/ready, fire, aim: the AI tool trap

Two colleagues, one long run in ed tech

James Carman: Welcome to the Forward Slash Podcast, where we lean into the future of IT by inviting fellow thought leaders, innovators, and problem solvers to slash through its complexity. I'm your host, James Carman, and today we're talking to Jason Swafford. Jason Swafford is a technology executive, entrepreneur, and founder of Drag6, a strategy advisory firm focused on AI, technology leadership, and organizational effectiveness. Over the past 25 years, he has led global engineering organizations through the internet, agile, mobile, cloud, and AI revolutions. After co-founding an AI-native lending platform, Jason now advises executives and business owners on how to build organizations that execute, innovate, and thrive in an era of rapid technological change. Welcome to the podcast, Jason.

Jason Swafford: Thanks, James. It's nice to be here, finally.

James Carman: Great to have you. I should add, I also consider him a friend. Jason's a great guy. We worked together at Cengage Learning for quite some time. So I learned a lot of good stuff from Jason. I consider him a mentor.

Jason Swafford: That was a good time. I enjoyed that stint. It went a lot longer than I planned. I was like, I'm going to go into this ed tech thing for about five years, and twelve years later I was like, that was a good run. Education is a great place to spend your time.

James Carman: It's funny, I was reading the bio and it's like internet, agile, mobile, cloud, AI. You've been through it all, man.

Jason Swafford: I know. I hit all the key words.

James Carman: How did you do that at only 27 years old? That is so weird.

The lending platform, and why small businesses do not speak lender

James Carman: So you're now doing strategy consulting in an AI age. Have you exited out of your lending platform? Is that still a thing, and you're doing this other thing as well?

Jason Swafford: No, it's still a thing. Let me back up a little bit and take you through what I was doing. A couple years ago when I left Cengage I started that company. I actually started a slightly different company and it's morphed and pivoted multiple times. If you think about it, like three years ago, AI started becoming really a force in our lives, blowing our minds every couple of weeks with new things coming out. I've been involved in commercial lending as an investor for years, and I've worked with brokers and lenders for years and watched them struggle with trying to support small business owners in the borrowing process. I had this idea, I had it for a decade. I started to execute on it and I pivoted a few times through it. It's still alive, and you know how startups are, we're still grinding.

James Carman: What's the elevator pitch as a lending platform? What makes it distinct from others?

Jason Swafford: It's really focused on the small business borrower. Small businesses don't speak lender, and lenders generally don't speak borrower. If you've ever been through the Small Business Administration and the SBA lending process, there's a lot of jargon, there's a lot of documentation, there's a lot of financial information that banks need in order to lend, and that the government needs in order to lend. And those things aren't necessarily easy for small business owners to get together. The secret with small business owners, and I've been one, I've owned gyms and some other small businesses over the years, is we don't necessarily have our financials together like a corporation does. We don't have a team of financial people and analysts on our team to help us pull those things together. Oftentimes we don't know what a pro forma is. We don't know what EBITDA means. You'd be surprised how many small business owners don't know what these things are. Helping the small business owners put their lending package together so they're more lendable is really the idea. So we've used AI to build the lending package for the small business owners so that they can take it to lenders. And we've also created a mechanism for us to push it to those lenders, so think about brokering it out to multiple lenders to see who's interested.

James Carman: I think you could be right place, right time. I would imagine with AI there's going to be a lot of new businesses starting up, because some of those barriers to entry that used to be there, of like, oh my goodness, it's going to take me eight months of hardcore heads down coding to put this thing together. That's no longer the case. So you're going to be busy, I would guess.

Jason Swafford: The reality is also that, because of the impacts and influence of AI, people coming out of college today, it's a little bit different than it was in the mid-90s when I came out of college. The job market is a little bit tighter, and AI is definitely impacting those entry-level jobs a little bit more than the jobs like we have, where we've been in the industry for a long time, we have a lot of critical mass behind us, we have that advantage. But I look back at myself. I didn't know anything. I got out of college, I knew how to program a little bit on an AS400 because I had to do it in order to get my statistics analyzed for my psych major. That's it. That's all I knew. I was taught on the job. So our hypothesis is you're absolutely right. There's going to be a lot of small businesses starting up, and they're going to need help, and they're going to need access to tools like this to help them get their information together and get the funding they need to get started.

James Carman: See, that's the difference between you and me, Jason. When I got out of college, I knew everything. I had everything figured out. I knew exactly what to do in every situation. Nobody could tell me anything. The thing I don't understand is how, with all of this experience, have I forgotten all of that stuff? I don't understand it.

Jason Swafford: Well, it happens. My dad's told me several times he's forgotten more than I've ever known.

"I gotta get the tools": the conversation happening daily

James Carman: As we coach folks and advise in this era, it feels like when you give them advice on what needs to be done, it feels surprisingly simple, but they're like, nah, that can't be it. Some of the stuff is back to the basics. But it feels like everybody's just wanting, I gotta get the tools, I gotta get the tools, gotta buy all the tools now. Is that your experience right now? Are we past that yet, or is that still what you're seeing?

Jason Swafford: I definitely don't think we're past it. That's the conversation I'm having literally on a daily basis. That's the easy answer. We're inundated with information about tools. The tools seem to be the decision to make. I go on YouTube, go on any platform right now, you'll see people presenting and demonstrating tools. I may get a hundred emails a week from people that have tool companies that want me to buy into their tools. And there's nothing wrong with the tools. I think all the tools are applicable in some way, so at some point you're going to leverage them. The place we get lost is, we're going to have evolution of tools at a rapid, rapid pace. And if you don't have your strategy right, if you're not intentional about how your business or your company is looking at your AI strategy longer term, then what tools you purchase may or may not be the right decision. I think that's the key piece, actually bringing it up a couple levels, so that you can make good decisions around your tooling.

The failure rate, the J curve, and R and D nobody labeled

Jason Swafford: What you notice when you hear people talking about, well I tried this, I tried that, or you see industry reports where they say extreme numbers. I've heard ninety-five percent of AI projects fail. I've heard that a few times. I've actually seen some data on it. I'm not sure that's actually true. I think there's some nuance in that conversation. But I think that we're failing because we're not sure how these gambles or these bets we're making are related back to our business in general, our overarching business strategy. So I think that's why we see a lot of turmoil around that.

James Carman: It's kind of fascinating that you don't really see in other domains that people will just go buy a bunch of tools and then say, okay, now what can I build with that? It's like, I need to build a thing, I have a goal in mind, what tools help me get there? It's usually the reverse in most other contexts. You don't just buy a bunch of hammers and nail guns and say, okay, now what can I build? You say, I want to build a house, what tools do I need to go do that? But it's fascinating how us technology people are. We do love our tools. That failure rate is really interesting. I just did a keynote at Cincy AI Week, and Stanford released their enterprise AI playbook back in April and they talked a lot about the failure and those sorts of things. They studied successful implementations of AI and getting through that J curve, and sixty-one percent of all of those had an initial failure. So I think the key is, if you're going to fail, that's okay. This is all changing around us. Don't throw your hands in the air, learn from it. Figure out what's working, what's not, and take another run at it.

Jason Swafford: I agree. The failure rate is a bit of a news bite in my opinion, because the J curve is real. We're just, as an industry, at the bottom of the curve in a lot of ways. We're failing a lot because we're trying to figure it out. I literally put a presentation together for someone a couple days ago, and I'm not going to bore you with a presentation, but the point was, we are in this fast R and D cycle and we don't even realize it. And we think that we're falling behind if we're failing right now, because it feels like everybody's making progress in this AI space and we're afraid we're falling behind. And the reality is, no, we're all just trying to figure out how all of this is going to work. I don't have a problem with any of the R and D or any of the trial and error and failures and successes. I think the way it's couched, or the way it's related back to the business and the business strategy, is the problem. As long as you're doing it intentionally and in context of your business and what you're trying to achieve and what your goals are and how you want to be seen in your industry from an AI forward leaning position or not, that all matters to the decisions you make, and how free you make it for your employees to take risks and try things and fail and learn.

Start at the business model, not at efficiency

James Carman: Let's say you do start in the right way and you have that strategy. How do you coach folks to actually use AI to start tackling those things, to fill in the gaps where maybe you could never overcome a hurdle, where things weren't possible before but now they are?

Jason Swafford: Where this needs to start is back at the business model, to be honest. With the folks I've been working with lately, the conversation has really gone from, everybody wants to optimize to get more efficient. That's where we're all thinking. How can I get more throughput? How can I get more work done? And they're thinking, can AI accomplish that for me in some way? Can I give my team access to Claude and miraculously they'll be more efficient and effective at what they're doing? So that's where everybody's at right now. But if you take it up a couple notches, and this is where I've been advising mostly, let's look at what AI is doing to your business model in general. Is the introduction or the disruption of AI going to impact your business, and how? Is it going to impact your business in that there's now opportunities you didn't have before that you can take advantage of? Or is it potentially going to impact your business more negatively, in that it's impacting the moat you've created, or what you thought was that moat around your business? Maybe your advantage for years has been that your technology is a decade ahead of everyone else's, and that your distribution strategy is far superior to someone else's, and now AI has closed the distance between you and the competition. AI has made it possible for someone like me that has an idea in the commercial lending space to erode part of that, because I can jump in and in six months have an application that may be able to compete at the small end and chip away at the edges. So everyone has to be intentional about this and realize what the possible impacts are, how they then want to think about that from their organization, and what kind of actions they want to take off of it. So I start there. That's the initial conversation. And it literally is a business conversation. There's really no AI to it, other than me introducing what could happen with AI.

Inertia, ankle biters, and what AI does to a moat

James Carman: Is there anything around established organizations? They've got their processes and procedures, they've done all that stuff, so they have a little bit of momentum, or inertia even. Whereas a startup can be a lot more nimble. There are two things about inertia. It's hard to get it started, but it's also hard to change its direction if you want to. So for that established company, are you advising them to lean into some of the moats they have and strengthen those more? Because when it comes to your systems, it's going to be hard for you to completely pivot and be nimble like a startup could, because they don't have that inertia already built in a direction that's different than where you want to go. How are you advising folks to think about it?

Jason Swafford: Even before AI, I think, and AI exaggerates everything right now, it amplifies everything, so if you asked me five years ago I probably would have the same answer. The startup is more nimble, they have less to lose than the big players. If you're a multi-billion-dollar publishing company that builds ed tech solutions, you're not very nimble just by sheer size and reach, but you've got so many other advantages. If you're a startup and you have zero revenue, you have zero customers and you have this idea, you can pivot all over the place and find places where you can nibble at their heels. But the speed at which a startup today can have impacts on a larger company's revenue is much more significant now. It changes the financial considerations. In the past, a big company may say, I don't care about the ankle biters, they're fine, we're going to keep our eyes on them, we might learn from them, they'll do our R and D, we'll watch what they're doing, and if there's anything interesting we'll do something about it. Now they're actually a little bit more of a threat. They can make their way into the market a little bit more effectively.

Innovating at the edges, with guardrails

Jason Swafford: Getting back to your point, are there things you can do when the flywheel's moving and you've got so much inertia that it's hard to pivot? The answer is yes, there are things for you to do, but you've got to break it into smaller chunks. The one place that I think larger companies can actually take advantage of AI, and it's probably a little bit contrarian to what most people are thinking, is where we've tried to constrain the edges of the business, the business units, from being quote unquote innovative and building their own things. Let's say, engineering on the fringes of the company rather than allowing IT to handle those pieces of it. I think there's an opportunity with the introduction of AI to start to allow that to happen, but with some guardrails. And I think this is where the pivot from our way of thinking about IT should be, to start to open up the ability for your marketing department or your product department or your sales department to run a little bit with solutions they may be coming up with themselves, that actually allow them to do better business and to be more nimble like a startup. There's opportunities to do that.

Shadow IT is friction looking for a way around

James Carman: I've heard folks talk about that, innovate at the edge within your organization and protect your core business functions. This topic comes up a lot, and I don't want to turn it into a whole shadow IT, shadow AI conversation, but you hit on something interesting. I think we've thought about shadow IT and shadow AI the wrong way for a long time. I think that can actually be a superpower. In some sense it wasn't the same problem that it is today. Somebody might have spent a weekend coding something up, but you could only get so far in the past. You're doing ankle biting at that point. But now maybe the salesperson, they're biting at the kneecap. They can get a lot farther in a weekend with AI than they used to. So I do think there are some ways we need to be leaning into that shadow AI. Don't take it out of the shadows, make it, this is part of our business. As you said, this is an innovative approach. You are innovating for us. Bring all that stuff out in the open and foster it. That's a learning culture. I think we're going to need that. Our knee jerk is you gotta shut it down, because that's what we've always done. But I think we're wrong.

Jason Swafford: I agree. I think it's a pretty valid discussion to be having, because this is one of the places where we have to change the way we think about things from an IT organizational perspective. And let's be honest about why shadow anything happens at the edges. It's because they're trying to get around friction of some sort. They need something. And you know how the conversation has always gone. The business units need something. There's only so much finite capacity and budget for us to support the businesses in some way. There's not endless amounts of money or capacity to do things. So we need to prioritize what work we can do for these folks on the edges. Some things are just not going to get done. They get lower priority, they get deprioritized, and that's just the reality of it. That still feels like friction to the business units. And then internally, from an IT perspective, we spend a lot of time in our organizations going, man, if I had more capacity, if I just had two more developers, if I had five more developers, if I had a QA team, we could get so much more done. So we've got this dissonance between these two things. They want more, we want more, finite budget.

Governance becomes enablement

Jason Swafford: Now we have this scenario where we can allow them to be the front lines a little bit. Start the rapid prototyping. Get early validation from your teams, even if it's internal. It doesn't have to be a product that goes external. Get internal verification. Start using it. But let's have some guardrails around it, some rules around it that say once you get a certain amount of critical mass in your organization, we have to start having conversations about this. And from an IT organization perspective, I think this is an opportunity for us to start building our agentic workflow force that sits around this enablement that we're doing for the business. The business is enabled, and we allow them to leverage agentic capabilities or skills that we allow them to have access to, but we have some checks and balances in place to make sure they're not doing things that they shouldn't be doing. Let's check to make sure they're not introducing security vulnerabilities. Let's check to make sure they're not publishing personal information that they shouldn't be, or exposing it in ways they shouldn't. Let's put some things around that for them, give them tools, and let them run a little bit. Let's see what happens. And that's way more scalable now than it was when we had to have people doing that, because we can have AI help us do that a little bit more than in the past, where we would have architects or solution architects running around making sure people weren't breaking things and causing problems.

James Carman: There are two things there. One is, I think AI is almost a Rosetta Stone. In the past, the business would have to verbalize and tell us, this is what I want. They're trying to articulate it to us, and there's a paradigm mismatch in how we talk about things. There's a language barrier there. But now they can actually build something. This is what I'm trying to build, this is the best I could get to. So we can see an actual product, and it may be rough, but I see where you're headed with that. I think it'll be a good communication tool. The business can do some high fidelity mockups, or even a working prototype, and give us a sense of where they're headed. And like you said, they can play around with, I like this button over here versus here. They can do some of that early work, and then we can help them make it a real product. The other thing, and I think this is the other big conversation I see going on right now, is governance. You talked about the guardrails and let's make sure they don't break security. I don't know that we thought about governance that way in the past. I know some highly regulated organizations don't think of it as an enabler. But what you're talking about is building those governance tools up front, making the right thing to do the easy thing to do. How do we right-size that governance so that it can actually speed folks up and not be a bunch of checkboxes and ServiceNow tickets?

Jason Swafford: That's what I'm testing right now. And I want to be transparent, we're in early days of these things as well. We're trial and erroring things. But the current hypothesis around governance is helping enable IT to help enable the business to do these things, because they know their business better than we're ever going to know it. They know what they're trying to do better than we do. So if they're in Replit, or if they're in n8n, or if they're trying to automate their business processes and they can do it, then we can learn from that, or we can give them tools. One client I'm working with, I said, let's give them a repo that they can work in. Let's give them an agent that spins up a repo for them so they don't even have to know how to do it. They just ask, it spins it up for them, and now they've got their Claude Code environment, if we want to give them that, or whatever environment we're allowing them to use. And now they've got a few skills in there and they've got some other agents in that ecosystem for them. So the governance has now become enablement. To your point, you just said it a minute ago. Governance has become enablement. Yes, we've still got guardrails around it.

The handoff point, and who supports it after that

Jason Swafford: And then we've got to be intentional about the handoff decision points. So if someone's building an app for themselves and a couple other people on their team, and they've got a small pod, and they're using it because they can hit an MCP and pull some data out of Salesforce and massage it in a way that makes it really customized to what they're doing, and if that person that built it leaves and goes somewhere else the next day and it falls apart, and no one uses it again, who cares? It is what it is. Let them do that stuff. If that thing becomes bigger, and all of a sudden not just this pod but three other pods want to use it, or maybe there's an opportunity to scale this for the rest of the company, those are the important points where IT has to get involved. What does this mean? Who's going to support this thing? Because it's not going to be the salesperson that created it. They're not going to support five different organizations. They were comfortable building it for the couple of people on their team and making sure things work and fixing a bug here and there, and that's fine. We can allow things like that to happen. But once it starts to scale, that hardening, that operationalizing, needs to happen. There needs to be an intentional mechanism in the organization for doing that. I believe, and again it's a hypothesis still, that what's going to happen is the business is actually going to understand what we've been trying to tell them for years. They're going to understand it better because they're going to feel it. We've been saying, look, we can build this for you, but do you really want us to maintain this for the next 10 years? This is a cost. We're not going to be able to do anything else because we're going to be maintaining this thing for you. Do you really want that? And they're like, yes, I want it, I don't have to maintain it, build it. And we do it. But now they'll feel a little bit of that pain. They're like, is this really worth maintaining? Do I really need this? So we'll feel each other's pain a little bit more, I think. It might actually bring us together a little bit.

Where it breaks: CVEs, and twenty people building the same thing

James Carman: I love that idea. I think that's a real possibility. The areas where some of this might start to break down, and I love what you said about give them the repo, give them the tools to build these things, where things start to get a little crazy is if that salesperson builds something for their pod of three or four people and they're using it and they're happy. There are a couple of things that can go wrong. They're not caring for the thing like an IT team would. So for instance, if a CVE comes out, something that's now a danger, a risk to the organization. As long as we make these things easy to do and they're spun up and they see the light of day and they're not done under someone's desk, we can lean in and say, hey, you've got a CVE lit up, let's help you fix that. Or hey, you need to upgrade your library, whatever the case may be. The other thing, and this came up in the conference I was at this week, is for bigger organizations, one salesperson builds something and then you've got 20 different salespeople building that same exact thing. So that idea of keeping these things above board and communicated, people talking about it and sharing freely, we've got to solve that problem, so that salesperson one built something and then two weeks later salesperson two doesn't just decide, I'm going to build the same exact thing. It'd be, wait a minute, hold on, salesperson one over here said they were doing something around this, let me call them, let me talk to them. So that visibility of what's being done in the organization, I think that's key. And that's another reason why we need to lean into this shadow stuff. Let's put a light on these things and celebrate these innovations.

Jason Swafford: I agree with you. I think it's a balance. This is one of those things that us IT folks are always optimizing. Let's optimize. We don't want redundancy. We want to control cost, control reach, control risk, control exposure. All of those things we're trying to do, and it's all valid. There's nothing invalid about it. But this balance of, if five salespeople have built roughly the same thing, what is actually the impact of that to the business? And if we have the right tools in place that are going to scan it to make sure that there's no security risks, that are going to scan to make sure that we've got things meeting the conditions and criteria that the business has in place, and we're able to do that across a broader swath of these one-off solutions, maybe some of that isn't as big of a risk as it has been in the past, and we can allow some of that to happen and then consolidate over time. But your point is well taken. In general, building ten things that do the same thing is probably not a great idea.

James Carman: I'm a software engineer, so we've always had things beat into our head, the DRY principle, do not repeat yourself. If you do something two times, go and make a library for it. And I've learned over time, because when I knew everything coming out of college, I would immediately want to fix that DRY situation. But what I learned was sometimes some duplication is fine. We saw it at Cengage, where we would try to create standard component libraries for, let's say, assessments. You want to do a multiple choice. Well, five different people would have five different opinions about how multiple choice questions should work. So we had to put a bunch of configurations into the one uber model, and it's like, no, just have five different flavors, because it's really not that complex of a thing anyway. It's a multiple choice question. But there are some times where you've built the five lions and they're all great, but if you were able to combine all five of those lions, you got Voltron. You got something super powerful. I had to get some sort of a reference in there, and that was a very ham-handed one.

Ship It or Skip It: the Chief AI Officer

James Carman: I think we're ready for our next segment of the show, which is Ship It or Skip It. A lot of companies are spinning up a new role in their C suite. You're seeing the CAIO, Chief AI Officer, having a C suite position. What are your thoughts on that? Do you think this is a good idea?

Jason Swafford: If you ask me this question on different days of the week, I may have different opinions about it, to be perfectly honest. I think in general it's a good idea, because organizations need somebody focused on this. What my role has been a lot is, I am the person focusing, so that the CTO and the CEO and the other C suite folks don't have to focus on AI to figure out what AI is going to do in their organization. Generally though, what I think is interesting about the position is where it should sit in the organization. What kind of role is this? Is this a resource role? Is this a technology role? Does this eventually just roll into technology? Which, as a technologist, it feels like AI is a technology. But the jury's still out on whether or not AI will ultimately be something worth having its own structure for. It's going to be like saying we need our chief electricity officer, or something along those lines. I think eventually we get to a point where we don't actually need this, because it's going to permeate everything within the organization.

Jason Swafford: But in general, you need a strategy, and you need to know how AI is going to work within your organization, and you need focus. I'm advising a bank right now. We haven't gotten into actually doing work for them yet. But when I first started talking to them, they said, we've got to get in front of this AI thing, we don't know exactly what we're going to need to do, there's all this compliance we need to deal with, so we put together a committee and they meet once a quarter to talk about AI. And I said, you're talking about AI once a quarter? That's probably the problem. You're only gathering a few people from a few different departments together every three months to talk about what you should do for AI. I just don't think realistically that's going to work for you. So you need someone that's going to take that lead in your organization. So I think that CAIO role is honestly going to be a short-lived thing, maybe a couple of years, then it's going to be worked into the organization a little bit differently. I do think there's an interesting debate to be made about whether or not, as we get into the agentic workforce, there is this resource, almost like a human resource function, that is your agentic workforce function. And there may be a controversial CAIO role that is more like an HR lead.

James Carman: I've heard that, and the phrase is, is it anthropomorphizing? It's funny that the industry is trying to anthropomorphize these agents, like they're now human-like almost. I'm very skeptical of that whole idea. But I'm with you on the Chief AI Officer role. I think you have a couple of options. Somebody needs to learn it, as you said, and somebody needs to own this. But you've got to get your head wrapped around it. The CTOs and the CIOs in these organizations are already busy. They already have a full job that they're doing. So you could do one of two things. You could pick off some of the responsibilities of your CIO or CTO, whomever you want to have own this, and delegate that to other folks and peel them away and give them time to wrap their head around this AI thing. Or you could bring in an expert from the outside and delegate to them and say, okay, you figure it out. I guess it depends on the CIO or CTO. Do they feel comfortable delegating that responsibility? Some people are just geeks and they want to learn it themselves. But I'm with you. I love the chief electricity officer. I was thinking chief Microsoft Office officer, or chief water officer. I think over time people will be more comfortable, it'll just be part of our DNA, and a dedicated role will not be necessary. You may have an AI strategist or an AI engineer, but I don't know that you're going to need a Chief AI Officer at the C suite level.

Jason Swafford: I think that's where it's going to progress to. I don't know how quickly, but that's where it'll go.

Ship It or Skip It: vibe coding

James Carman: One we've talked about a lot: vibe coding. It's funny, you hear different versions of what vibe coding is, but I'll frame the question. What I consider to be vibe coding would be folks who don't have experience with any sort of software engineering skills in the past whatsoever, and they're using Claude and kind of YOLOing it. They're one-shotting these things and just letting Claude or Codex or any of these tools build software for them, and just, let's ship it. That's what I'm talking about. What's your take on that? Are you pro vibe coding? Are you anti?

Jason Swafford: I'm generally pro-vibe coding, I guess. I don't like the term vibe coding, to be honest, but I'm pro-people having the opportunity to get their ideas out of their head and be able to build. As someone that's never been the best coder, but has had ideas, I've always had to rely on these fascinatingly skilled people that can actually take ideas and turn them into things. And my whole life through, I've had this frustration oftentimes of, I know solid architecture, I know how to build solid software, but I'm not the virtuoso that's going to sit at the keyboard and write slick code. It's just not my skill set. But when you introduce these tools to someone who's got the ideas and has the ability to then turn them into something, I think that is really compelling, and I think that's going to skyrocket us into some great places. There are going to be some unbelievable products and ideas that get innovated over the next five, ten years. What I don't like about it is the sloppiness of it. The sloppiness of just build something, think it's scalable, think it's secure, think it's something that can go to market. And if you're not listening to your technology partners that actually understand what good software looks like and behaves like, you're going to get into a difficult situation, especially if you start trying to push those things out to customers or to large scale groups. The potential for problems is significant. That's why you need to have, again, guardrails is not my favorite term, but the guardrails around how you're going to allow people to vibe code within your organization. It can't just be a free for all. You've got to have some constraints around it.

Who actually gets hurt by mistaking a prototype for a product

James Carman: The consensus seems, and this is my thinking currently about where we're going to land, and I say this every time, I don't want to sound gatekeepy or elitist, but I think what organizations will learn is that vibe coding, the uninitiated using these tools to build things, is great as that resource on the edge. Allow folks to express themselves and build these semi-workable solutions. They're not enterprise grade yet, but they're able to build these things and express themselves. I think that's a great R and D tool. I think it's great for innovation. But the mistake would be if your actual software production pipeline and your software production capacity in your company is completely comprised of folks who haven't run things in production before. That's where I think people are going to start running into those barriers and they're going to learn some lessons the hard way, because the allure of, I can pay someone fifty thousand dollars and they just sling stuff together with Claude Code, that's tough to compete with on cost. I think those hard lessons are going to have to be learned. But I really hope we do settle in on leaning into the innovation, leaning into the expressiveness of these tools, and helping people communicate with one another. And then once it comes to, these are the things that I'm going to charge people money for, or really supporting my core business, the software that's built for those types of things, you're going to need more rigor around that.

Jason Swafford: I'll give you an anecdote from a couple experiences. I am actually not as worried about the larger companies. The larger companies that already have established productionized IT teams that do this stuff and have rigor, I'm a lot less worried about them. I think they need to figure out how to innovate on the edges, and they need to figure out how to use their business units that are willing to put their capacity toward it as their R and D. But they're going to follow the same rules that they've always followed, where they R and D, then they productize, then they optimize, and then they build efficiencies into the systems just like they've always done. I think that system's still going to work really well for them. What I worry about is the mid-size, the smaller companies that don't realize that that rigor is important, and they hire, to your point, or someone builds an application on a weekend and then they put it out in front of their team or in front of their clients, and then real problems happen. They didn't realize that it was still a prototype that they think is productionized, that hasn't been optimized or scaled or secured in any way. They haven't followed the rigor that these companies that you and I have worked with for the last couple decades already have.

James Carman: I think the savvy companies are wise enough to not go down that path. Now there are going to be those exceptional cases that do attempt this, but those will be the counterexamples of why it doesn't work, and other people are like, whew, glad we didn't do that. But a lot of our business is rescue projects. I saw a tweet, if they're still called tweets, the other day about a guy surmising that years in the future there's going to be this new thing, like an AI rescue engineer. All this AI slop that's being built for years and years, five, six, seven years in the future, they're going to need all those engineers to come back and rescue all this stuff. Maybe there's some truth to that. I always think about Demolition Man when this comes up, when we had to thaw out Sylvester Stallone so he could fight Simon Phoenix, because nobody else knew how to do it.

Jason Swafford: That is one reference I understand. I know the story.

The lightning round

James Carman: Let's move on to the next segment of the show. I don't know if we fully prepared you for the lightning round. This is the meat of the show. This is what everybody tunes in for. From my understanding, what we hear from our audience is they kind of just fast forward through all the technical mumbo jumbo stuff to get to this. We don't even edit the first part. If we screw up during the first part, we don't even edit it, because nobody's listening anyway.

Jason Swafford: Okay, so this is the important piece of the show. I'll try not to disappoint.

James Carman: This is where things matter. There are right and wrong answers to these questions, and they will be graded. We have an algorithm that we run this through. It takes a couple of months. It used to only be about 12 days, but token usage is expensive now, so we kind of have to pace ourselves. It's twice as much token usage now with Fable. We'll get back to you on what your score was. I don't know the right answers, and I don't even know how the algorithm works. They tell me it's world class. I say all this because a lot of people come into this and they think, lightning rounds, these are just silly questions. That's not what we're talking about here. This is the real deal. These are high stakes questions. I want you to mentally prepare yourself. Are you prepared?

Jason Swafford: I'm getting nervous. Sure.

James Carman: Number one question, again, high stakes. Do you snore?

Jason Swafford: Yes, I do snore, unfortunately. I don't believe I do, but my watch tells me I do and my wife definitely tells me I do.

James Carman: I get smacked in the middle of the night every now and then and I look over and my wife's like, I know she did it, I'm not smacking myself, but she goes back to sleep. How do you smack a person in the face and go back to sleep?

Jason Swafford: I get a kick to the ribs once in a while while I'm sleeping, to wake up and roll over.

James Carman: Name a primate besides monkeys and apes.

Jason Swafford: Does a bonobo count? I don't think it's technically a chimp. It might be considered a chimpanzee, but I don't think it is. They walk more upright. They almost walk like humans.

James Carman: Like I said, I don't know the right answers to these things. A bonobo. Really? I'm now going to look into that. Are you a sourdough or a wheat kind of a guy?

Jason Swafford: That's an interesting one. I think generally the taste of good sourdough I really enjoy, but I probably eat wheat more than I eat sourdough. So take that for what it's worth.

James Carman: Or maybe neither. You're a fitness guy, so maybe you're like, no carbs. I feel like it's a PR problem for the foods with the word sour in them. Other than sour patch kids. Like sour cream, you know what I mean? Like, is it spoiled?

Jason Swafford: Some of the best breads are sourdough. I don't even know how they make sourdough bread. You can get some crappy sourdough at the grocery store, but I've had some really good homemade sourdoughs that are unbelievable. And they're supposed to be better for you, from what I've understood. There are some enzymes or something.

James Carman: That's what I hear. Gut health, something or other. How many cups of coffee do you drink per day, roughly?

Jason Swafford: Six or seven, maybe ten. A lot. I have a fancy coffee maker. I've been drinking coffee since we've been sitting here. I get up around five o'clock in the morning and it's game on until about eleven, and then I try to cut it off.

James Carman: Are you a black coffee guy? Are you putting stuff in it?

Jason Swafford: My first coffee of the day has a scoop of collagen in it, a chocolate collagen, but after that it's all black coffee.

James Carman: I was buying that Black Rifle coffee and I like it, but I read the other day that it's not one of the more clean ones. Now I've got to find a different coffee. Everything's going to kill you these days.

Jason Swafford: I drink Bulletproof. Bulletproof's coffee seems to be relatively clean from mold. I can tell. I get congested.

James Carman: Have you ever seen a kangaroo in person?

Jason Swafford: Yes. And I believe it was at the Toledo Zoo. It was a long time ago. You know what, I think I saw a kangaroo at the Detroit Zoo, if I'm not mistaken. I've never been to Australia, so I've never seen one in the wild, only in captivity.

James Carman: You mentioned Toledo, so this tees up this question. If you were given an all expenses paid trip to Cleveland, Ohio, would you take it?

Jason Swafford: That is a great question. All expenses, I guess I would, sure. I mean, why not? I guess I would go, maybe if there's a basketball game going on or something.

James Carman: They've got the Rock and Roll Hall of Fame up there. So, Cleveland Rocks. This is one we've been asking people on the Forward Slash for quite some time. It's kind of a little bit of a social experiment. Do you like the smell of gasoline?

Jason Swafford: No. I do not.

James Carman: No, okay. That's fair, which is weird. For a long time we had everybody saying yes. And typically the answer was something along the lines of, yes I do, because it reminds me of going on hunting trips with my grandpa, we'd stop and fill up the truck. It was always an associative thing. But I don't think anybody was like huffing gas or anything.

Jason Swafford: I have recency bias probably. When I'm pumping gas and then I have gas on my hands, it drives me crazy, and then I smell it the whole time I'm driving. So I think that's probably my recency bias not liking the smell of gasoline.

James Carman: And sometimes it's bad. How does that work? It's not like I'm squirting my hands and washing them off with the gas when I'm done. How does it get all over the handle under normal usage of these tools?

Jason Swafford: All I can think is that when people are pumping gas and it keeps kicking off and it doesn't want to pump any further because it's got a mechanism that says you've got enough gas in your tank, they keep clonking it until it overflows, and then it gets all over the thing, and then they just stick it back in the slot, and when you pick it up you've got gas all over your hands. I have to take a page from my dad's book at some point in my life and get gloves that I keep in my trunk. Maybe when I turn like sixty-five I'll do that eventually. But that's my dad's trick. He's got his gloves.

James Carman: My daughter has this little silicone thing that goes in the holder for the cap inside the lid, and she puts that on. What temperature do you like to keep your thermostat at?

Jason Swafford: At night I like it to be cold, so I like it around sixty-five at night. And then during the day, right now it's pushing eighty outside here, so it's probably around seventy-two degrees inside. In the winter I keep it about sixty-eight during the day, let it get down to sixty-five at night. And then in the summer I keep it around sixty-two, sixty-three during the day, and then I like it cold when I'm sleeping, so I crank it down and freeze everybody out of the house.

James Carman: I like to sleep when it's cold too. On a scale of one to Mariah Carey, where would you rate your karaoke skills?

Jason Swafford: Oh my God. Like one. I'm horrible. I'll give you a little quick anecdote. We had a charity event at Cengage, and the event was, three of the leaders in the Farmington Hills office, and of course my team paid the most money possible to get me to have to sing karaoke. And I had to sing Right Said Fred, I'm Sexy and I Know It. And they videotaped this, and I know this video is floating around, so hopefully no one posts that video. It was terrible. And I had to do it at Dave and Buster's, which was even more embarrassing.

James Carman: Out in public, so there were people outside the company. Those things are humbling. We had a dunking booth for one of our all team events. We do a summit every year where we bring everybody in town. As a leader you try to do a good job and try to be nice to people and do the right thing most of the time. But man, I'll tell you what, when they get that opportunity to dunk you, they're going to come after you. They're not like, we'll take it easy on James. It felt almost like it was personal. At some point I think the guy ran up and just pushed the button, even though he missed the stupid thing.

Jason Swafford: For a good cause. Yeah, that's personal for sure.

James Carman: I probably did do something wrong, but you know. Last question. Do you Instagram your food?

Jason Swafford: No. I'm not an Instagram my food kind of guy. I will say that when I travel, especially for business and I'm going to a nice dinner, my wife's a foodie, she loves food stuff. I never Instagram it, but I always take pictures of all the food that I'm eating so that I can share it with her. Sometimes it makes her mad because she's jealous, but most of the time it's like, look at this ridiculously awesome meal I'm having at this fancy steakhouse. Doesn't it look nice? I'm not a big post-things-on-social-media guy. It might just be my age. It's just not my jam.

James Carman: I'm more of a share-with-my-text-chat-people pictures that I think are funny. And you don't share those photos as a nanny nanny boo-boo. You do it because you think this looks cool. You're off in another city and she's taking care of the kids and all the things back home. I had one that I sent home. There was a place in San Francisco, I don't even remember the name of the place, but they served a pizza and then they would just dump salad on top. And I thought, okay, I have to try this, this just sounds crazy. It was really good.

Where to find Drag6

James Carman: That does it for the lightning round. Anything coming up you've got going on? Are you going to be at a conference or anything? Or do you have a one man show on Broadway?

Jason Swafford: Nothing specific right now. We're working on some things, so maybe in a couple months when we get things rolling, we'll come back and share some stuff. But nothing specific. We're just in the trenches, working with clients and trying to build a business.

James Carman: And Drag6, how can they learn more? Where can they find more information?

Jason Swafford: Really easy. You can go to dragsix.com. You can look me up on LinkedIn. I think my LinkedIn profile is jswaf, J S W A F. So you can find me there, and you can reach out to me, either DM me on LinkedIn, or come to our website, or you can reach out to me at jason@dragsix.com, and it's drag the number dot com.

James Carman: That's pretty easy. We will definitely make sure that when we post this episode, on LinkedIn we link out to you so folks can come check you out and see what Drag6 is all about. I'm excited for Drag6 for you, man. You always do good stuff and I'm looking forward to seeing what you get into.

Jason Swafford: Come check out what we're doing. We're really excited too. It's an exciting time. Honestly, I've been through a lot of stuff in this industry in the last twenty-five years, and this is on par if not more than all of it. So I'm glad I'm getting a chance to do it. It's going to be an interesting ride.

James Carman: Well, you're a healthy guy, you've got a ways to go. You're not drinking that mold coffee, so you've got many years.

Jason Swafford: I'm trying to stay healthy, trust me. I'm trying to keep things together.

James Carman: That wraps up this episode. If you'd like to get in touch with us here at the Forward Slash, drop us a line at theforwardslash@callibrity.com. See you next time. Jason, happy to have you on the show today. It's been a really great conversation, and as always I enjoy talking with you about it. Thanks for coming on the show.

Jason Swafford: Thank you. It's been a blast.