Buy, Build & AI: Your New Software Strategy for Energy Leaders

Energy leaders are facing a new reality. The classic buy versus build software debate is no longer enough. These days, it's all about harnessing the power of build, buy, plus AI to future proof your business. That's what we're gonna talk about today. So let's start the countdown. From LinkedIn as well as YouTube. The CG hour is a presentation and discussion with industry experts as well as leaders around hot topics in business as well as technology. So if you are in the live audience, please introduce yourself in the comments. Let us know where you're tuning in from and introduce yourself, build your network, and connect with others in the comments. So today's discussion is all around Build, Buy, and AI, the new software strategy for energy leaders. And we are going to introduce you to our expert panel here. Over to my left, we have Joel Wolf. He is the founder of Wolf Global and CIO and former VP of Technology at Medallion Midstream. Welcome Joel. How are I'm doing great. It's great to be with you guys. Awesome. Please let our audience know a bit about your career Joel. So for the last twenty plus years, I've been senior most leader in IT for multiple organizations in the oil and gas industry. The past few years, I've mainly focused in the midstream sector where I've overseen information and operational technology as well as cybersecurity. And then currently, as through Wolfe Global, I'm helping C suite executives with their strategy around IT and business in general, as well as CIO and CISO services. Wonderful. We're going to get a great business perspective from your side, Joel. Thank you. And over here on my right, I have Nate Richards, CEO of NRx. Nate, welcome to the show. Thanks. Great to be here. Can you please share a little bit about your career with our audience? Yeah, I started a software consulting company in two thousand and six that helped oil and gas companies primarily in a similar spectrum of the energy space that Joel has been in from the service provider side, building, buying software solutions, doing integration. And then later in my career, now I run a product company. So I've kind of moved to the product side of the software world, providing a sales enablement platform for the retail energy space. Awesome. Welcome. And last but not least, CG Infinity's very own Tanya Shepherd. Welcome Tanya. Thank you, Fanny. Thanks for having us. Absolutely. Tanya is the SVP at CG Infinity. Can you share a bit more about your career, Tanya? Yeah. So Nate and I are reverse in our career path. I spent fifteen years running a software company, that was built for the energy sector and really went through many sales cycles where companies were going, do we build, do we buy and having to debate that with them. And one of the things that I missed as being a product company was being able to not only help a client implement the software, but truly operationalize it, help them redo their business processes around it, get it truly adopted and get it integrated into the whole ecosystem because that's what makes software successful. So in the second half of my career, I now am part of a consulting company where our practice does just exactly that. We help them select software, but we also help them integrate and optimize their operations to take full advantage of the software. Awesome. So as you guys can see, we're going to have a broad perspective across this industry panel here. I want to welcome some of the folks that are joining us live. I see here Tracy Morgan from California. Welcome Justin Wilson as well in the comments. Feel free to drop any questions and comments throughout this discussion and I'll be sure to bring it up to our guest. So before we get started, we wanted to play for you kind of like an introductory introductory video to kind of set the context for this build, buy, and AI debate and discussion. So let's roll that right now. When you build, you're choosing to become, to sort of insert into your P and L, a single customer software company. To me, the framework is mission critical versus non mission critical. Okay. Right. Does it differentiate you in the market to have your own IP on a thing, whatever that thing is? Yeah. Is your market so complex that the only way is to really build it yourself? That's sometimes that happens. We see that in gas, complex industrial gas all the time. Sure. Or yes, you might know the expertise, but your time is actually better spent even though it's mission critical. Maybe it's not market differentiating and maybe the best thing is to go to an NRx or you know one of your competitors and say what do you got yeah because I don't want to devote all my enterprise resources to building this. I thought I would kind of share that's that's my lens I thought I would open it up and see if you disagree I think those are great. I would add maybe two things. One is time to market. You know, if you're building, you're starting with a blank sheet of paper in the coding world, you're like, in the corporate world, you're starting with HR recruiting or, you know, assessing consulting partners, you know, doing a deal with CG infinity to kind of how's this going to get staffed? What's my timeline budget, etcetera. And so you look at, okay, when would I even get to MVP in a real user's hands doing something good for my business versus a buy option, you know, generally is going be a faster time to market. And I'm sure there are implementations we can all look at that nightmares wound on for years, never got live. It's not always sometimes a vendor's fault, sometimes a customer's customer's fault. So right. Yeah. Not pointing any fingers there Yeah. So shared responsibility. And I think the other thing I would look at is total cost of ownership, which is when you build, you're choosing to become to sort of insert into your P and L a single customer software company. So that is R and D, that is OpEx support and move, add, change, bug fix. That is every next version, every feature you want in the future. And the technical debts the need to upgrade and maintain platform, right modernity, right? So when dot net whatever comes out or Python whatever comes out or there's some hole is found in the current version of blah, what you're going to, what's it going to take to move to the new thing, which adds no value from a feature perspective, but as a vendor, like that's just our problem. Like no customers are like, oh, here's some more money to upgrade the platform. Like, no. If you're just dropping in, you're watching the CG Hour. Today's topic is all around build, buy and AI, your software strategy for energy leaders. And thank you for those joining us in the feed. See Galermo from Bedford, Texas. I also see Todd Cackley, CIO at Textron here from the DFW area. So welcome everybody. I want to start the discussion here with Joel. Joel, as you kind of saw from that video and across your last five years basically in midstream, upstream, all that. How has the software decision making process changed over the last few years? Yeah. You know, and everyone knows what happened five years ago. And, you know, really the the amount of innovative new technology that's come out, the speed that it's come out and the maturity that it's come out the last few years has accelerated enormously, right? And so I don't know that it's really changed, particularly if you had a cloud first methodology, which many of us have had for over a decade. But the amount of mature solutions that are available to you to not only help you make that decision of buildbuy, but that are meeting many of the requirements of business tasks and solutions that they need. Yeah, and then throw in AI now. You've got multiple tools. AI is the cornerstone of everything. So yeah. Yeah. Nate, from your perspective, this build versus buy used to be versus, right? But how has that changed from your perspective? You know, think just came off of Joel's point. If you rewind the clock twenty years ago, you had monolithic solutions that sort of were everything where or build it from scratch. It was very bipolar type of option. So you had no what we've gone from is this monolithic market to a more modular market. So you can say, hey, I want to take these modules from vendor A and I want to integrate them with these modules. So there's sort of a best of breed option that wasn't really out there as you get more of a diversity of service providers and also just better integration. So solutions used to be very closed. So you'd buy a large monolithic application and the data would go in and it doesn't ever come out. And now this idea that my data would sort of be held hostage by a vendor is just not a nonstarter. And so software solutions tend to be more open, APIs, connectivity, sometimes built in integrations already in the industry. And so that those variables lead you to a much broader spectrum of options today as you look at build versus buy. Yeah, and let us know in the comments too from your perspective in the audience. Are you team build, team buy or team hybrid? Where are you at right now? Tanya, from your perspective, you've been on the business side. You're now on the consulting side. What do businesses kind of get wrong or underestimate when it comes to this kind of build by decision making? I don't know if it's the business or the organization overall, right? You can have an organization full of very smart people that know the business that you run today. But you by not by building it yourself and not going out to a software vendor, you lose the opportunity to get different perspectives, to understand maybe how different markets or different companies accomplish the same goal, right, and the same business but in a different way. And so what I see a lot of times when companies do go the build route, and there are still some out there that do that, they're like, I don't like buying because it's a black box. They then have to come back and revisit what they've built when they expand into a new market or put out a new product or go after a different customer segment. And so that rework takes time, money, effort. But if you find the best of breed solution where you can have that flexibility that you can pivot as a business with more integration, more configuration versus having to sit down and write the business cases that are in Susie's head. Right? You get more value overall, I think. Joel, do you want to build on that? Have has there been any situations where you've bought the software and then it introduced more complexities or you've built it and then there's more complexities down the road? Yes, I think we've all gone through that, particularly early in my career, you know, back to your monolithic statement, it's, you know, you dealt with a lot of legacy systems that that were just unable to integrate. And that's what it's all about today, right? So I think it's less of a risk today for most organizations, particularly if they're choosing, you know, cloud native solutions. So yeah. Nate, as you kind of work with different clients, have you seen kind of that regret of one's decision versus another and how do you even fix that? Yeah, there's maybe two quick anecdotes that come to mind. One is I had a customer that had built their own CRM. And at the time, you know, maybe they started five or six years ago. They thought, hey, they had a, they hired this great guy who knows the industry, similar to the, what Tanya's described, knows the industry, has grown up in the industry, knows how to code. And they're like, this guy is going to build us the best system ever. And guess what? That guy built them a system that worked great and they loved it until that guy left. And so this is kind of another risk, which is, you know, because you're not a software company, your job is not hiring and retaining software developers and maintaining that IP. And so they kind of got stuck on this island where they try to hire other people to work on it. The way some of the tools the guy used, you know, they didn't know. And so also just think the ability to manage someone when you're not in the software business who's building software, don't really know what they're using and I guess they told us it worked great in it. The buttons clicked. So kind of getting stuck on this island and the ability to support that over the long haul with the changes in the business is kind of one kind of dead end type scenario. Another one is thinking that, well, went with this big name vendor that's like a well known platform and we got a great integrator who told us they could implement it for us in six months and do all of this stuff that we're currently doing with this spreadsheet and that thing, which sounds really great, but it kind of can turn into another build option as well because you find out that you're really developing software just within a platform. And so what you thought was a buy option can kind of start to look a lot more like a build and DIY option over the long haul and has some of those worst of both worlds. You're committed to the platform, you've got a ton of IP **** ** in it, and yet it limits your options to kind of come in with other solutions down the road. Great examples. Thank you. Tanya, I mean, obviously budgeting comes into both these decisions or all decisions, right? Are there certain decision making factors that you use to guide clients on navigating that huge upfront cost so that they have a long term value? Yeah, budgets always a question, right? And so do you and people look at it going, well, I've got to pay this software vendor this a big amount of money versus I can have staff. Well, promise you that staff comes with a big dollar as well. It's just spread out differently and you, you know, you handle it differently from accounting perspective. But the other part of it is, is Nate's right, the software you choose, the partners that you choose, are critical to the success of where you're at. And so you can buy great software, but select a crappy partner. And it's no different than buying crappy software and selecting a great partner. And sometimes, right, you get a better result with crappy software and a great partner because they can help you work around all the issues. So it's it's not one or the other. It's a combination. And you want, as part of your decision making, not only look at what problem is this that I'm trying to solve with this decision, but how does that then interact with the rest of my ecosystem? Because if you get something that doesn't play well with others, then you're stuck. You're you're still on a different island, right? It's still an island though. Yeah, that's actually a very good point and to Nate's point as well. We'll talk about platforms a lot today, but there's different types of platforms. And if you silo yourself in the wrong platform, you're just going to have more difficulty down the road. So we talked about open source and APIs, and there are many platforms out there today that are providing that for you. And really, it's you got to think beyond the use case that you're dealing with today. You got to think across the organization and choose correctly. Think also just the relevance of that partner's knowledge to your industry versus just knowing the tech side. So some, not naming name, big name platforms have very well known partners that implement those platforms and have demonstrated success across a whole range of industries. But when you connect with them, how well do they really know your area, right? Energy is this massive spectrum, right? From hydrocarbons all the way down to the electricity meter and all sorts of side routes, LNG, gas, right, all sorts of things along the way. And each of them is a career, a lifetime worth of knowledge. So I think just saying, we know energy CRM, let's say for example, like a platform and a term doesn't really mean that they're ready to help you be successful with what you're trying to accomplish. So I think getting into that more detailed Being honest about your organization too, are you as an organization able to truly articulate your business needs, your scenarios? Or are we is there somebody got to come pull it out of you? And if we've got to come pull it out of you, then the person pulling it out of you better know the business because when you start using terms, they're gonna go, well, did you mean this or did you mean that? Because I can interpret that requirement in multiple ways and I want to understand which one you mean, right? And if you don't start with solid requirements, and I always preach that while we're all speaking the same language, we're not. Right? A single term can have multiple meanings to multiple people. And if you don't explicitly get that down right first, I don't care who you are, you're going to you're not going to be happy at the end. Yeah. Joel, any learning lessons actually from having to select multiple vendors and things like that over the years? Well, you know, so, spending most of my career in the mid market space, you know, we're very agile and, you know, tend to look at technologies that are both best in breed, but also potentially best in breed. And, you know, but choosing in a way that you're lowering your risk somewhat from a budgetary standpoint and an adoption standpoint. So again, speed to market is arguably number one or number two when we're talking about these things. So having the willingness to fail fast is important. And again, the technologies today allow you to do that a lot more easily. Well, Nate, you brought us a video with some timely statistics, right? Can you set up the video before we play it? Yeah. As recently as last week or two, there was a study that's been ongoing about organizations implementing GenAI pilots in their enterprise and what type of outcome data are we starting to see. We're seeing kind of almost into the year mark of people launching these pilots and saying, hey, what can we really deliver to the business? And some really interesting takeaways, one is that most of them are not producing the type of ROI that were anticipated. You're even hearing this in AI investor circles like, hey, is this overhype? Are we on the hype cycle here and need to start to be cautious? But the three positive notes that came out, I think are probably the most salient. One is that vertical SaaS, so software that's aligned to your industry as opposed to kind of the horizontal platform offering is producing a better outcome in terms of ROI. Second is that purchase solutions are getting practical applications better than DIY, which is maybe a little germane to our discussion today. And then maybe a little self serving for me, but ask the MIT guys. And then the third thing I think is about the human element. I think it's maybe the best takeaway, which is that direct connection to the line of business managers versus central AI lab. So if you think about two ways you could go about implementing AI, right? You could say, hey, we're going to assemble this team of experts and they're going to go into the lab and we're going to give them all our requirements. We want you to fix these problems and they're going to come out with all these great solutions. That's the central AI lab versus, hey, we're going to park a couple of experts in these key departments. They're going to sit with the line of business. They're going to observe, learn and then quickly iterate, right, close to the business, close to the user. And that's driving much higher adoption rates and better outcomes. So I thought it was three really interesting points to where we are today. Thanks for setting that up. Well, let's roll that video right now. Did you know that a study from MIT reports that most AI pilots are failing to show value? The key takeaways were, one, purchased vertical solutions succeed more often than internal builds. Two, the biggest ROI came from automating back office work and three, Adoption was driven by direct connection to line managers. Enarex makes it easy to buy and sell energy in competitive markets. We provide a suite of software and data services for energy suppliers and brokers to drive sales, and help them effectively compete for and win customers. Enarex drives ROI through sales enablement and proprietary energy market data. Supplier Hub provides sales and sales ops workflow as well as commissions management for suppliers. BrokerHub delivers deal flow and commissions management in addition to digital sales for brokers. MarketHub is both API and portal based connectivity for energy transactions. Finally, DataHub is a market data portal for NRx's proprietary license compliance, pricing, and competitiveness ranking data. When you partner with Enarex, we deliver off the shelf AI solutions for energy sales, automating the back office work of sales support, and bring deep connection to brokers and suppliers. The result? NRx powers one out of eight commercial energy transactions in the US market. For a total of over seventy terawatt hours in deal flow per year. If you're looking to drive more sales, engage with brokers, or streamline your energy selling process, consider NRx as a trusted solution provider. Connect with us today. Visit NRx dot com for more information. As you can see, there's definitely more context there and I want to give a shout out to some of our audience members here. Todd Cackley, CIO at Textron, he mentions the ROI challenge is more on the I than the R nowadays. Most companies over invest in AI before understanding the value. So that's a key thing. Would you want to add up to that? Yeah, think well in most of what you're you're seeing today is is is generative AI use cases. But some folks are trying to build their own models as well. But I think the challenge comes is that if you use the same siloed methodology when you're taking a new technology like AI and your technology stack isn't future ready to consume that or to provide that integration, you're going to be challenged with it. Sure. Tonya, I want to come over to you because obviously before any of the AI talk, have to talk about data and data cleansing, right? Garbage in garbage out. I said that exact same statement to somebody yesterday. Garbage in and garbage Expand on that please. Yeah, so. I deal with data every day with our clients and and bad data causes bad results. And so, but it's we have so much data now that how do you isolate and figure out where the bad data is and what the correct data is? And when you have desperate systems that are aren't talking to each other, you know, we've got a contracting system over here and a billing system over there. Oh, by the way, there's a human in between that's trying to get data between the two. You're gonna have mistakes, and that's just the way it is. And so, you know, I wanna I'm I'm like, can we fix the in between first and that will clean up the data? But every day I'm like, I'm not allowed to look at data anymore because I always find problems. And then and then the project to go fix it. But it's it's hard to capture your arms around. And and sometimes they're so nuanced that you don't know it's bad data, but then people are starting to believe the results of the data, which to me is just as harmful, right? Or more so if you're going to make decisions off of bad data, which is, you know, AI can go in and round it all up and provide you with a story. But is that story correct? And and and unfortunately, a lot of times I see people just going, well, here's what it told me versus does it make sense? Can I do a gut check on it? Right? It's the same thing when you hear something crazy off of AI going, no, the sky's not purple, right? You can feed it bad data and it will give you bad results. And so knowing when to and how to validate the data that you're getting out is really important. One hundred percent. And Joel, I want to come back to you because there's tons of mergers and acquisitions happening in the energy sector, and you were involved in one yourself, right. So when that happens, you got a whole bunch of tech stack that is coming together, data coming together, any learning lessons that you can share with our audience around how to navigate all that? Yeah. I mean, just to piggyback off what Tanya was talking about, it really comes down to the data. And so you could have two energy companies that almost exactly the same business in the same sector, everything else, but they could run their business completely different. The way that they identify, categorize and name assets, for example, could be completely different. Way and here's another key point is that there are there's a good chance that there is a fair amount of data that is still being manually acquired and manipulated before it reaches a system of record. And so you have to fully understand that and plan accordingly. Secret Excel sheets. That's right. That's owned by one person. Oh, they're not so secret. Nate, can you, would you like to build on top of that what Tonya and Joel said? You know, I think a lot of organizations, especially with this AI thing and kind of going off our guest comment, which is that the problems with the eye, too much eye and then expecting this massive return, right? I think a little bit of, you know, it'd be like having a car, a car you've had for a while and your gym clothes are in the back and you're like, my car smells. I need to get a new car. So you go get some new high-tech car and you throw your gym clothes in the back and you're like, wow, this car smells too. Now I have proof that it's the car. The car is the problem, right? And so the gym clothes are your messy data, right? It doesn't matter. You can put it in a souped up high-tech package, but the data quality just comes out and now you're just getting more certain about bad conclusions. I think there's Seinfeld episode in there. That's true. That's really hard. Yeah, yeah. It's a great one. And I think that need to, organizations underestimate something Joel said I wanted to highlight, which is that is their current tech stack prepared to be sort of AI friendly? You might say, well, we have databases. Well, GenAI has a very limited amount of space to store things. And so just saying, well, we'll just stick our database in AI. Like that's not how it works. And there's you have to be very clever about getting the information where the AI can work with it. And so that becomes kind of the new challenge now. So it's not just we have data, we have clean data, but do we have it in a way that's kind of compatible with the way that that GNI needs to consume it? And that becomes a whole new investment absolutely. Yeah. Tanya, from your perspective, building on that, right? What are some things to consider when it comes to data strategy and data governance? Oh, if we could all start over and start with clean data, that'd be great in automation. But I had a client not too long ago talk about we need to replace our systems. And I said, well, okay, so let's think about as you're doing that, can we develop a single source of truth for data that everybody feeds off of? Because to Joel's point, most systems have some representation of a customer, let's say, but they use different terms for it. And different people have access to different systems where they're changing data. And now you've got syncing problems where if you can have a single source of truth that this is where this customer information is residing and who owns it and who's allowed to change it. And every other system can look at it, right, and use it, then you start to get to a point where data makes more sense across the organization and that you can start to build models that help you utilize that data. But it's very hard to do when you're now kind of company that's five years old and you've got all the processes running and holy crap, how do we get to the point that we can centralize? It's, you know, and then people go into data warehousing projects and all of these other projects and they, they don't start with the basics of let's just map the data. Let's just figure out what this same data element is cross caused, caused or called across all the systems. And where should it really live? And that's a big exercise. And then from that, adding on to that, I'd like to hear from Joel, right? At the end of the day, data also has to come from users that are putting it in the correct way or in the right formats, that. Talk about data from a user adoption and maintenance standpoint, Joel. Yeah, and that's very key, right? So particularly if you're dealing with legacy systems or a lot of manual processes, right? Shifting the workloads to your knowledge worker to be responsible for data integrity and quality is not an easy task. But there has to be ownership across the board. And depending on where it resides in the organization, the only way that you'll get real adoption is if all of your stakeholders throughout the chain understand the ultimate outcome. So it's order to cash, right? And how they the tasks that they're doing on a daily basis, what do they mean down the chain? And once you start to educate folks in that process, when they see a hiccup in the process, they know where to go. They go, well, I'm going to go two rungs down or I'm going go one rung up and I'm going to fix this now instead of at the quarterly business meeting. Most organizations though have employees that know how to do their tasks. Yes. And that's all they know how to do. They don't understand how data gets to them and what the downstream to them looks like or what they're impacting when they get it wrong. And so we go in and ask questions, okay, going, well, why are you doing this? I don't know. Because somebody told me to do it. What happens if you don't do it? I don't know. Somebody told me to do it. And so it's getting that education. And when you're doing that constant hiring and you've got turnover. It's gotta be a big part of your data strategy. Yeah. Well, and it's also just gotta be organizationally part of how you live as an organization. It's gotta be that important. Culture of that stewardship at the data level. I think organizations tend to reward people for production, like volume of work completed as opposed to quality of work completed because there's an underappreciation generally of what role downstream these little ripples in the data start to create. At a retailer, if you quote one thing and then contract for something else and then enroll something else and bill something else, the materiality of that, because you can have one thousand of those things out there and then realize, wow, we've been underbilling millions of dollars for years. You can't go back to those customers and collect that most times. I've been in data migration projects where just for a single month for one contract, it was a miss of fifty thousand dollars or for one customer for a year, it was a million dollars. And so those aren't small numbers when you look at it in the concept of, you know, a larger picture. The source of these things are they have, there's some good systems that do one of those things really well. They bell, they book, they quote, right? But that line of sight, this idea of coordination around that single version of the truth is typically or export to a spreadsheet, someone QCs it in Excel. Then there's some FTP uptake process into some legacy system or some middleware that shoots it off to different places. No one person really looking across, right, looking at their, their piece in the chain. And so that stewardship across the processes, I think is a, is a big opportunity. Let us know in the comments out there, any data horror stories that you want to share that you can share, no names will be mentioned, right? So that we can learn from it and lessons learned from your data horror stories. So we could talk about data all day long, obviously, but let's switch to integrations, right? Especially with AI coming on board, multiple tools. Nate, I want to start with you around integrations. There's a lot of terms right now when it comes to that, right? So for those of our audience that are not familiar with all the terms and even myself, like APIs, MCPs, define those for us in a general term and then tell us why they're so important. Okay, so the way that traditional software talks to another piece of software is we call an API. And it's just an interface that both systems understand and allows them to pass information back and forth. We'll keep that as our definition. What's different about AI is that's the structure of that definition is absent. So in a gen AI world, let's just talk about that so we don't get too off in the weeds. Everything is just a piece of text in a stream. And so the way that you actually speak to an external system is something called MCP, kind of an emergent standard. And what it does is you basically tell, I'm going to use ChatGPT as an example. That's not really what it is, but I'm going to use that because everyone is familiar with it. So you basically tell ChatGPT, hey, you have a tool and the tool is called get customer data. And if you need anything about the customer, you just tell me, I want to use the tool called get customer data and you tell me what you want. And then I'm going to come back and give you what you asked for. So effectively, MCP is, it creates an imaginary tool inside of the context of the AI and it tells it, hey, this is how you use the tool. And then you move on with your life of what do you want to do. And every now and it'll say, hey, I need to get this customer data. You said you would give it to me. Now that's what lives in the AI side. But you now have to bridge the gap back into your traditional software world, which is API driven. And so that is kind of a missing link right now for kind of traditional systems, even a systems that I would call you know, up to par from a technology perspective, this is very emergent. And so this, how do we weave together our traditional databases, APIs and now get it into the Gen AI world is something that is really a very frontier thing happening right now. And I'll say one more thing about it which is that what we're doing is basically providing these interfaces like we would ordinarily provide APIs going forward. So we're saying, hey, if you want to talk to your data in our systems, all you have to do is link to this from your ChatGPT. It doesn't have to come through us. So we see that as kind of an interesting hybrid where customers can own their IP. They can build out their own prompts and things in their tool of choice, right, but still have access to the data and functionality inside of a vendor system. So just one model that we think is kind of providing some of that build feel with a buy kind of option to it. Yeah. And as we've been discussing here and even in our previous discussions, right, it's not so much build versus buy anymore. Almost buy to build. We have a video to expand on that a little further. Let's roll that now. As an organization, you have to decide sort of what's the base minimum of, of stuff that you just don't want to take care of. So buy and customize is another one. Yeah. Okay. So there's like a continuum also. There is. I think it was twenty years ago, you literally had to build from scratch or get SAP, like Oracle or something monolithic all in one ERP ish. And it probably wasn't tailored to your industry. And like all of that was on you. And then it's evolved to like, we have platforms like Salesforce where you get like a piece with some sort of generic objects and then you're like, okay, it doesn't know what a meter is. Okay, let me start coding up a meter object. It doesn't know what usage is. Can we code that up? Right. And you're kind of building inside of a platform. So that actually segues into something you and I were talking about a few days ago around your view as build to buy. So can you just dive into that a little bit? What do you mean by Buy to build. Buy to build. Yeah. Buy to build. So I think that, like you said, right, there's multiple platforms out there now that give you, and quite frankly, they come in tiers of how much functionality they give you and relatively how much then you're going to pay for it. Right. And so as an organization, you have to decide sort of what's the base minimum of, of stuff that you just don't want to take care of like EDI or right. What's the common stuff that you want this system that you're going to get to have in it in order to take care of the standards and the, the updates that, that you don't want to have to be apprised of. And does it give you enough flexibility to work within the workflow that's already in it, but customized to how you want to run your business. So your customized pricing or, a customized workflow where, yes, I may want to calculate charges for this invoice, but I don't want it to actually be an official invoice until someone's approved it because it's over five million dollars right? So to be able to put in your customizations, does the tools that you're looking at get you there? More and more software companies are going that way, but not all of them have yet. And just because you've looked at one and it didn't give it to you doesn't mean that there's not one out there. And so have you really looked at everything and have you asked for the vendor going, hey, I like this much of your platform. I don't really need this, but it's missing this. What's the likelihood of getting that? As you can see, there's a whole plethora of options there. I want to call attention to a comment we got from the audience from Nick. He works in cybersecurity and he's seen time and time again, the tools and vendors that organizations buy to bring that software, but then it sits on the shelf and it's not integrated. So millions and millions of dollars wasted there. But as we look towards the future, right, of the next three years even, I'm not even gonna go to five or maybe even just the next year. Right? Like because it's moving so fast. So hard to see. Joel, from the energy industry, right, where do you kind of see software strategies moving towards in the next year, let's say? Well, I think we've been talking about it. I think we'll see organizations to continue to adopt cloud centric platforms to help them transform their business. I think you'll still see a mixture of of the build versus buy. But if if you're if again, if your technology stack is not future ready for the AI explosion, you're going to be caught off guard. So the best way to do that for a lot of especially mid market organizations is to introduce and maintain these platforms that are going to help you do that. Tanya, from your side, what are the risks of companies not preparing for this AI revolution and what can they watch out for and prepare for? I see companies struggling with it today and it's on an extreme. You have people that are like, that are so into AI like Nate, right? They can talk to you about it all day. They use it in almost everything that they're doing. And then I see people that are struggling with, wait, I can use AI to do that? How do I even start? What do I put in? Which tool do I use? And they're so overwhelmed by it that they don't want to touch it. And unfortunately that's probably eighty percent of the organization because it's the technical guys that really want to be in AI. The people that are doing the operational things to run your business aren't tech savvy, right? They know how to do the tasks that they've been assigned to do. And now introducing AI to them and asking them to adopt it seems really like foreign and scary. And so I think, we'd talked before the show about, you know, the centers for AI within a company, or do you distribute it to the individual businesses? And I think putting someone that is passionate about AI with the business, with them, understanding what they're doing and starting almost like drip marketing. Little things to them that helps make their jobs easier on a day by day basis or a week by week basis. Hey, I see you've been running this report every day. Let me do this for you. And it will give you these three things. Oh my God, that would make it so much easier. And don't even, you don't have to use the word AI, right? As a technical person, use AI to give them the tools they need. But if you don't tell them, they'll start adopting it way quicker than if you try and just say everybody use AI. We've gotten to a point where we shouldn't say AI almost. Almost. With some people, absolutely. Yeah. Nate, your perspective risks that companies should start preparing for? You know, I think looking out year to three years is there's this conversation going on right now, which is sort of the death of SaaS. So the death of platforms and so in favor of what? What will people use? And the idea is that there'll be this agent oriented use. So if you think about that, rather than you go to a screen and you put in some data and then you click save and then you go to the next thing and then you run a report. So that sort of might be any kind of software we use today. Instead, what you have is this sort of these specialized agents that do bits and pieces of this work. As an example, we make a commissions platform and one of our customers runs kind of like an eighteen point inspection before they send out all of their payments, sending out several million dollars. They need to check that nothing looks out of line. There's no objective rule that it's out of line, but they want to get in front of a human and get a feel for what's changed since last month and some things like that. So a year ago we would have had to build them a custom like eighteen point export this sheet and export that model and send that to this person and send this email, discrete sort of features that are custom features for them. But now what we've done is we've built sort of MCP native into our solution and they can actually make agents that they write or we can write for them. And they basically can do build this workflow to be done by an agent and escalate to a human as needed. And so I think that type of change in how humans interact with software, not to give it as a threat, but as little helpers, copilot sometimes is the word that's tossed around that sort of help them on this mundane bulk export filter, make pivot table type stuff. I think humans will kind of get out of that business. Your own personal assistant. You say, hey, I really love to have this little script, but I don't know how to program. I'm a commissions analyst. So I think that this chat GPT thing really helps bridge the gap of these people that we were sending to Power BI School, you know, two years ago. And now they can just sort of talk like they do to their chat GPT when they want to plan their vacation, except they're talking to their commission system. So I see that type of thing on the horizon as a real tangible transformation coming. Yeah. And that's what we're seeing in a lot of the software development now and SaaS solutions as well as introducing their agents. They don't really call them that, but that's kind of what they're doing. Right? So they're helping transactionally. They're helping the employee do the work, but they're also helping with that data integrity as well. Joel, I'm curious from a business perspective, because you're in front of users all the time, are cultures of the organization and mindsets of the people ready for these kind of AI agents and to be their co pilots? What do you think the user readiness is for agents? I think it's low right now. For asking the questions of how do you get up to speed and not fall behind, That's that's really the crux of it, right? So not only do organizations need to be doing right now to dedicate resources to this effort, but they need to be need to be committed to increasing the AI fluency of the entire organization because it's not going to just be the technologists that are going to do the traditional utility compute and I'm going to get you a network and I'm going to get you a laptop and and and so forth and so on. Every single person in the organization is going to not only interact with it, but have an influence on how it's used. Unfortunately, I also think it's generational in adoption, right? You're gonna have, I think the Gen Zs, even the millennials are gonna be quicker at adopting it than people my age, Right now there would be some exceptions, but you'll have some that will, just can't get there or maybe don't want to that crazy AI stuff that, you know, they're gonna, they're gonna take over the world. There's a trust factor too, right? Right now, you know, if you got in really early to this idea of hallucination being a regular occurrence and telling you something that just completely falsely Became jaded. Right? Right. You're like, wow, I can never trust this thing called AI again. And, you know, I think we're seeing ways to start to cordon off those hallucinations from certain types of problems, as well as certain types of problems that are very prone to that, like questions of fact versus summarization. AI is really good at that. Highly accurate. Like here's a whole bunch of data. Just give me the thesis. What does it mean? Versus, has this ever happened? Right? That question is very prone. So there's some education, I think, around people of like understanding right tool, right job for AI versus just like this is a stick and I'm just going to whack everything, you know, with the same stick. I think there's still some learning curve and some trust that needs to be built, from the AI community that we can protect users from these types of downside risks. I want to stay on that a little further, like AI literacy at the organizational level, right? Open to any of you to reply. Any tactical tips that you would have for people on how to build a culture of AI literacy or training tips or upskilling tips? So within our organization, we've been doing, almost tech talks and and providing examples of here's how I've used it in my project. Because I think that the more use cases individuals see, the more they go, oh, I never thought about using it for that. And I think it's just getting exposure and seeing success with it as well. And also talking about lessons learned. And so part of what we're trying to share are best practices in how to structure a prompt when you're doing prompt engineering, you know, putting in that I want you to play this role, or here's the, here's the assumptions I need you to make, or here's the audience, or here's the criteria, or here's the guardrails that I'm going to put around your answer become very important. But if you've never been exposed to that, you have no idea that you need to put those things in. And it's that education level that becomes very important. Yeah. There are some organizations today that are doing that. So similar to, you know, HR training, cybersecurity training, there's mandatory AI fluency training that's being introduced. And I think you're going to see a whole lot more of that because again, you know, it's going to matter to everybody. We took one senior. So I would say senior versus junior when you first start in terms of pick any, doesn't have to be a developer, any role, I would start senior, putting these tools Senior role in. Yeah. Versus like, hey, here's a person, someone new. They'll be, they won't have a closed mind and they'll be the great person to bring these ideas. The problem is they just don't know enough really one to hit high ROI type use cases. And two, like what would be relevant outside of their bubble. And then also the ability to recognize something that's not right, where they're going down a rabbit trail. This is not going to be productive. Senior person knows all of those things and you get just a huge multiplier on that person. And then what else did, we took one senior developer from all seven of our product teams. They meet weekly for ninety minutes on almost like an out of band sprint out of our product development life cycle, building only AI features, and they're not allowed to get any other tasks in their product teams. They still sit in their product teams, so they still have connectivity. But we sort of cordoned off some resources to say, hey, we're going to be intentional about doing this. And the cross pollination we've gotten from that, because those little lessons like must versus should language and prompts. You can't say you should never do this because should means usually not, ought to, must means never. So those little subtleties, even we have some non native English speakers that struggle to get there because it is so subtle in the way that you write those prompts. All that type of learning, having it live in a, in kind of a community of practice without going full central lab. Right. Right. Where these people are in the back and don't, they'll come out with the great solution in six months. Yeah. You know, it's Joel, you mentioned like we have to make it matter to the person. Right? And I think of my mother. I know we're talking about energy here. She reached out to me and she said, help me get on chat GPT so it can help me look at my medical labs. That's a great And she's seventy four years That's And so it mattered to her and English is not her first language, right? And so she kind of forced herself to learn it, right? Because her medical labs matter to her. So whether it's their job or their role in the organization, if it matters enough, they'll learn it, And then also I want to point out a comment from Galero in the the feed here. He also mentions adoption will be driven by a company's leadership as well. So it's not just the mindset of the employees, the mindset of the leaders. If the C suite doesn't drive adoption, then it will just be patchwork inside the organization. I mean, we still see organizations that limit the tools that resources can use for AI. And, and that obviously then will squash adoption. And I get there's a, there's a balance between, protecting your data, but there's ways to do it, right? And still get to being modern in your technology. That's right. Okay, Joel, final words for our audience when it comes to build, buy and AI. Well, they all go together, obviously. No, I mean, again, you need to think about your technology stack and is it future ready? So every single solution that's brought up, every business case, you need to think through it properly of not only what am I solving today, but what am I building for tomorrow? Build for that future. Nate? I truly believe in an industry we talked about M and A where we have a consolidation pressure. There's no more comfortable middle. You're either innovating or you're not going to survive. So AI is a natural path and being kind of mid market, you've got resources, but you've still got agility, right? As opposed to bureaucracy, we need IT to vet this tool for eighteen months before we can put one license on the ground. You know, you can move quickly, maybe take a little bit more risk, but you can navigate that. I think there's just a huge advantage and embracing it fully. And then you really have to go and ask yourself those same questions you asked two years ago, but now with a fresh mindset. Is this the right solution? Are we right on the path? Does this align with where we want to be as a business in three years? What do we see it like to look to work here? And we have the right tools in our people's hands. It's a great point. And in fact, I would, I would say even sooner than that, right? Because it's changing so fast. Yeah. I love that there's no comfortable middle. Yeah. And that Nate, you're just full of these and then think of data, messy data as your smelly Jim shorts. That's gonna stick in my mind now. I'm leave you with that. We'll move into the next card. To get that My card does not smell all blue, Roger. And Tanya, last but not least, your final takeaways for our audience. Technology is great. It's only as good as the people, the data and the processes you put around it. And so an organization has to be smart about the choices they make. Yeah. And they're never easy. And sometimes you don't get it right the first time. But that's okay. You got to keep moving forward. Yeah. Well, thank you, Tanya. Thank you, Nate. Thank you, Joel. Absolutely. Thank you for all of you out there in the audience for contributing your questions and your comments and your insights. I want all of you to also note that on October twenty second, that will be our next CG hour. And on that show, we're going to talk all about how to accelerate your AI journey from the data process and people side. So please stay tuned, save that date. Thank you to everyone here. And just a reminder to call out, our sponsor here, CG Infinity, where their slogan is all about people first driven to transform. And we'll see you on the next episode. And as we roll out, just want to share a little bit of CG Infinity's mentorship program as we're talking all about people here. I'll see you next time. You can meet on a walk, you can meet on teams, you can meet face to face, but really you're gonna meet people who are here for you. They're your cheerleaders, they're your sponsors, they're your coaches, and they are going to have your best interest at hand. I I think with giving feedback, you know, if it's, you know, constructive or, you know, or positive feedback, it always has to come from a perspective of caring. When there is feedback that may be perceived as negative or hard to hear, I always like to make sure that we focus on the path forward. Especially with sandwich technique. I think we all know what I'm talking about. So I know like each person prefers getting feedback, different ways. It's easier to just talk it out, get feedback. And, if you incorporate it seriously in your professional life, it helps you manage your folks. Everybody has perspective. However, if you're only talking from your perspective, you're never going to answer or be able to understand everybody else's perspective. I think we're all afraid to say, I don't know or oh, that quite didn't work out and having, you know, this relationship so that we could talk it out and work through it was immensely immensely helpful. Work on what matters in the moment and ideate together and I've really enjoyed and benefited a lot from just hearing a different perspective for whatever problem I'm approaching. We get heads down on our projects and we're not thinking what's the next thing for me, and your mentor can help you do that. Yeah. They say, okay, great. That sounds like all that's going great. Have you thought about doing this? We all get upset, but like the most important soft skill I've learned is kind of just controlling my emotions and kind of looking at it. Just give it thirty seconds, give it a minute and just wait for your emotions to dry out. But it was actually you in a conversation that told me, is it the message, the messenger, or the delivery? Open Door Policy has helped me so much, especially with talking to people and business people. So I grew up having a mentoring program in every company I've been at. So I'm a firm believer in the mentoring process.