It Might Look Right: AI and the Human in the Loop with Scott Wintheiser

PIVOTING TO TECHNOLOGY ADOPTION
EP135: It Might Look Right: AI, Websites, and the Human in the Loop
Guest: Scott Wintheiser, CEO & Co-Founder, Lightburn
Host: Donna P. Mitchell, The Transformation Authority™
Published: Wednesday, November 4, 2026

FULL EPISODE TRANSCRIPT

Donna P. Mitchell (0:01): Welcome to Pivoting to Technology Adoption. I'm your host, Donna P. Mitchell, The Transformation Authority™. Every week I sit down with executives, strategists, and innovators to tackle the one problem no one's solving: why 70% of technology transformations fail. Let's pivot. Good morning, good afternoon, good evening, and welcome to Pivoting to Technology Adoption. My name is Donna P. Mitchell, and today I have a great guest, Scott Wintheiser.

Donna P. Mitchell (0:32): And Scott is involved in a lot of the transformations of the past. He is currently CEO and co-founder of Lightburn, and at the current time, a Milwaukee-based company. And he's going to share with us exactly what they're doing because he's got a lot of history, just like myself, about 28 years. And I'd rather have him tell us what he's doing now and what brought him to pivoting to technology adoption. Scott, say hello to your audience and ours.

Scott Wintheiser (1:02): Hello. Thanks for having me.

Donna P. Mitchell (1:05): You're very welcome to be here. Tell us exactly what you're doing now. You mentioned websites, you mentioned transformations, digital transformation. Give us some insight on what you're doing now.

Scott Wintheiser (1:17): Sure. So we're an agency out of Milwaukee, as you said. Um, we've been around 28 years. We started as a traditional design and dev shop. We would build and design websites, you know, very simple things at the beginning. Um, now our thing is really, we work primarily with manufacturers heavily on a continual improvement type cycle. So we work with them to constantly improve their website, their digital experiences for their customers and their buyer journeys. And then we put a lot of effort into digital transformation so that we can create as many, take as many experiences that may be currently offline for their users and move that online, whether that's product registrations or warranty registrations or, you know, customer support, finding a dealer, finding parts for your product, for the products that you've purchased, you know, things like that.

Scott Wintheiser (2:09): So we're really working with largely manufacturers to try to improve that digital transformation process and make it as streamlined as we can for them.

Donna P. Mitchell (2:21): Thank you for that. So I remember reading some information that unlike some of the others in digital transformation, instead of always adding on, you will develop or do a continuous process from what is already there. Explain that to us, please.

Scott Wintheiser (2:37): Yeah, so along with continual improvement, one of the things we really focus on is right-sizing their ecosystem or their digital, what's often called a stack. What we often see, especially in manufacturing, is we have clients that come to us with, you know, they're larger, sophisticated businesses. They're very good at what they produce and what they design and engineer, but their digital maturity may not entirely be there. So oftentimes they try to compensate for that by having oversized software licenses.

Scott Wintheiser (3:09): So for example, the CMS that manages your website or the CRM that manages all the leads that are coming from your website. We find that they tend to assume that they need a very complicated, large-scale product because they're potentially a larger business that has, you know, heavy amounts of revenue. So they believe that they need that more, you know, fancier license. And what we do with them is figure out that, okay, you're not even using 80% of what you're paying for in those licenses.

Scott Wintheiser (3:40): So rather than continuing to pay for those licenses, let's put— redirect that money into actually creating better experiences for your customers, which is what, what you're trying to do, and less money on licensing that you don't need. And then over time, help them grow back into maybe a more robust platform.

Donna P. Mitchell (3:59): So when you say a better experience, can you expand on that?

Scott Wintheiser (4:02): Sure. So for example, if you are a manufacturer and you sell your products through a dealer, oftentimes people know that they want your product, you might be an industry leader, for example. But because you're not selling direct or through your website, oftentimes one of the pain points is for somebody to try to find the best place to go buy that product. And so that, that dealer locator type experience can often be less than ideal, partly because the list of dealers exists in some third-party system or some CRM that you have in-house or some system that might be, you know, 15, 20 years old.

Scott Wintheiser (4:43): And it probably changes constantly. And so we try to work with manufacturers, for example, to get that information online in a way that makes it very easy, not only for your potential customers to find, but also nowadays for, um, like answer engines. So AI search, for example, to make sure that if somebody's trying to find your product through AI, that they're actually getting the correct dealer information. I'd say that's, that's a typical example of where we often start with a manufacturer is that dealer locator type information.

Donna P. Mitchell (5:13): What size businesses are you actually working with in the customer base?

Scott Wintheiser (5:17): I mean, it depends on the product, but, you know, we have some heavy clients that sell worldwide, but their, their customer base may not be huge. So for example, they might work with municipalities or something like that. And their, their typical customer might be looking for a pretty large investment if they were to purchase. Otherwise, we have clients too that are selling, you know, products that are a few thousand dollars and they're selling them through dealers and they might have, you know, a few hundred dealers in the US, you know, something on that scale.

Donna P. Mitchell (5:48): So how do they end up really going the wrong direction? How do you end up helping them fix if they're really off track?

Scott Wintheiser (6:00): Sure. So the first thing we do is try to make sure that whatever information they're their customers are getting through the website is as accurate as possible. Because as many— most manufacturers know, you know, the bigger their dealer network is, the more often those dealer networks change, whether it's locations or contact information, or a dealer no longer sells their products, or 3 new dealers do sell their products. So making sure that wherever that data actually lives within their organization, that we're making sure we're exposing that on the website, fast and as correctly as possible, so that it's as current as possible.

Scott Wintheiser (6:37): So we're helping them streamline that flow of data, like trying to get the information because oftentimes, for example, what will happen is a client will come to us and all of that dealer information lives in an Excel file somewhere. Well, you can't really power a website with an Excel file. So streamlining that process of getting that information out of their whatever internal system they might have and onto their website is as quickly and accurately as possible on a regular basis, that's often a starting point for us.

Donna P. Mitchell (7:02): So once you say it's a starting point and you decide to move forward with the company, how does that work process develop? What type of company do you see? You start in one place in space and then you move to another?

Scott Wintheiser (7:17): Oh, sure. So often, like, we'll fix that problem, and then typically what we'll do is we'll manage the customer journey. So, you know, our clients' customers, what they're experiencing in terms of what they're looking for on the website, what they're finding, what they're not finding, and then we'll find the next thing that we might need to add. So dealer locator might be the first thing, but the second thing might be something like trying to find the right replacement parts for whatever, you know, whatever item you might own, or finding information about how to use the product better, or contacting customer support.

Scott Wintheiser (7:55): Because you have questions about the product or something like that. And so what we're doing is on a 90-day basis, we work with clients to identify what's the current pain points for their customers. Why are those pain points? And what can we do to, to make those pain points go away? And then we typically set, set out a plan on what we can update on the website, what we can improve maybe somewhere else, um, and try to fix those pain points. And then 90 days later, we do the same thing. We just try to create the idea that, that your website's never really done. You're just constantly improving it. So what needs to be done this quarter?

Scott Wintheiser (8:28): Fix it. What needs to be done next quarter? Fix it.

Donna P. Mitchell (8:31): So let me ask you this question. So you work with the websites, but it sounds like you start working with the infrastructure as well, correct?

Scott Wintheiser (8:38): A little bit. Yeah. We tend not to own it. Um, typically their in-house IT department would own that. You know, they're, um, but we certainly work with them to figure out, okay, what information can we be sending to your systems so that you, your team is as knowledgeable as possible? And what information can we be getting out of your systems so that your prospects and potential customers are as knowledgeable as possible? And then, you know, trying to complete that loop.

Donna P. Mitchell (9:04): So it sounds like in that manufacturing world, websites are really still important. A lot of people try to say websites are not important. You don't need them. But at the end of the day, there's still a lot of business in having a website as a business. Is that correct?

Scott Wintheiser (9:19): Oh, absolutely. And I think going forward, they're actually going to become more important, um, because people are relying on AI search so much. When people were doing Google searches in the past, you know, they would typically search for a few keywords at a time. Now they're asking full-on questions, like long, sometimes multiple sentence questions. On to try to help them find a product, for example, or a solution.

Donna P. Mitchell (9:44): Can you give us an example?

Scott Wintheiser (9:47): So a classic example I've heard oftentimes is like shoes. In the past, somebody would have searched for running shoes, which is a pretty simple search, right? Now they might search for something like best running shoes in cold weather, or best running shoes in cold weather for you know, people with low arches or something like that, much more detailed searches. And so what AI is doing, because AI can actually answer those questions, the way Google Search used to work is they would go out and they would just look for pages that had those keywords on 'em and they would assume that those were pretty good matches.

Scott Wintheiser (10:25): Now what they're doing is AI search actually does, it's what's called fan out. So when it searches, it does, it basically does a bunch of research for you. So it'll go out and it'll try to find the best shoes and then it'll go out and it'll figure out, okay, what are the things about low arch support that would, I would want to look for in shoes? And it would gather, it does all these, all of this research to develop a much more elaborate question or a much more elaborate answer for you. So instead of just saying, well, this site has a lot of information about shoes, it actually figures out what it is about shoes that would be good for your specific case.

Scott Wintheiser (11:02): And does a lot of research on multiple sites and then comes back with a cohesive answer for you. The reason that's important to manufacturers is that if your website isn't answering those questions explicitly, it's going— the search engines or AI search is going to find those answers somewhere else. And if it finds it somewhere else, you have no control over what it's answering and your product may not be included. So we're actually seeing that for manufacturers and for most companies, the need to get a significantly larger amount of information actually on your website.

Scott Wintheiser (11:36): It's not so much about making your website look great, or look, you know, it used to be all about branding, right? It was all about what does your website look like? Nobody really cared what was on it. Now it's the exact opposite. So obviously, you still have to get your brand across, and you still have to have a site that's enjoyable for a human. But you also have to be able to answer massive amounts of questions with a lot of product data, and that product data has to be accurate. And so a lot of clients, especially manufacturers, don't have that. Pricing is another great example.

Scott Wintheiser (12:10): So if, if you're a manufacturer that creates a piece of equipment, oftentimes one of the things people are going to ask first is, what does that equipment cost? What is this? What does this product cost? If your website's not answering that question, even if they search by brand name, right, they might say, what is, what does Brand X's equipment cost? If the company's website doesn't answer that question in some way, the AI search is just going to make it up. It's either going to do research, or it's essentially going to guess based on other information that it has on what your product costs.

Scott Wintheiser (12:41): And that's the answer it's going to return back to the user. So having accurate information, you need the answers that AI is providing so that you're controlling those answers more than just letting AI make it up.

Donna P. Mitchell (12:56): Okay. So for those that are listening, if you have a website and it has all the information on it and you've got the chatbot, the person is usually on the site to interact with the chatbot. Basically, you're talking about the information still needs to be on the site, whether the person is on that website or not, so it can still give the answer to AI. So when whoever's on whatever AI they on, it can reach out and figure out what's going on either, even in the Google search, in the URL.

Scott Wintheiser (13:22): Yep.

Donna P. Mitchell (13:22): So it's even more important to make sure you have more information on the site, not just when they get on the site and you got your little chatbot there. You need more than a chatbot.

Scott Wintheiser (13:29): Yeah. And even that chatbot on your own website is generally powered by the content from your website. So if people are asking, you know, very detailed questions about your own, through your own chatbot, that chatbot still has to get the answer from somewhere. And generally, it's from your own website. So if the information is not there, it's either not going to give an answer, or it's going to give an answer that may not be what you want it to be. And so going back to your original question of like, are websites important? They're significantly probably more important, but in a way that's different, because it's not where people are necessarily initially hearing about you or initially learning about you.

Scott Wintheiser (14:10): By the time they actually visit your website, they've already asked an AI search or Google search with AI overviews, a whole bunch of questions and gotten pointed in your direction. And so they're much more informed. And now when they're coming to your website, they're typically a much more qualified prospect in the first place. So they're usually farther down that buyer path or buyer journey by the time they even visit your website than they would've been, you know, 10 years ago.

Donna P. Mitchell (14:38): So it seems like you definitely improve a website or process improvement versus starting all over again. Is that correct?

Scott Wintheiser (14:49): Yeah. Yeah. So in the old days, you know, our clients would come to us and they needed a whole new website from start to finish. We would redesign the entire thing, often on a new platform or new CMS, you know, completely new look and feel and create a whole new website and we'd put it live. And then it would just slowly decay over time because it was kind of put up there and then the business gets busy and they don't put a lot of effort into it. Now, most of our, a lot of our clients actually come to us with a site that's okay. They have some buyer journey problems, you know, their customers can't find the right information or whatever.

Scott Wintheiser (15:25): And so we start with what they have and we're just going through that continual improvement of trying to make it better. Yeah. So it's like if you're buying a house, you don't have to tear the whole thing down. You can just, you know, start improving the kitchen and then start improving the siding or the roof or whatever. And then over time you have a whole nice, you know, finished product. Yeah. You don't have to gut it all together. Right.

Donna P. Mitchell (15:46): Okay. But is there, in today's world, they got the vibe coding, they have all these different sites, you got AI, are there some platforms that are more friendly to use and build on than others?

Scott Wintheiser (15:58): We think so. We think there's a couple platforms that we like. Umbraco, which is a lesser-known one in the US, is open source, much like WordPress, which is also open source. And the reason open source is so important is because the source code for the website is open, it's essentially available to the public, it's also available to AI. So if you're using AI to build or update your website, it's easier to use a platform where it has full knowledge of how the website works.

Scott Wintheiser (16:31): Whereas if you're using some more expensive proprietary platform, it may not know how to work with the software as well, because how the inner workings of that software are sort of hidden from it. And so we see a real big shift in moving towards open source platforms because they're so efficient and accurate for AI to work with. So in a lot of cases now, we can publish entire sections of a website without writing any code or even without logging into an admin and adding content.

Scott Wintheiser (17:02): You know, the AI is taking the content that was developed somewhere else and the designs that were developed somewhere else, and it's pushing all of that up into the— into your website now without having to have a developer or a content manager even touch the site directly.

Donna P. Mitchell (17:17): Is that like HTML or something like that?

Scott Wintheiser (17:21): Yeah. So it can do all the HTML, it can populate all the content into the database. It can do everything that a developer or a content manager in the past would've done. Um, the area that's a little different is, so, you know, whether they have us do that or they try to do that in-house, now the, the thing that's really, um, where people are spending more time is how do we do, how do we make sure AI isn't going off the rails and, you know, posting the wrong information or posting information that's not supposed to be public or any of that kind of stuff?

Scott Wintheiser (17:55): So—

Donna P. Mitchell (17:55): That's exactly where I was going. How do we make sure that they don't decide, hey, let's have a party, let's play, you know, Kumbaya or something like that together, and they go rogue?

Scott Wintheiser (18:07): Yeah. Yeah. And it's very easy, you know, when you say go rogue, it sounds very nefarious, but it doesn't have to be. Um, I think a lot of times AIs are trying to be helpful, you know, by expanding on your content or adding additional information or, you know, just thinking that it knows what you're trying to do. And maybe it doesn't have the full context, and so it shares a little too much information or it positions something slightly different than what you were hoping for. And so using AI to update your website is great.

Scott Wintheiser (18:38): And we certainly recommend it and would be happy to talk to anybody about it. But it's important to have guardrails in place, both from making sure that the models you're using are returning consistent results. So that once you get AI, say, connected to your website, what it does today might be different than what it does 3 weeks from now, because AI models keep getting updated. And so putting guardrails in place to make sure that you have predictable results and, uh, predictable workflows to make sure that, okay, if it does add content, that's great, but maybe it's, maybe there should be a human in the loop that is at least approving it and making sure that it's on brand and it's not sharing anything that's sensitive and that kind of stuff.

Donna P. Mitchell (19:24): So for those that are listening, what have you seen that concerns you? With AI, whether it was good or bad? It might have been something good that it tried to do, like you said, help, but it still kind of concerned you that it thought about doing it.

Scott Wintheiser (19:40): Yeah. Well, it's a good example. We've had— we know some cases with clients on their own systems where they're using AI, where it knows maybe a little bit too much about your business. And so if you're writing an article, it might start including details that maybe aren't ready for the public yet, especially if you sell a product or a, say a product that has, that maybe comes out with a new model every year.

Scott Wintheiser (20:12): It's very easy for the AI to maybe get confused and start sharing, you know, product highlights, for example, that are for next year's model. And it's already trying to include that in your article of like, you know, and don't forget to buy product X because it's, you know, got these 2 features, but those features in reality aren't ready for another 2 years. And so it'll start to get confused as to like, okay, we're working in this sandbox. Don't include the data that's over there yet. That's not ready for publication yet.

Donna P. Mitchell (20:42): But if you don't have someone reading before you publish, in your workflow, then it goes out is what you're saying. But if you have a human in the loop at that point before publishing and taking the next step, you might be able to catch it.

Scott Wintheiser (20:57): Yeah.

Donna P. Mitchell (20:57): But for those that are listening, it can happen automatically if you give them too much leeway.

Scott Wintheiser (21:01): Yeah, absolutely. And there's, there's ways to put those guardrails in even before human in the loop of just making sure that you understand the context of the information that your AI has before you ask it to start doing a task. So if your AI has access to competitor data or something like that, making sure that it doesn't confuse that with data about your own product, and so that it's not accidentally, you know, adding information about your product that isn't accurate.

Scott Wintheiser (21:36): And so making sure you're really containing that sandbox that it's playing in. It's a lot easier to control and making sure that it's not making up information that might sound great, but might not actually be accurate.

Donna P. Mitchell (21:51): So if you have folks listening now that are totally confused, they kind of know they need to put some guardrails in or a human in the loop, but they're not sure what, when, or if IT is doing the right thing and they're a CEO or COO. And where do you come in? Where does your company come in? How far do you play? In the sandbox, or do you throw that to IT and you're more on the business process improvement side?

Scott Wintheiser (22:17): Well, we typically get involved once it becomes, uh, once it gets to their marketing team. So our clients are generally the marketing team at a business. Um, we can advise them on internal flows of like, okay, it's cool that you're creating these blog posts about products or whatever. Here are some things that we're seeing with those blog posts that might be problematic and then helping them trace back maybe why that happened. It could be that they don't have a strong voice and tone set with their AI, so their AI is writing in a style that doesn't match their brand.

Scott Wintheiser (22:52): Hmm. It could be that they're not including enough information, so the writing that gets produced is sort of generic at best, or it could be that it's including too much information and then we're seeing like, okay, it keeps adding details about products that aren't accurate or it keeps adding, you know, whatever information to your, to your blog posts, for example, um, that are, don't seem relevant. They don't, they seem different from what you would've done in the past. And so helping them trace back where's that information coming from, from into their workflow, and then advising them on workflows once it gets to the website to make sure that it's being approved and there's still some human accountability for what gets pushed out onto your website.

Donna P. Mitchell (23:35): Because when you talk about the it, the AI, who's really responsible? Who's responsible for all of that being done correctly?

Scott Wintheiser (23:45): It depends. So we're typically responsible for it once it gets to your website. We can help you become responsible for it. Obviously, the marketing team is generally the ultimate person that's accountable for it. It's their website. It's not our website. We're helping you build it for you. And only you would know whether it's the correct information or not. So we make sure that there's a human in the loop during that publishing process. And then we help them figure out, at least identify who's accountable for the information before it gets to them.

Scott Wintheiser (24:18): So, you know, because our clients are marketing departments, we can only go so far into their organization as to helping them set accountability and and determining workflow. But we can certainly help them find some common tripping points or pitfalls.

Donna P. Mitchell (24:34): That's good to know. So, who do you normally work with? What size organizations?

Scott Wintheiser (24:42): We're general— our businesses are generally, our clients are generally between $15 and $100 million in revenue, sometimes much larger. Depends. Sometimes we'll have manufacturing clients that are, you know, in the billions in revenue. But their website needs are maybe not what you would think of for a, you know, billion-dollar company, they might be selling to a dealer market. And so their target audience is fairly smaller than you might think for a business that big. We're really trying to find clients that are, that value their website on a continual basis.

Scott Wintheiser (25:17): So manufacturers generally, but clients that are, that see the need and constantly removing pain points for their end customers. Whether they're a small business or a much larger business, you know, the process for us is pretty much the same.

Donna P. Mitchell (25:32): What are the pain points you're really seeing frequently today? What trends are you noticing?

Scott Wintheiser (25:36): Well, the big one right now is not for clients, not for end customers. So our clients' customers not being able to find the information they're looking for. People with AI search, they're getting used to the idea of getting accurate information very quickly. And so if that information is not on the website, you know, they're often going to a competitor. So those— that's one of the big pain points for our clients. And then another issue is often for our clients themselves, their pain points are, it takes too much effort to get this information up to the website on a regular basis.

Scott Wintheiser (26:09): So publishing new content or modifying content, that process is too painful for them because it's, there's a lot of hoops they have to jump through, or there's a lot of manual labor that needs to be done. And we help remove that so that they can actually improve that content workflow and make sure that they're getting as much content to their website as they can.

Donna P. Mitchell (26:31): Hmm. Yeah, it sounds like for some people it could be quite a bit, but at the end of the day, everybody has to be compliant sooner or later with the guardrails that are being discussed. Do you think AI can be controlled, or do you think based on what you know and have seen with websites, you have some concerns for the future? From your journey, from what you've seen, from who you talk to, I'm trying to really get a feel.

Donna P. Mitchell (27:03): And sometimes my clients and, and the audience and everybody just kind of want to know what's real and what's hyperbole and what are you really seeing and what do people really feel about what they're seeing, I guess is what I'm trying to share.

Scott Wintheiser (27:22): Sure. Yeah, I think, um, the one thing that we see with clients, for example, is they tend to underestimate what AI's already capable of. They sort of assume of like, well, that's too complicated. There's no way that AI could do that for me. Um, or it would take too long for us to, to teach it to. And the reality is that that's just not true. Right. Yeah, AI can get very smart about your business and your needs and your problems very quickly and very accurately.

Scott Wintheiser (27:53): What the risks are is, you know, at least in the world that we live in within, you know, the clients that we have and the marketing tools that we use and, you know, primarily building websites, I think the issue is assuming we have some clients or prospects that actually trust AI a little too much. So they think like, well, AI told, you know, we asked AI to build a whole new website for us and it did. And it pushes it out without, you know, they push that live without really understanding what it built because they trust the AI to have gotten it right.

Scott Wintheiser (28:24): And so they're not testing it. They're not, you know, looking for pitfalls that a human developer or human engineer would find because it looks good to them and they don't really understand what's behind that deliverable that was given to them by AI. So they don't really know what to test for and they just trust it and throw it out. But in reality—

Donna P. Mitchell (28:46): What should they be testing for? Don't leave us hanging. What should they be testing for? Can you get a little bit more specific?

Scott Wintheiser (28:52): Sure. So for example, if you were, if you were a non-engineer and you wanted to build a dealer locator, like we talked about earlier, you might have it, you know, you might vibe code something that looks like a dealer locator and it might work most of the time. But what you're not realizing is that the way it was built is filled with, uh, code that might have vulnerabilities in it, or it relies on data that isn't updated on a regular basis, or um, it's accepting, uh, uh, inputs that aren't, um, accurate.

Scott Wintheiser (29:29): You know, things that a human developer would test for because of their experience with building software. Claude can be very good, but if you're not an experienced developer, when you ask it to do something, you don't really always have the ability to know if it's right. It might look right, but it doesn't mean that it is right.

Donna P. Mitchell (29:48): So does that mean with vibe coding or some of the AI website development tools of today that you could be really opening yourself up to some malware or some type of hacking and— Potentially. Cybersecurity issues and not know it because it looked like it's done?

Scott Wintheiser (30:07): Absolutely. Yeah. And I think the bigger risk is, um, you're actually creating more what's typically called technical debt. So, as you continue to work on that product, like add new features or make it look different or whatever you wanna do with it, the bigger that that site gets, the more confused AI will get about it, and it'll start to include functionality that you might've forgot was even there. So behind the scenes, it's doing things like emailing data to somebody or saving something to a database, and you sort of forget all that stuff's happening in a bit of a black box, and there's not really anybody, any human that really understands what's happening in that black box anymore because AI built the whole thing.

Scott Wintheiser (30:52): And, um, you're sort of trusting what, what's inside of it.

Donna P. Mitchell (31:01): Okay. So now I got a question. So if you're using HTML, doesn't that kind of narrow it down a little bit more? With those points on HTML, that's a little bit different than, than the vibe coding and a lot of the other sites.

Scott Wintheiser (31:15): Sure. Yeah. Okay. So like if you're, if you're a small enough business and you're just trying to get what's essentially a web version of like a brochure on a webpage and get it out in the wild, AI is very good at that. And, you know, most likely you can look at it, you can read it, and you can make sure that everything on there is okay and you're probably fine. And in fact, for a very small business, I would recommend you do that because it's very inexpensive. If you get something that has a lot more of what we call moving pieces, you know, data that comes in or out from third-party sources.

Donna P. Mitchell (31:48): Yeah, yeah, yeah, yeah. Tell us, tell us about the moving pieces if you can and the tools that you like to use. Is that okay?

Scott Wintheiser (31:56): Sure. So, you know, with our contract manufacturers, most of the moving parts are what's called a PIM. So product information, which is typically a third party or an external database of all their product information. There's CRM. So there's obviously, or an ERP, which would have— would be an internal system that they're using to manage their customer data or prospect data. And then there's the website. And so which generally is built on a database.

Scott Wintheiser (32:29): And so those 3 work with each other. So in a perfect world, as somebody comes to your website, if they're a returning customer, you already know a bit about them on the website because that data's in the, in the CRM. And then they do something on the website or they buy a product or they download a form or they fill out a form or something. And then that information ends up in the CRM and back in the ERP. And, and those 3 systems are constantly updating each other. So that internally they know what's happening on the website, the website knows what's happening outside of the website, and your sales and marketing department know both of those parts as well.

Scott Wintheiser (33:03): So that you can start to reach out to the customers that are spending a lot of time in this area of your website and maybe haven't bought that product yet. So send your sales team over there and figure out why they haven't bought it yet.

Donna P. Mitchell (33:18): That's interesting. So what is it you haven't told us that we need to know? What did you want us to know for why you wanted to be on Pivoting to Technology Adoption? Because you shared some great information. I really haven't had anybody talk about websites. And, or, or after you, well, I do have one other question before you answer that. Um, once you're working with a customer, do you ever end up with a customer where you've done all this great website and you've given them everything they wanted and then they still don't use it?

Scott Wintheiser (33:47): Sure. Yeah, all the time. All the time. We, uh, not so much anymore because we're better at finding clients that value it as an ongoing investment and less of a one-time purchase. Um, but in the past, for sure, I mean, clients used to treat it as a, as a to-do list of like, we need a new website, we got a new website, check, we put it over there and then we forget about it. And then, you know, 3 years later they wonder whether or not their customers aren't using their websites or are complaining that they can't find information. Well, it's 'cause nobody's updated it in 2 years or they're not updating it enough.

Scott Wintheiser (34:21): That certainly happens less so now. We're better at finding clients that understand that before, um, we even start working with them.

Donna P. Mitchell (34:32): That's—

Scott Wintheiser (34:32): But going back to your other question or your last question, I guess, um, I think that if there was something I'd want your listeners to know, um, especially around your questions with AI is I wouldn't be afraid of it. I would definitely start, if you haven't already, I would definitely spend time trying to just play with it and ask it to help you on things that maybe you assume it's not gonna be able to and see where it goes. 'Cause we've always, we always see people do that.

Scott Wintheiser (35:04): They kind of go into it assuming that it's just not gonna be able to do what they want it to. And then there's always a line that gets crossed of like where it comes back with, uh, with a result and you're, and you're sort of shocked as to like, Oh, that's pretty incredible. And then it's down the rabbit hole they go and it's, you know, now you can't get 'em to stop using AI. Um, so I would definitely recommend that. And, um, and don't be afraid of it. You know, also don't trust it, it just out of the box. You know, make sure you're always thoroughly vetting what it's creating because the more we see people use AI, the bigger the output that they're generating is.

Scott Wintheiser (35:43): So it might be instead of a 1-page website, now it's 10 pages for their website, but they may not be spending enough time making sure that those 10 pages are accurate because once they get to a certain trust level, then they just start trusting everything that comes out of it because it looks good enough. So you still have to vet it. You still have to, you know, do quality control for sure.

Donna P. Mitchell (36:02): What do you think everyone needs to stop doing? Trusting it?

Scott Wintheiser (36:11): Yeah, I guess so. I guess that would be it. Yeah. Trusting it without reading it.

Donna P. Mitchell (36:15): Yeah. I think there needs to be pushback or you need to question, why did you say this? This is this, or— Right. I think there needs to be a conversation back and forth, even if like they were live versus just rolling with it. Otherwise it hallucinates. It starts giving information that's not accurate. Then you gotta ask it, is this accurate or are you hallucinating? Or are these factual? What about my integrity? And then it'll come back, oh, I did put this together. Yeah.

Scott Wintheiser (36:45): Yeah, exactly. Or, or it's including something that you may have, like information from a different chat you may have had with it that was about something similar, but it's a little different. And it starts bringing in information that really shouldn't be included in, in your current context. And so just trusting what it's doing. I think is a mistake for sure. And it's something we see a lot.

Donna P. Mitchell (37:07): Boy, this is really a nice conversation about websites and everything with AI. And I'm really so glad we met, Scott. And I'm hoping you come back with some additional information that we need to know about websites, because it sounds like websites are the thumbs up right now instead of a thumbs down.

Scott Wintheiser (37:30): Sure, absolutely. I think it's— I think we're back. You know, there was a time where it was less important, and I think now we're going back the other way where it's more important than ever.

Donna P. Mitchell (37:41): How do people reach you?

Scott Wintheiser (37:41): You can find me on LinkedIn. I'm probably the only Scott Wintheiser there, or on our website, lightburn.co. So that's L-I-G-H-T-B-U-R-N.co.

Donna P. Mitchell (37:49): Well, thank you so much for sharing all the information, and thank you for listening to Pivoting to Technology Adoption, because we are shaping tomorrow together. Thank you. Thank you.

Scott Wintheiser (38:00): Thanks for having me. It was a good time.

Donna P. Mitchell (38:00): Thank you for listening to Pivoting to Technology Adoption. If this episode gave you something to think about, subscribe, leave a review, and share it with a colleague who needs to hear it. My new book, Pivoting to Technology Adoption: Mind the Gap, is available now on Amazon. Visit pivotingportal.com to work with me directly. I'm Donna P. Mitchell, The Transformation Authority™. Until next time, mind the gap.

ENDNOTES AND SOURCES
[1] Figures Scott Wintheiser mentions in this conversation, including how much of their software licenses companies use and his clients' revenue range, are his own estimates from Lightburn's client work. They are not independently sourced.

GUEST
Scott Wintheiser
Scott Wintheiser is CEO and co-founder of Lightburn, a Milwaukee-based agency he co-founded in 1998. Lightburn works primarily with manufacturers on continual improvement of their websites, digital experiences, and buyer journeys, moving offline customer experiences online and right-sizing the technology behind them. Lightburn is an Umbraco Gold Partner.
Website: lightburn.co
LinkedIn: linkedin.com/in/scottwintheiser