An AV and integration-focused podcast broadcast weekly

This week's episode of Automation Unplugged features Mike Wilson, Co-Founder of Bethesda Systems. Mike last joined us on episode 306 in May of 2025, so if you want the full story on Mike and the company he's built go back and give that episode a listen. Today, we're zooming right in on AI.

In this episode, Mike and I discussed:

• How AI image editing helps his team solve real client challenges — from removing private artwork and family photos before sharing project images, to swapping out unlicensed content on screens.
• How he uses simple ChatGPT prompts to turn a designer's scale elevation drawing into side-by-side visuals, helping clients make decisions faster and reducing costly change-orders after installation.
• And how Bethesda Systems is building an AI culture from the ground up, including weekly leadership meetings with hands-on AI demos and shared licenses across the team.

What makes this conversation so valuable is that Mike doesn't call himself an AI expert — and that's exactly the point. He's a business owner who started with simple experiments and turned them into practical tools that improve margins and client communication.

Whether you're already experimenting with AI or just trying to understand where to begin, you're going to get a lot of value from this conversation.

So settle in and enjoy my conversation with Mike Wilson. Let's get started!

SEE ALSO: #362: The Power of Vulnerability: Jason Zuniga on Peer Groups and Second Chances

Transcript 

Ron:

Hello, hello there. Ron Callis here with another episode of Automation Unplugged. Today we have a very special AI Insights version of Automation Unplugged, and we have a returning guest who has agreed to come on and share his perspectives and his journey in AI — the one and only Mike Wilson of Bethesda Systems.

Ron:

Mike was last on the show back in May 2025, just about a year ago, on show 306, so be sure to check out the full version if you want a more in-depth background on Mike, Bethesda Systems, and his journey there. For this show, we're going to zoom right in on the specifics around all things AI for Mike. So let's jump in and say hi to Mike. Mike, how are you, sir?

Mike:

Good morning, good afternoon, and good evening, depending on when you're listening. How are you?

Ron:

Amen. Out of curiosity, Mike, when do you find that you listen to or watch podcasts? Do you do it on morning walks, or on drives to job sites? Where do you consume content?

Mike:

Between 4:45 and 6:45, when the dog would like an afternoon walk. It's a great time filler for that purpose, so that would be my podcast listening time.

Ron:

Is podcast listening something you try to do all the time, or just sometimes? How do you think about it?

Mike:

I wouldn't consider myself a dedicated podcast listener as much as a specific podcast listener when I'm really interested in the content. Your podcast is definitely on my radar — I tend to listen to more than my fair share of yours. I have friends who have been on, people who interest me who have been on your podcast. It has been great for getting insight into the industry. And when I find a particular topic, I tend to listen to it. I like some movie podcasts, some sports podcasts. Podcasting is a great medium — it's just a matter of finding your interests, and there are so many out there.

Ron:

Amen, I agree. Now, for those who aren't familiar with you, Mike, and somehow live under a rock and don't know who Bethesda Systems is — not meaning to insult anybody out there — Bethesda Systems is a great operation that has been around for years. Mike, give us a quick nickel tour of yourself and Bethesda Systems, and then we'll jump right into it.

Mike:

Sure. Mike Wilson. I started Bethesda Systems with a business partner, John, in 2003. We're located in Bethesda, Maryland, in what I would call the DC, Maryland, Virginia corridor. We now do full high voltage and full low voltage. We specialize in working with homeowners to deliver what we call the full turnkey solution — from design to purchase to install to implementation to programming, and to one of the most important aspects, aftercare service.

Ron:

Okay, awesome. And folks, remember, if you want to learn more about Mike and Bethesda Systems, you don't have to go back too far — episode 306. Mike was kind enough to join me on that episode back in May of 2025, and we had a great conversation.

Ron:

Mike, are you willing to share what your journey in AI has been like? I'll preface this: ChatGPT came on the scene in Q4 of 2022, so in 2023 the world was starting to use this thing called ChatGPT at an accelerated rate. When this show releases, we're in mid-2026 — so what has occurred between then and now?

Mike:

Sure. Let me start by saying, as we go through this, I definitely don't consider myself an AI expert. I'm merely excited to offer my insight on some nuances that I find are helpful and work for people, but I don't want anyone to think I'm what someone would consider an expert. What's important, and why I wanted to come on here, is that AI is great for people who are just beginning, or non-experts, and it's a great way to get more excited about it.

Mike:

That being said, my first real understanding of AI — besides my children telling me that's how they're going to get their way through school — came when I attended a trade show buyers group conference, and there was this guy Ron who did a quick overview of how AI worked. That was my first true understanding of it. Then a couple of friends of mine in the industry who are really good at it started giving me different tips and tricks, and I've gotten to the point now where we use AI not just for returning emails, but for all sorts of stuff.

Mike:

And I can't stress this enough: if you're enjoying this podcast, continue to watch this whole series, because I use maybe 3% of what this whole AI world can do. I'm just hoping someone will get the bug, get an understanding of it, and then take it and run.

Ron:

Well, I love that that conference presentation was helpful. I want to say that was back in the spring of '23 or spring of '24.

Mike:

I don't remember. The only thing I'll say — and we could go into a whole different episode on this — is there's this thing called a Plaud, which is what you convinced me to get. It's a game changer, for a different podcast, but I just want to let you know I do have it on my desk.

Ron:

I'm proud of you. I love it. Here, I'll show you mine — I've got the Plaud pin. You're holding the first version of Plaud — I think it's called the Plaud Pro. My wife inherited that one. My son is a junior, and we're having all these conversations about college and applications, so my wife now uses it regularly to record those conversations and process them to help us through the college application process.

Ron:

All right, let's get into the nitty-gritty. We promised our audience that these episodes are going to be tight and to the point. So share one of the applications where you're finding a fun and helpful use of AI in your business and in your life. Let's present the problem, and then we'll share our screen and show how you work through it.

Mike:

Sure. We picked two quick things to review today. The first one is what I call how to solve the displaying-the-picture problem. Oftentimes we take pictures and we want to use them for social media, or to show a client a different example, and there are often parameters that prevent us from using those pictures — whether it's a family photo in the image, a certain piece of artwork, or a family memento — something in the space that the homeowner doesn't want other people to know exists.

Mike:

So for the first case, if we jump onto screen share, what you'll see is an image. I just took a regular image for this purpose, and what I want to show everyone is that there's this flowered piece of art over the fireplace.

Ron:

And let's be as descriptive as possible, Mike. We have folks who are on their morning walk, or their afternoon walk with their dog, and they don't see the image.

Mike:

I'm showing an image of a family room that probably has too much furniture — but that's a different discussion — and over the fireplace is the focal point of the room: what looks like a gold-framed piece of art with flowers. In this scenario, the homeowner told us that under no circumstances can anyone know they own this piece of art, but I want to show this picture to another client.

Mike:

Here's what I'm able to do — and again, this is very simple. We're going to use ChatGPT for this. There's Claude, there's Gemini, there's Copilot — there are a million different options — but right now we're using the simplest form so people can get involved. We have a simple chat window. We take the picture, put it inside the prompt, and say something kind of fun. For this scenario: “Let's replace the image in the picture with dogs playing poker.” The AI will analyze the image, analyze our request, and then create a new image for us.

Ron:

While we're waiting for it to process — and we are doing this live, folks, so if you watch on YouTube, you'll see it happening right before your eyes — do you have a preference between ChatGPT, Gemini, or Claude? What do you use, and when?

Mike:

I actually do. I tend to like Gemini a lot for more in-depth picture work. I like Claude when I'm doing some type of scope — I really like putting our proposals into that. And ChatGPT I like when I just need something quick and want to get it done.

Mike:

As we come back, you can see it took the piece of art and replaced it with dogs playing poker. We'll zoom in so you can see it a little bigger — and voilà. Now let's go a step further: I want to remove that plant on the table. For whatever reason — in this analogy, we just want to remove the plant on the table.

Ron:

Yeah. You want to further refine the image — move a plant, move a piece of furniture. You want to tweak the image before you publish it to your website or social media.

Mike:

So I'm going to do a little something funny here. I'm going to say, “Remove the flower pot on the coffee table.” I don't know if it's a plant or a flower pot, but I want to use the term “flower pot” because I want you to see that the AI engine will normally do a very good job of understanding what we mean, as long as we use the terms “coffee table” and “remove.”

Mike:

While that runs, just to finish the earlier question — there are some unbelievable things you can do with Chat. You can create your own tools, you can have things built for your company. But everything we're doing here is the simplest form: just open up the prompt and go. And there we go — I'll zoom the image in for you, and now you can see the flower pot, or plant, has been removed from the coffee table. The dogs playing poker are still up there, and now I can present this image to my client.

Ron:

I'm going to pick on it — did it add the candles to the table?

Mike:

I don't believe so. We can double-check.

Ron:

Oh no, they were already there. You're right — it just deleted the one item. That's great. So Mike, are you regularly tweaking images as needed to help with your social media content? Is that a typical application for you?

Mike:

Yes. Another good example of tweaking an image is an image on a screen that you don't have licensing for. If you want to show a generic football photo instead of something from the NFL, or a movie image you don't have access to, we do it for that. Or the fireplace — if you want to make it look like the fireplace is lit, or make it look darker outside. There are a lot of tweaks you can do to photos with this, and it's just the tip of the iceberg. I encourage everyone to start to play with it.

Mike:

I'm sure we'll see comments about what I could have done differently or how you can do it differently, and that's totally okay. I purposely picked this for the podcast so people can see: let's take some simple, silly stuff, start to massage it, and see where we get.

Ron:

I love it. Now, you've prepped a second example.

Mike:

Yes, I did. The second example is one I feel we use a lot. I'm showing you a scaled elevation image of a television. This is something a designer will normally draw up. They have no idea what TV you want to use, so they build a scale elevation — which is what you see in this image — and put a random television in there. There are scale dimensions on it, there's a fireplace at the bottom, and on top is basically a rectangle that looks like a 16-by-9 television. I don't know what size it is.

Ron:

Okay.

Mike:

So now we take this image and go into our prompt. This homeowner — I don't know if it's the husband or the wife or the two husbands, whatever it is — they really like big. They want to get an idea of what an 85-inch TV would look like over this fireplace.

Mike:

So I'm going to enter into the prompt: “I've attached a scale elevation drawing of a TV over a fireplace. I would like to replace that TV with a larger one. I would like to use a Sony K85XR70. Can you please show me this updated drawing to scale with this model?” Now it will analyze that and create the image for me.

Ron:

While that's rendering — how do you think about the use of AI at Bethesda, Mike? Do you operate on a team license on any platform, or is it individual personal use? What's the across-the-board use case of AI, or do you have a plan yet?

Mike:

I think the plan is always evolving. Right now we haven't established a full team. I share a license with my assistant, I believe my two project managers share one, and we're always learning. As you go more in depth, you can understand how to build pre-designed prompts so you get consistent results. We can go to much higher levels with this stuff, and I'm excited to continue. We have a leadership meeting every week, and we try to dedicate one a month to a 20- or 30-minute demo on different things Chat can do.

Ron:

It's a great way to share. All right, what are we seeing on screen?

Mike:

What we're seeing now is a much bigger television. It's a Sony, and you can see in the image how much bigger the television looks. Now you get a better idea of what it would look like over the fireplace, and the client can see the size of the television and say, “Okay, now I understand.”

Mike:

We'll do one more thing, and hopefully it works the way we want. I'm going to say, “That's great, but can you also show me an option with a smaller TV? I'd like to use a Sony K55XR70, and can we do it to scale for the smaller television?” We type that prompt in, and off it goes.

Ron:

In your experience, Mike, does showing a client or prospective client a visual aid, such as you're creating here in Chat, help the sale — help the client make a decision more quickly?

Mike:

I think it helps on two ends. One, it helps the client make a decision more quickly. But even more important, it creates less concern after the fact. There's less discussion of “I didn't understand,” less “Oh, that's not where I thought it would be.” What's really important is we all try to work on good margins. We're all in this business to make money, but it's a competitive industry. One of the advantages you can gain with this simple use of Chat is a way to be more efficient. One of the best ways to generate more revenue is to be more profitable, and one of the ways to be more profitable is not to rehang something or take something back.

Mike:

The image came back, and as you can see, the TV is centered in the space. We can get more advanced and ask it to lower it or move it, but now you can see the difference between the 55-inch and the 85-inch. You can print these two images — a larger version and a smaller version — bring them to your client and say, “Look, this is roughly to scale what the difference between those two TVs would look like.”

Ron:

Love it. I love this example — I think it's easy and fun for people to go and immediately test. I want to wrap us up here, and I'd love for you to share, from your perspective: what can people out there do to accelerate their use of AI, or the benefit of it in their life or their business?

Mike:

This is my favorite thing to tell people: if you want to learn about AI, ask AI to teach you about AI.

Ron:

What do you mean?

Mike:

A great example — if you went into Chat and really wanted to do this, ask it: “Hi Chat, this is what I'm trying to do. I'm going to upload an image for you, and I want to show three different TV sizes. Can you help me by telling me what I need to tell you to do that?” And it will do it. If you ask it, “What is the best way to show an image to a client?” it will do it. There's so much trial and error in AI that you learn and come up with your own unique approaches.

Mike:

There are all sorts of things people do: put plans in Chat and ask it to count the number of lights. Put your proposal into Chat and ask it where there are opportunities to save cost. Put a proposal into Chat and ask it to write a scope. I showed you the Plaud — record a meeting and ask AI to do a summary with bullet points. We could go on forever about all the stuff AI can do.

Mike:

I know what we did seemed very simple and very rudimentary, especially to people who are more skilled in AI, but the point is you have to start somewhere. If you're sitting on a tennis court just bouncing a ball up and down, you may look silly if Andre Agassi is next to you serving. However, he started that way too. I don't want anyone to be intimidated or afraid to start by doing something rudimentary and silly, because that's how you get to the next level.

Ron:

Amen. Mike, for people who want to get in touch with you — where can we send them?

Mike:

I'm on all the socials: @BethesdaSystems on Facebook, LinkedIn, and Instagram. I have a wonderful website provided by One Firefly — shameless plug. Feel free to reach out if you have any questions, and I wish everyone a lot of luck. I promise you, if you just do one little thing every week — just something simple — I guarantee you can monetize it in some way.

Ron:

Love it. Mike, thank you for joining me on AI Insights. This was awesome.

Mike:

Thank you very much. Take care. Good luck, everybody.

Ron Callis is the CEO of One Firefly, LLC, a digital marketing agency based out of South Florida and creator of Automation Unplugged. Founded in 2007, One Firefly has quickly became the leading marketing firm specializing in the integrated technology and security space. The One Firefly team work hard to create innovative solutions to help Integrators boost their online presence, such as the elite website solution, Mercury Pro.

Resources and links from the interview:

Show Notes:

https://www.bethesdasystems.com/

https://onefirefly.com/expertise/podcast/item/au-306-michael-wilson-on-the-business-of-high-end-integration