need to stream clean customer data to your entire stack,

including your data infrastructure tools, head over to rudderstack.com to learn more. I'm curious, and you can maybe reach back to your time of user testing, or maybe even, like, so far at Ultricks, what were some of your aha moments with these AI-enabled workflows that maybe you yourself or people on your team were creating at user testing or you guys are doing at Ultricks? Like, are there like one or two that stick out where you're like, wow, like that's really cool?

Yeah, I think generally speaking, what I see happening today is when companies are experimenting with AI, and we did exactly this, you know, I mentioned that we were doing a bunch of this in my last company. You know, you license something like a check. at GPT or a Claude or a Gemini for everybody, and you get sort of those personal productivity use cases, you know,

helps me rewrite emails and stuff. That's great. And then you get a lot of, what I'd say, sort of like creative derivative works. You know, I load a bunch of marketing documents in and I can get a marketing document back.

And that's valuable too. What I can't do in most of these systems is ask it detailed questions about how my business operates. You know, I can't ask it to cut territories or, you know, give me an extra iteration of the marketing plan or what will my commission be if I close this deal or, you know, what's the close rate then on open bugs in my data platform?

It just doesn't have that data. And I think there's an opportunity to build those pipelines and build those models so that it can sort of answer those questions. But again, I don't think that's going to be a flip-the-switch pointed at the database. And so I think one of the big opportunities is to figure out how do we help companies and business users, number one, sort of bring their data to AI, to build that set of canonical data. that they can go into an area where you can learn. I think the second, though, is bringing AI into your data flows. And so, you know, for example, a lot of our customers use Ultricks to do, you know,

end-of-month reconciliation, right? Budget versus actuals, for example. Why is the last step of that not? And then I have a trained LLM that looks at this month versus the last four months and says, hey, in a rolling five-month basis, what were the anomalies this month? Is there anything showing up in our budgets that we should be aware of? What was the impact on tariffs in this month's budget? That should be an agent that closes the books.

And that agent is really that same Ultricks workflow you've had with AI applied to it. So I think that's what I'm really excited about is this idea that I think world-class companies are going to build their next generation of agents on top of their strategic cloud data platforms using a lot. logic layer like Ultrix and using large language models, what I don't think they're going to do is build that on top of all of the business logic of the last 15 years and all their application stacks. And so I think that to me is really compelling. It was one of my big lessons when I was building stuff in chat GPT a while ago was I don't want to call six APIs of my old business applications, pay to use all those APIs, wait for it

to run all that old logic. What I really want to do is go into my data lake and say, hey, I understand my. sales data or my marketing data or my finance data, you know, grab these eight fields, pull this together, do a workflow and then do cool AI stuff with it, right? I think that to me is the future. Yeah, I mean, it's funny. It's almost like the last cycle, at least in the data space, got cut off because we got to Cloud Data Warehouse, we got like some adoption with a lot of companies with that. And then there was a lot of buzz around like, okay, we're going to build apps around the data platform. And there's going to be all these apps that come up

and all these startups that are cloud warehouse native. And, you know, we're going to see the, see just every category, CRM, ERP, with the idea that like, why would you not want to essentially onboard an app and give it access to your data versus like storing all your data in the proprietary system that you're going to move off of in four or five years and then again, four or five years. And, you know, and then perpetuating the cycle with a lot of, you know, a lot of companies switching between providers. And then that kind of got to, then it got cut off with AI, partially funding, right?

Like a lot of the funding gets sucked out of that realm and dumped it to AI. And then partially, like, it's just the, you know, the nature of something that's more revolutionary. But I like what you're saying is I think that pattern can still continue, but it just looks a little bit different because of the AI use case. But I think we can still head in that direction. It just looks different with the, you know, AI agents. I completely agree. I think it could be an accelerant.

And I think this is also common in text. where the first iteration isn't what ends up being the end that we go to. I don't want to simply replicate the same kind of declarative logic that I used to buy from my large multi-tenant SaaS vendor on top of my cloud data platform. So I think two things are true. I very much agree with what you're saying. I think one is almost like an inverse strategy. I want to have a data-driven, agent-driven strategy for how I solve problems in my business,

not how I rebuild my business logic. Yeah. And I think that will be a very different thing. I will have, I think I'll have agents that do territory carving and deal management and deal pricing and forecasting and all these things. And that's going to work very differently than how I buy business application software today. I think the second thing is the cost of writing software is going down dramatically. And so again, I think the idea of buying big packaged monolithic software versus like, I can just solve this problem myself. I have this big cloud data platform. And to your point earlier, I've already pulled most to my data into this now. Like, I grab this centralized, smart data platform. Why don't I start building whatever combination of data-driven agents and some lightweight applications? And I'll just own, you don't want to have other vendors own all my data. You know, there's a big tug and tug of war going on even right now in the press about, you know, who gets to charge me for access to my own data that's in these applications. I think pretty quickly companies are going to realize, like, I have all this data in my cloud data platform. Yeah. Like, I'll just go

there and I'm going to, I'll use it and I'll solve problems in a new sort of AI forward way. So I completely agree with you. And I think this will be an accelerant for people who lean into that strategy. And I think they'll do it not by buying a new set of packaged apps on top of their cloud data platforms. I think they'll roll out the ability for the business to solve problems with that data using AI and create a bunch of agents. And I think that'll be the next wave of transformation we see inside companies. Yeah. So we tell us where alterics fits in there because I know you have a good answer and strong opinion. And I also, I mean, go back to something you said, I can't remember in the show or maybe when we were talking before the

show is like, you know, you believe the people that are actually going to help like define and build these agents are the citizen developers. So, so tell us how Autrix fits in. Yeah, I don't even know that these folks think of themselves as citizen developers. Yeah, I think it's fair. I mean, I get what you're saying. I think there are really smart people all over your business. that understand your business. They understand merchandising returns or closing the books every quarter or the marketing workflows that you have.

And I think those people are going to say, oh, like if you gave me access to the data that I need without me writing much of Python or SQL, but I can just work with the data. I understand that data. And you gave me access to agentic AI capabilities. I could change the way we do, merchandising returns or closing the books every quarter.

And so I think empowering those folks to do that iterative innovation, right, and to really work through how that can change. And, you know, again, having a control process in place where they can publish those agents and do that kind of stuff. I think that's the real transformation. I don't think it's going to be the job, frankly, of everyone in IT to sit down and like reimagine merchandising.

Like, I think that's a really tall ask. Now, instead, to say, hey, imagine agentic AI and what it could do, how do you empower the business now that you've built this cloud data platform and you understand how AI can work, and how do you partner with them to reimagine business processes? I think world-class IT teams are going to do that really well. But that, to me, I think, is the value prop. And again, it goes back to this is what we've been doing for 20 years,

is going to business users, these analysts and saying, hey, like, you know how to do rev-ops. Like, let's give you tooling to get you access to all the data to solve problems. I think the new way of solving problems is going to be right and automate to play an agent, right? Or write a, you know, write a workflow that, you know, put, some code into a vector database. There's some data in a vector database that lets the agent, you know,

answer this question for your field or whatever. I think that's where this goes. We are getting close to the buzzer here. John, do you have any final burning questions? Yeah, I think well, I mean, as far as as, you know, altrux goes, I think it'd be

awesome to tell people that maybe are interested in checking it out, like what the right, you know, what the right path would be for that. Because I think we have a number of business users and more technical business users are like, how do I, what do I do? How do I check this thing out? Yeah, well, I mean, first of all, I mean, we're a large company with a very large footprint. So number one, you might find out already that you're allowed to use all this inside your company.

So like that's a great point. That might be one. The second is like every company all the way back to the beginning on websites. Like our website's a great place to go learn about the product and connect with our team and sort of see what we're doing there. We recently had our big user conference, which was inspire. We put a bunch of that content online where we were really talking about, you know, how do you leverage your cloud data platform? How do you build an AI did ex-airinghouse? I think those areas are probably where I'd send folks, but most importantly, you'd reach out. We're happy to come in and talk to people. Awesome. Last question for you. Any before we wrap up here, you middle of your career got into product management.

I clearly very passionate about that. What's one of maybe the kind of key learnings you've had from like building products to CEO work that you still like as a CEO is just like, you know, something that's just still really valuable to you today. I think two that really stand out. Number one, I think good product managers are all about solving problems,

not building features. And I think that actually translates to this conversation today. If data teams are thinking about how to solve problems for their business users, I think really good things happen. If they're just building cool features, stuff sits on the shelf. So I think that's number one. I think the second is what I love about product management is it's this job where you're sort of responsible for everything and you're in

charge of absolutely nothing. Like you don't run engineering, you don't run sales, you don't run marketing, but if the product's terrible or the product doesn't sell, it's your fault. And I think there's a lot you learn about how to communicate to people and how to get aligned on what we're doing, why we're doing it, why everybody should care. And again, I think that's a great analogy for getting anything done in business. And I think I've been thinking a lot about what does that look like in this transformation to AI? Like how do IT, the business, compliance, legal, like, how does everybody get aligned on like, here's where we all need to go together? Because there's a lot of people around the table that could all sort of say no or do it my way or I'll do it myself. And

when you're a product manager that happens, sort of nothing happens. If you're a product manager and you sort of figure out like, how do I get everybody aligned and going in a direction? Because again, I can't just tell them to. They have to want to do the thing that I'm getting them to do, really amazing things happen. And so I've tried to take that mentality even into my CEO jobs of like, even though now I am in charge and tell everybody what to do, it's a lot better if everybody wants to go in that direction. And I think, for sure. I think having that mindset for this next transformation to AI inside companies, I think is a really powerful framing of like, hey, are we all sort of going in the same direction because we want to? Yeah. Yeah, so that's a great

word. Well, Annie, thank you so much for joining us on the show today, sharing your insight. It's been a fantastic conversation. And yeah, we hope to reconnect soon. Absolutely. I really enjoyed chatting with you guys. I hope the list of all enjoy it. And thanks for having me on. The Datastack show is brought to you by Rudderstack. Learn more at rudderstack.com.