Refining Your 2025 AI Strategy with RapidScale’s Duane Barnes

Scott Kinka: 

Hi friends, and welcome to another episode of the Bridgecast. Super excited about this one. My guest today is not only a business associate but also a long-time friend, Duane Barnes. He’s the president at RapidScale. We’ll talk about who RapidScale is and what they do in a couple of minutes. But Duane, welcome to the show.

Duane Barnes: 

Thanks, Scott. Good to see you as always.

Scott Kinka:

Yeah, fantastic. You are. And where are we seeing you from so that our friends online know you’re in Raleigh, North Carolina? It’s a cloudy day where you are, but there’s a view there. Fantastic. Before we jump into your career, tell us a little bit about you. You’re in Raleigh-Dorham. Know where the kids are in life. Just get a little bit of background so we get to know you.

Duane Barnes: 

Raleigh, North Carolina. Yeah. Yeah, I’ve got a 13-year-old or a 12-year-old daughter, almost 13, usual 12-year-old daughter stuff, soccer, basketball, sports, piano, a little bit of everything. My wife has also, of course, been with her for a very, very long time. We just celebrated 23 years of marriage six days ago. So, all is good here in Raleigh. And, you know, we’ve been here since 2007. So, I moved up from Florida and Never looked back.

Scott Kinka: 

Amen. Fantastic. So, you are the president of RapidScale right now. RapidScale is a Cox company, right? So I want to get into all that, but just talk a little bit about your journey. I mean, you started, you really began on the technical side from my understanding and you and I have had some knockdown, drag-out technical conversations over the years. But just kind of walk me through, you know, being opportunistic and starting, you know, as a programmer really and landing as president of RapidScale.

Duane Barnes: 

Yeah, I always gravitated toward technology from a very early age while tinkering with computers at home. My very first job was in dial-up tech support, as you might imagine, was the heyday of the ’90s. You know, it is lots of fun, but I always worked for entrepreneurs and co-founded a couple of small companies; some made it, and some didn’t. But the idea there was just always wanted to be in control of destiny, so to speak. You know, I did spend a little bit of time at Disney, and technology was there. There’s a great learning curve from the entrepreneurial business to the largest of the largest enterprises on the planet at the time. So got a lot of, you know, resiliency and, and continue business continuity, DR type, you know, they can’t go down experience. For example, if you’re at Disney, you don’t want the bookings engine for their travel team to go down. So massive, massive data centers. Then, I transitioned back to the entrepreneurs and stayed there. Over the years, entrepreneurs tend to sell their companies. And most of the time, I’ve moved on. But this last one was a unique one. RapidScale was a great, entrepreneurial, fast-growing company. But when it was sold to Cox, I assumed that as anybody would, you’d maybe stay a year or two and leave, but the Cox family. It’s just such a unique group to work for. They’re very diversified, very much about leaving the world a better place for the next generation. All the assets are really designed to give back to the community and enable businesses to compete on a global scale, which is really RapidScale’s mission. And all of our sister companies have something to do with clean tech, green tech, fintech enablement, and, you know, people around the world living a more prosperous life. And that’s just not normal. You don’t get that with a PE firm. You don’t get that with a VC firm. Don’t get that typically in founder-led businesses. You’re just trying to grow at all costs. So you really do. It really allows us to connect with a greater purpose.

Scott Kinka: 

That’s awesome. I’m going to come back to the RapidScale story in a minute. Then, we’re going to jump into our topic of the day, which we worked on during the pre-call. The story is interesting. You’re working effectively as an entrepreneur to some extent inside one of the largest privately held companies in the US. If not the largest, you’re right up there at the top. And that’s an interesting mix. I mean, tell me. Tell me something that you know now, right? That you wish you could have gone back. You could tell your younger self, if you will, from this super interesting experience.

Duane Barnes: 

Well, I will tell you a couple of things. One is, you know, I think particularly when people are younger in tech, either startups or early-stage businesses, everyone’s sort of trained to ask for stock options, ask for, you know, RSUs or ask for equity. And almost no one ever gets it when they ask for it, right? There really isn’t such a thing as sweat equity unless you were just very lucky to co-found the business with a great visionary. I would tell you, follow the best leader you’ve ever worked for from one place to the next, and you will be very much fulfilled. And I think you’ve got a lot of followers. I’ve got a lot of followers, but I think that’s, that is the thing that will fulfill you as working for a great leader. The company itself is going to vary from place to place, but I look at the Cox family, and I think there isn’t a better opportunity out there. Just isn’t. Sure, I could go take private equity money and start a company or VC money and start a company, but to get the scale that we have access to here, it’s just an unbelievable opportunity. I mentor a lot of people and do lots of speaking circuits. I get asked all the time, you, what piece of advice would you give someone? It’s, you know, the idea has got to be great. Don’t get me wrong. You have to be passionate about what you’re doing, or it’s just going to fall flat, but you really have to be partnered up with a great leader. And, that will, you know, make, make things come to fruition more times than not. Cause you can just watch the people that follow that leader from, know, typically not job hopping, but you know what I mean? For example, every five to seven years, they go do something else, and the same group of people follow them from one place to the next. And it’s successful every time. And that’s not, you know, an accident.

Scott Kinka: 

Yeah. Certainly. I have been fortunate enough to have been surrounded by some people that I’ve followed around and have followed me around, and I get it. 

Duane Barnes: 

I know you have.

Understanding RapidScale’s Business Model

Scott Kinka: 

I mean, it’s serendipitous in a lot of ways, too. And I think. You know, it’s also interesting that you make connections along the way. And, you know, sometimes they’re a partner, and sometimes they’re a competitor. And sometimes you bring them in and make them your partner in the business. You’re right. Think, you know, finding your work family is the way to, ultimately, breed success. That’s great advice. Talk to me a little bit about RapidScale while we’re at it because we’re going to dig into this topic. Maybe our listeners aren’t familiar. Certainly, they’re familiar with Cox, perhaps not with RapidScale. Just give us the elevator pitch there, if you would.

Duane Barnes:

Yeah, we really are trying to. We have two phases of our business, and for the lack of an easier way to describe it, technically, we think of the private cloud and the public cloud. And we view that from a lens of how we help companies simplify IT, right? So make the complex simple and make it the easy button they’re looking for, which allows them to unleash innovation, right? So, a lot of companies are tied up in tech debt. Investments they made, you know, in the pre-call, we mentioned VMware, these enterprise license agreements, things that they’ve just spent money on. And so they don’t have the free money, the capital to go do things that are more interesting and more innovative. And so that’s what we do. We come into a business, we talk about all of their challenges, and we figure out the best way to leverage their existing assets, whether it be on-prem. In a private cloud and in the public cloud, and help them get to a business outcome. And we try to do that in a very consultative way, a very long-term view of how we can partner with a client. It might start with a PS engagement to do an assessment. It could be a full data center migration or somewhere in between, but the idea isn’t just to sell a widget, right? It’s very much a; we want to become a partner in your business. Think of a mini GSI, for example, a mini Accenture or a mini PwC, but that’s just focused on cloud and technology and managed services. And, you know, let us get one successful project after another under our belts with you or the client and take it from there. And, you know, companies evolve over time. They get bigger, they get smaller, they merge, they transact. And we’re there to flex with them. Just like you would do in the cloud. You consume more when you’re busy; you consume less when you’re not busy. We have flexible engineering, flexible DevOps teams, and flexible AI engineers who can come in and really add value to the company. So we really, for all intents and purposes, stopped talking about speeds and feeds and data centers and locations and networks and all of this stuff. Those things are all table stakes. Now it’s about driving. Business outcomes on the investment in

Scott Kinka: 

Fantastic. And then, you know, I get your comment about the line card. Just give me an understanding of the aspiration in terms of what you deliver. Just run me quick, like the categories that you’re in from a technology perspective.

Duane Barnes: 

Yeah, for sure. At the highest level, think of all things AWS, Azure, Google Cloud, and VMware. And when you layer on cybersecurity and professional services around that, the professional services are really deep. You can think of anything from a data engineering assessment of where all the customer’s data might sit and how best to get that to a place where you can leverage machine learning models to just have workloads on a legacy VMware environment, and I need to get them out of a data center to a full-blown AI deployment, potentially using existing software stacks or building it from scratch like a virtual assistant or integration with a contact center or integration with a CRM and everything in between. So there’s really not much we can’t do from a technology perspective in any of the three major clouds along with the legacy VMware stack.

Scott Kinka: 

Gotcha. Gotcha. All right. Well, let’s transition to chat about AI. Mean, know that you guys, you know, given what you just shared, there’s quite a bit of investment going on on the customer side, certainly on the platform provider side around AI. We’re in the hype cycle right now. Curious. You know, we’ve talked about hype cycles a lot on this show. Where are we in the hype cycle in your mind?

Duane Barnes: 

I think it’s not at the top yet. I don’t think we’re at the peak of the hype cycle yet, personally. I think it’s going to continue to be a lot of hype. I would say that there are certain companies, every day, I subscribe to several newsletters that tell you every morning about all the new AI tools that were released the day before. So that’s why I still feel like it’s in the hype cycle. Companies are being founded daily. And a lot of it’s just software plugins and things that make people’s daily lives a little bit simpler. I think that enterprises are still grappling with how they can leverage AI to get a return faster somewhere in their business. And so that’s really where we focus. It could be known, supply chain, know, streamlined supply chain, or streamlined in a sales organization. Uncovering unlocks potential in your product, whatever product the customer goes to market with. And that’s really where we see all the actual use cases coming from. Most times, eight to nine times out of 10, the customers we talk to can’t even leverage AI today if they want to. They don’t have their data properly organized in one of the major cloud platforms to apply a large language model. or any type of artificial intelligence to even predictive AI versus generative AI that just doesn’t have the data there, right? So if we really work with them on data science and data engineering to get their data, extract it from their CRM, from their ticketing system, from their back office and get it into the cloud so that then we can apply models to it to drive the outcomes they’re looking for. And those are usually nine to 12-month-long engagements just to get them prepared. Paired to leverage AI. Once we’re there, massive savings, massive increased functionality, increased productivity, a lot less of the job that people don’t like to do, data entry, work. So think legal vertical, for example. You might have made 10,000 contracts in the lifetime of a business, and you are searching all of those contracts for one particular clause and then building a new contract for a new customer based on all that legacy work that’s been done. Then, the risks, liability risks, or whatever the firm is looking to protect are quickly brought to light. So we’re not paying attorneys to read these things word for word, but we’re using a model to predict the, you know, ultimately where all the risks and challenges lie and point that out and draft a new version of the agreement on the fly using artificial intelligence. So, these are just things that are very commonly done so that we can help customers speed up those processes.

Customer Engagements and AI Use Cases

Scott Kinka: 

So there’s a lot there. I mean, you talked about use cases, you talked about preparation and data, you talked about length of time and project. Let’s unpack a little bit of that. Let me take a step, like one step up. Mean, obviously, we’re in the hype cycle. Like you said, your belief may not be quite there yet. I actually said our summit was a couple of weeks ago, and one of our keynotes was about AI. And I said from the stage, we’re in the era of go get some. And in a lot of ways, kind of where we were with cloud, right? Maybe, you know, eight, 10 years ago, you know, boards of directors saying, go do stuff, no use case. Like it’s a tool looking for a problem, but they have to be able to tell investors and, you know, that they’re leaning in. So, let’s just flesh out the use case conversation. You gave us a good one about legal for a moment, but I think what I really want to know is less like, let’s find examples and more like. Where really are we? For example, if every conversation we’re having is being dominated by AI, you’ve talked a lot about what RapidScale does. How often is the customer coming in with a fully formed idea around what they want to do that you’re executing? Or are you trying to find the nail for the hammer in a lot of cases? Walk me through that process.

Duane Barnes: 

That’s a fantastic question. I would say less than half the time does the customer have a fully fleshed-out idea. They have an idea, more times than not, or at the very least, they’ve been told, as you mentioned, by a board or a PE firm or some type of third-party advisor to look into leveraging AI to solve some problem they have. So they either know the problem they have, or they have an idea of what they might want to solve. Butt they have no idea where to start, more often than not. They sometimes even have been picked out the tool, like Copilot or something like that. They’re like, we want to use Copilot. I’m like, why? We’re happy to sell it to you, but ultimately, if you don’t integrate it with any of your data, it’s just a chatbot at the end of the day. There’s nothing really useful there. So that is the short answer to whether they have the ideas or not. And then, you know, Sometimes we do, we have the hammer, and we’re looking for the nail at the customer’s request. And so we get involved oftentimes, not even with IT. Usually, we’re talking to the marketing team, the sales team, the legal department, the supply chain department, and, yeah, looking for ideas really for efficiencies. Then, in the manufacturing and maybe agriculture space, I’ll use agriculture as an example. A lot of ag companies have tons of IOT data. I mean, you’ve probably seen that in your goings about, you know, they’re collecting data from sensors, data from humidity, temperature, soil samples, whatever it might be. They just collect. They’re not doing anything with it in most cases, right? They have it because they think they’re going to need it at some point. And the data is valued.

Scott Kinka: 

Yeah. Or they’ve got some kind of governance need to say they checked the sensor or whatever. So here’s the log. Great. Okay. Awesome. Yeah.

Duane Barnes: 18:11.696)

Yeah, exactly. However, we found that we can leverage models to predict crop yields. Right. And so we can look at all the historical data, all the historical crop yields, and then we can predict the next crop yield based on current data. And that’s really valuable to an agriculture business. You can use things like computer vision AI and harvesting equipment. Right. So this is kind of a funny story that I think people can relate to is the potato and golf ball story, right? You have a potato farmer who happens to be next to a driving range in the middle of the country. And dirty potatoes and dirty golf balls kind of look the same, right? The harvesting equipment. Potatoes are soft and don’t damage the equipment. Golf balls are hard and damage the equipment. We can install this very inexpensive camera and the harvesting equipment, link it up to an LTE or 5G network, and then train the machine to skip over the golf balls. Right? It is not that complicated at the end of the day, but with current technology that exists in any of the public cloud platforms, we can actually pull that off and make it happen. It also saves the company from downtime and loss of productivity due to damaged equipment. And it’s not as expensive as one would think. It’s easily deployed. We just have to upload tens of thousands of images of the two things that the A and B test, so to speak. You want A, you don’t want B. You want A, you don’t want B. And you train the computer to train the equipment to do the thing that you want it to do. So, those are just a couple of use cases where real-life scenarios and real-life deployments are involved. Most people think of AI as being this super high-tech, cutting edge, scientific, know, sci-fi stuff, but it’s so simple to just deploy nearly off-the-shelf solutions for very basic business, daily technologies, fleet vehicles, and all kinds of stuff.

Challenges with AI Tools like Co-Pilot

Scott Kinka: 

There’s a lot there. I do want to come back to acting like a consulting company, just given both of our histories, but I want to stuff that to the end. Let’s stick with the AI side for a moment. You mentioned Co-Pilot. I mean, it’s a really good conversation. There’s been some fairly damning, you know, things that have been out in the press of late about Co-Pilot. You know, one Microsoft employee says the platform doesn’t do 75 % of what the advertising claims to do.

I don’t know if that’s true. It’s, and a lot of it has to do with some of the, you’re giving it and how you’re using it conversation we were just having, but let’s just talk to that real quick. Like you guys have a big Microsoft practice. You operate around that. It’s probably the easy button, button step in, to get some economy around AI right now. But why would someone say that? Why is the press bad? Like is this a tech problem, a people process training problem, or a little bit of both? Do you have thoughts about co-pilot in the workplace right now and where we are?

Duane Barnes: 

Yeah, I think you could say the top three, whether it’s Gemini, chat, GPT, or co-pilot, all have the same problem. If you’re not a prompt engineer, you’re going to get what you put into it. Right. And so prompt engineers are very good at tailoring. It’s sort of like the old days when you had to search for something on the internet; you actually had to structure your sentence to like with the result you were looking for, and you tried over and over again until you found the thing on the internet. Now Google just figures out what you’re thinking and gives it to you in real-time. We’re not there yet with the.

Scott Kinka:

Yeah. Because Google’s been listening to you in your house all week. So it has context, but that’s another topic.

Duane Barnes: 

Exactly. But that’s where we are with AI, right? You have to know how to ask these GPTs the question in the structured format, then the follow-up question and then the filter to that question to get the answer you’re looking for. The average employee does not, is not a prompt engineer, nor do they even speak close to a structured language in everyday English. So I would say they all have that problem at the moment. They’re all getting better daily. But all of them have consumed the entire public internet of data at this point. So they will only get smarter as more data is generated on the internet. We’re also having privacy conversations and legal conversations about what is okay to upload to these systems. But back to the copilot, I really do think it can be usefully deployed. But you need to do an actual deployment, not just install the license and tell the users how to use it. You need to connect it to your SharePoint, right? Connect it to your OneDrive, connect it to your email, and connect it to all of your document repositories for it to truly be useful. And then there are tons and tons of plugins with tools like Salesforce and Slack and other things where you can have it mine all of that data as well. If you do a proper deployment, it can save you tons of time and productivity, but most people just use it to rewrite their emails, check grammar, and use Word. I think you’re not going to get the return on the expense of that license if that’s all you’re using it for.

Scott Kinka:

Yeah. You mentioned plugins and co-pilot, and I guess you could say this about any GPT, but I’m going to cordon off co-pilot for a moment, right? You mentioned something earlier about the value of any, you know, large language model based on the information that informs it. And certainly, situations are useful in the business when business information is provided to it. Is co-pilot Does Copilot think about that differently than kind of the public LLMs? In other words, if I were to hook up Copilot to my Salesforce, right? Is that data cordoned off for my own use or is it, or is it informing their public LLM? Do I think about that differently than, you know, chat GPT and Gemini as an example?

Duane Barnes: 

Yeah, so for the everyday user that’s not on a very specific enterprise agreement, it is going to be training the public model. Now, there are very large enterprises that can negotiate an enterprise agreement with Microsoft to get a private version of the model. And so their data is completely cordoned off. But you have to be spending millions and millions a year to even get Microsoft’s attention to carve that out for you. And, of course, you pay handsomely for that model to be allocated to you. So I would just say, you just go to buy Office 365 from GoDaddy or RapidScale or Rackspace or whoever, and you just install the basic enterprise license. It’s training the public model. So, we spend a lot of time educating clients on policy and making sure that they have internal policies. For example, our parent company Cox, if I go to chat GPT’s website, the very first thing I get is a splash page from our cyber team that says you’re about to leverage an AI, a public AI model, and you’re agreeing that you’re adhering to our policies and procedures around the data that you upload to it and making sure it’s not company data or sensitive data or anything like that. While that doesn’t deter somebody from doing it who may not understand what they’re doing, It’s a very good safeguard for most of us who would say, I should think twice about uploading a spreadsheet with all my customers to the GPT model and having it rebuild the pivot tables for me.

Scott Kinka: 

Okay, what do I do here? I mean, truth be told, most businesses aren’t even there in terms of governance. I mean, a third of CEOs report that they don’t have any governance steps in place around AI right now. You know, talk about why that’s such. Let me take a step back. I’m gonna rephrase the question. We talked to CEOs who, you know, and CIOs who are going down an AI journey and trying to figure out what their AI journey is going to look like. And I asked that question, and sometimes they look like they’ve got two heads, like I have two heads, you know what, well, I haven’t started AI yet, so I don’t need a governance policy. And I’m like, yeah, that’s not what’s going on. Right. Because copilot’s active, open, know, chat GPT is basically free like your employees are doing it. Talk about that risk a little bit. You know how often you have to get the governance thing squared away before you start solving for use cases.

Duane Barnes: 

Great, good timing. I was just on a panel last week withTrustt Bank, so no, they responded.

Scott Kinka: 

Yeah, I know them. I actually had some AI conversations with some of the folks from Trust Bank.

Duane Barnes: 

Very, very brilliant. Their head of AI for their fraud department just came over from AP Morgan. So they have the heavy hitters in the space. This was literally half of the hour-and-a-half conversation around this question. And what’s happening is sort of teasing out the point you were making a few minutes ago: people are using it, particularly the millennials and Gen Z, who are using GPT every day for all kinds of things that aren’t business related. But then they get the idea of, like, well, maybe this can help me fix this Word document that I’ve been working on for my project at work. Or maybe I can use it to teach me how to build a better PowerPoint or something like that. They go download a plugin, and then another plugin, and then another plugin. And they’re not reading the privacy policies of any of those, of course. And so. All of a sudden, you now have company data from three, four, or five different third parties, including all of the public largely language models. And, you know, depending on what it is, you may say, okay, well, who cares? It’s just spreadsheet data. But the moment somebody puts employee data or personally identifiable information, you’re now at risk as the company, the employee; maybe it made a mistake, but you, as the company, are on the hook for compliance. Your industry may not even be in. But if data is out there and people can connect it back to a human, you very much are going to get fined by some governing body if that data is leaked. And that’s the challenge. As the panel I was on last week with Truist, their cyber team pointed out that the bad guys are using the same tools the good guys are using for fraud. And so it’s, you know, within seconds, these same tools can crack passwords, connect your social media profile and reverse engineer where you are in your role and then recreate your voice and your image and likeness and impersonate you for fraud. And so again, I just, it’s sort of like urging people not to do things that that’s gonna put your company data at risk or your personal data. And I always sort of dumbed this down for most people and say, Hey, would you just put your social security number in the Google search bar? Right. And just tell the whole world or post it on Facebook. And people are like, well, of course not. Like, well, why would you do that with your company today?

Scott Kinka: 

Yeah. Exactly. And I think there’s also a difference between having policies in place for CYA. We have it, and we are trained on it, right? And actually having something happen, right? Cause the damage is done there. You might not be, know, regulatorily or legally responsible because you’ve got a policy in place, but the damage is still done. So, you know, super interesting around that. You, do you, do you guys do governance work? Also, if a listener is on this thing right now, I would love that splash page that’s happening at Cox to happen in my business. Are those things that you guys get involved in?

Duane Barnes: 

Yeah, we can definitely consult there. Typically, if the customer has nobody, let’s say it’s a smaller company and has nobody in IT or no concept of a cyber group, we’d refer them to a third party that specializes in that so they can build policies from scratch like build a cyber policy or a security manual or something like that. But if they have a small team and they have some of those things in place and they just need advice on how to get it to the next level specific for AI, we can absolutely help them there and give them ideas of what we see other clients doing, what we do internally. Those are very simple safeguards because they are just like a speed bump in a neighbourhood or a yield time or something. Just give the user the pause to like; I didn’t realize that I shouldn’t be using this for company data. And that usually will help. Nine times out of 10 prevent the challenges we spoke about earlier.

The Importance of Data in AI Implementation

Scott Kinka: 

So yeah, two final questions. I mean, I could go on forever on this. This has been a great conversation. I just want to double-click on the data for a moment. You mentioned it earlier, and then we sort of got into governance and use cases and skipped over it. Data that we pulled for this other session that I did a couple of weeks ago from CIO Magazine said that only about a third of business leaders believe that they have the right data. It is in place to effectively enable AI, right? I mean, is that you, governance aside? I don’t think so, and it’s certainly not a technology problem. Is that the biggest reason why, you know, AI boondoggles fail at this point? There is just not enough good information to inform it.

Duane Barnes: 

I think a third is generous. Think some of those CIOs might be overly optimistic. So look, I think with the exception of a rare fraction of a percent of use cases, you need to be in the public cloud, right? I mean, unless you’re doing some kind of on-prem modelling at a pharma level where you’re trying to create the vaccine, right? You really need to be in a very specific data set. We need to be in the public cloud. And that doesn’t mean you’re forklifting your VMs and moving them to AWS. That doesn’t get you into a modern IT environment, right? So think about all the terms you and I used to deal with 10 years ago of serverless computing and containers, and like that, it’s table stakes. If you want to leverage AI, you have to use buckets of storage in the public cloud. You have to be in data lakes. Whether it’s structured or unstructured data, it needs to be near the GPUs and the machine learning models that can leverage it in the public cloud. Those models are being updated by the minute across the planet and by all of the public cloud providers. I think that’s not the case; it sounds simple, right? But really, what it means is, and I’ll use Salesforce as an example, they, of course, have their own AI. They call it Einstein, and it’s not bad. But if you want the data that’s in your CRM, and you want to correlate that with data in your billing platform, and you want to correlate that data in your AP or AR platform, and then you want to know how all of that impacts your customer service. Well, you’ve got to pull all of the data out of all those systems and put it in a data lake so that the model has access to all the data. Whether you stream it there in real-time or you store it there off, you know, know, nightly or use some kind of ETL process to get it there. It’s gotta be there. So, data science and data engineering work needs to be done first. And like I said, you said 30 % feel like they’re there. We’ve kind of feel like maybe 20 % are there, maybe a little less than that 15.

Very rarely do we walk into a business where they’ve got everything in a data lake. They’re using analytics and advanced algorithms to access that data and provide insights back to the business. Almost never. So there’s still some cleanup work that we have to come in and advise on. Then, we can deploy whatever you want. And that’s the key. And I think that’s the fun part, really, working with clients to figure out where the data is, where the importance is, what its significance is, tier it, classify it, tag it. It’s sort of like taking inventory in a retail store, right? All the things are out of place on the shelves every night when you close. You’ve got to get it all back together every day. Data is a mess. It’s a mess globally. It’s a mess in every company. And it’s garbage in, garbage out, as the old saying goes. And so, I’m just getting it. 

AI doesn’t care about the garbage, right? It can figure out the difference. The humans, you know, not so much.

Scott Kinka: 

It’s funny, too, because I think there are really two topics there, right? One is data availability, right? Which is in and of itself a great big project. If a business is in the middle of a giant analytics project like that, it’s the same issue that’s got to get done. You got to get the data in a spot. So availability is one. Accuracy is a wholly different conversation. Like, you know, I talked to customers who, you know, good tech isn’t going to fixed bed process, like at the end of the day, like if it’s not written down anywhere, you know, if you’re, if you’re somebody who’s out there going, well, I’m going to use AI to solve the fact that I don’t have my policies written down. Well, that’s risky. You know what I mean? I had a conversation with a customer who said, well, you know, our agents generally do the job. Well, we train over the shoulder. So, let’s just inform the model with call recordings. I’m like, well, that’s how, that’s why Air Canada’s bot made up a, you know, a customer incentive. You know, like that, probably not the route to go. I mean, maybe, at least, do an initial inspection and then scrub clean. But I mean, I think what it is doing is, you know, it all, don’t want to say it’s creating a moment of pause, but it’s, it’s definitely, I think in a lot of ways, if you’re trying to do this right, it’s creating a moment of like inventory and realism inside businesses to go. Are you okay, great? We can turn it on. Awesome. Is it going to do anything? Maybe not. Do we have a policy about whether you’re allowed to use it as an employee or not? Great. All right, what are the real use cases? All right, if we have real use cases, do we even have data to inform real use cases? It’s creating this almost macro level, not AI-specific review of the business in a lot of ways, at least in our experience. I wonder how… Sometimes, if you meet with clients, if you’re like, hey, this is a really great opportunity, but there’s so much unravelling to do before we get to technology that maybe it’s not a great opportunity. Have you guys run into those?

Duane Barnes: 

We have. I would say there are some other maybe external drivers that might not be as obvious to most folks. So the interesting thing is, particularly with the three public cloud providers that we all know and love here in the US and mostly around the globe, they all have massive incentive programs with their sales teams to drive AI consumption. Right? It’s an arms race at the end of the day. Because of that, they actually fund a lot of the work companies like RapidScale do for the customer. So if there’s a, if we can partner with the cloud, write down the use case provided to Amazon or Google or, or, or Microsoft. And there are reasons why you might go with one or the other. They will literally fund the PS work we do to drive the cloud consumption on the GPUs and the AI machine learning models on their platforms. And so that’s why I tend to say to clients is like, don’t assume you can’t afford to get there. Let us help you with that assessment. And we might still tell you, Hey, the juice isn’t worth the squeeze on this one, but there are some other areas through talking to some of your peers in the business or other leaders in your business that we think we can help with. But more times than that, we do find, Hey, let’s start small. Let’s don’t boil the ocean. Let’s start with one use case. Let’s build an RFP database for you and help you respond to RFPs faster instead of needing 10 people to do all of that work. Think of telecom, for example. Our parent company, E-Rate Season, comes around, and they respond to hundreds of E-Rate RFPs at the same time every year. We win some percentage of them every year, but we can now leverage AI to do most of that work in an automated fashion. Based on our data, facts, and things that we’ve done in the past, we do not worry about the machine learning system hallucinating and providing bad data because we’re only training it on our data. So that’s just the key: start small, prove the use cases out with the executive team, show the value, and then add on from there. And it’s like anything in IT. Once people get used to using it, they’re like, this is great. Can it do this thing now for me?

Strategic AI Adoption and Future Predictions

Scott Kinka: 

Yeah, right. I’m totally with you and Duane. I’m about to break our cardinal rule about the length of an episode that we talked about in advance of this thing because I have seven more questions I want to get into. We’ll save them for the next conversation. I would love to. This was super informative, and it was really great to connect with you again, obviously. And we’re, you know, super excited about everything that RapidScale is doing. You know, I think in particular. Your willingness to, I mentioned earlier, sort of act like a consultant, think, and a lot, both you and I have a lot of history on the service provider side. And I think through the last 10 years, you know, that business has been so enamoured with the platform. The platform would be the thing we were valued on that we stayed away from or tried to limit our engagement in the surrounding services. However, the reality is that customer value is in the surrounding services. So I think, To me, that’s been an interesting development in watching your organization as it’s become part of Cox, the story that you told earlier about customer outcomes and leaving the world a better place in the whole nine years. Not to get super esoterical about it, but it feels like you guys are like, yeah, we’ll go do all that work too so that the platform has value, not just make it about the platform. So it’s been super interesting to watch.

Duane Barnes: 

Yeah, we really operate on a, believe it or not, a 10-year plan. So if you think anybody that’s public, you’re quarter to quarter, and you’re measured by the most brutal analysts in the world. Right. And we just take the long-term view, right? We will outlast and outsurvive many of our competitors that will get acquired by private equity and merged and combined. And that’s not to say they’re bad companies or bad people. It’s just that I think that in this world of uncertainty and massive technological advancements, people are looking for, you know, assurances and stability. And I think you are kind of a rare business that can offer that in this amazingly interesting, technologically advancing world that we live in today.

Scott Kinka: 

Well, it’s a great way to end this portion of the episode. Let’s hit a couple of little things to find. I’m gonna shorten it up. I had three or four fun questions. We’ll make it two, so we get done on time here. The first one and it’s a loaded question, but I’m gonna ask you for a short answer. Give me a shameless prediction over the next 18 to 24 months. It could be in tech or sports. Try to stay away from making it about politics and religion. Know what I mean? But tell me something we can look back on maybe a year from now or 18 months from now that we can see if you were right.

Duane Barnes: 

Look, I think, and maybe this isn’t the Nostradamus level prediction, but I think you’re going to see a lot of liquidity in the markets here in the next 12 to 18 months, but a lot of still hesitancy, right? I think we had a long time of low interest rates and a lot of &A activity, a lot of PE firms speculating on startups. And I think we learned a lot during that period about what a real company is at the end of the day. I think it comes down to three things, and we need to get back to the fundamentals there. And I think we’re going to. I speak to a lot of bankers, a lot of PE firms, and a lot of potential. You fundamentally have to have a different product. You have to drive EBITDA, and you have to have a compelling story to tell the customer about what problem you’re solving. And if those three things don’t exist, good luck getting funding. And so I think as rates start to come down, there’ll be a lot of money coming off the table or going onto the table. And they’re going to be looking for well-run businesses with great ideas that help customers. So I think we’ve just been in this super-constrained, high-interest, high-inflationary environment for the last several years. And so I think it’ll be fun. It’ll be great to see what things come out of that and what advancements are made. But that’s my prediction for the next 12 days.

Scott Kinka: 

Yeah, and hopefully, we’ve learned, right? I mean, as the money comes back in, you know, there’s been multiple 2000, 2008, you know, I’d say the mid-2010s, like there’s been a couple of these periods when the economy was moving, interest rates were low, you know, investment was high in tech where we just sort of were happy to throw money at ideas without fundamentals behind them. You know, I mean, you’re seeing right now even, you know, announcements over the last couple of days, the big conglomerates going, wow, these Businesses over here we bought really don’t have anything to do with our core. So, I guess we got to spin them out. So, you know, we’ll see. I hope so as well. You know, we’ll look back super interesting, you know, what’s going on politically and how that might drive the economy. So we’ll, we’ll see. We’ll check that. All right. Two, I’ll make it two more quick ones. What’s on your end table right now? Like, what are you reading? Favourite business book, something we can put in the liner notes for people to hit.

Duane Barnes: 

Great one. I am about halfway through a book called Measure What Matters. It’s a fantastic book. I would have assumed you’ve read it once or twice by now, and I finished Chip Wars right before that. So it’s a great segue because it talks about the founders of Intel and Andy Grove and why Intel was so successful in both books but for very different reasons. 

Scott Kinka: 

Great book. Yeah.

Duane Barnes: 

So, if you’re interested in just implementing very simple things to help your company stay focused and drive results, it’s fantastic.

Scott Kinka: 

All right, now the fun one. Next, we get to our dystopian future. Whatever happens to be, you can insert that based on the kind of sci-fi that you enjoy, whatever dystopian future that is. But only one application continues to work on your phone. You get to choose which one it is. And be careful because the kind of future you’re envisioning may affect whether the app is useful or not, but what’s your one app?

Duane Barnes: 

That’s all. Man, that was a very tough one. Cause I’m not a huge social media guy. So, I hope that stops working at some point. You know what? I’ll apply this to photos and videos on your phone, right? I think as time has passed, it’s become easier and easier to use AI, whether you’re an Android guy or an Apple guy, to find photos with specific things in them when you’re looking for them.

Scott Kinka: 

Yep. Yep. Probably not a bad idea.

Duane Barnes: 

How many times have you been at a bar, and you’re like, man, I did this cool trip, and I saw this? You’re like, I can’t find the photo, right? I’m like, please make that better and better and better. And it’s really the only thing I care about on my phone at this point.

Scott Kinka: 

Yeah, totally. I love it. So, you know, when you’re hidden, you’re hidden in some cave with a, with a shelf full of canned goods that will hopefully last you for a couple of years, the one bit of entertainment you have is the ability to go through your memories in an AI curated way, is a good way to end this episode.

Duane Barnes: 

Memories, man. That’s what we’re all here for. Make great memories.

Scott Kinka: 

Duane, a complete pleasure, man. Always. It’s so great catching up with you. You know, we’ve got a shared history that makes these conversations fun to pull back into every couple of years. We will not wait so long for the next time to chat, and we’ll do it online again. Thanks for your time. I really, really appreciate it today.

Duane Barnes: 

Appreciate you guys having me. Love chatting as always, Scott.

Scott Kinka: 

Will do. Thanks so much.