Artificial intelligence adoption in the flooring industry has accelerated dramatically over the past several years, evolving from an optional business tool to what many now consider a competitive necessity. Yet implementation remains uneven, with many dealers still hesitant to embrace the technology. 

Recent developments highlight both the potential and the concerns surrounding AI in flooring operations. Maksim Nazarchuk, owner of Modern Estimates, has developed an AI-powered estimating service that can calculate square footage from photos and generate estimates within minutes. The technology represents a broader shift toward automation in an industry traditionally reliant on manual processes.

However, the rapid advancement of AI tools has also intensified questions about data security, privacy protections, and regulatory oversight—issues that continue to influence adoption rates across the sector. 

In a recent interview, Nazarchuk addressed common industry concerns about AI implementation, outlined current advantages and limitations of the technology, and discussed existing government regulations governing data privacy. The following excerpts explore these key issues facing flooring professionals today. The interview has been edited slightly for length and clarity. 

Floor Trends & Installation: There's a lot of fear surrounding AI and implementing it into flooring retail businesses. Let's talk about that. 

Maxim Nazarchuk: One of the big ones that keeps coming up is: “My work is being stolen from me. I went to school. I've built a career, and this thing comes along known as AI and it can do things quicker, faster. I can't possibly hope to compete with it. When it uses the data that it's given, including some of my own data that I've put out on the Internet, it doesn't give credit back to me for that data and so thereby it takes my job without any hope on my part to ever get credit for it and to earn money from that.” [That’s the] wrong way to look at it. 

AI doesn't take anyone’s individual data. It takes everything in aggregate. If you can imagine every single data point out there—books, videos, photos, everything you can imagine that explains the world around you. We tend to look at things individually as human beings. AI looks at things in aggregate to understand its surroundings. So, it doesn't see individual people. I know that when experts start explaining things in this way, it only gets scarier: “Oh, so I'm not even a human being?” It's very tempting to understand it that way, but really what it's doing is it's trying to understand its purpose. Most AIs out there right now are purpose driven. They don't have general intelligence. 

We have LLMs [Large Language Models] and different models, and they're the ones that understand language. Those are the most popular ones, and those are your Google Gemini, OpenAI, ChatGPT and the like. Then, you have video recognition, photo recognition and things like this. Most of the AI out there are purpose driven. 

AI is not going to take people's jobs, not anytime soon or in the near future. It's not going to take a flooring retailer's sales associate job or an estimator job in the sense that your employer or your client is not going to be able to replace a human. There are a lot of things that human beings do that AI cannot do. It is there to help essentially. 

AI is not a new concept. The theory itself was envisioned in the early 1900s. If you think about it, it's inevitable. As our world becomes more and more complicated, people have only so much capacity for attention on any given task. And as much as some of us like to pretend that we can multitask, what the study shows is you can switch quickly between tasks. You're not really multitasking. The human brain isn't capable of doing multiple unique things at once, cognitively. There's a lot of things going on in the background, but cognitively, you're not really able to do that. And as our world becomes more and more complicated with more devices, more information, more technology, more things to do, the capabilities that technology gives human beings to do things, even for entertainment and leisure, it becomes more and more complicated and difficult to keep track of it all. [AI] enhances our ability to process more data. Think of it as an extra engine in a car.  

Floor Trends & Installation: What are some of the advantages to implementing AI into a business? 

Nazarchuk: AI is really good at clarifying data and tracking data and following the conversations. For example, let's take estimating in the flooring industry. Projects range from anything that's super simple; you just come in, pull up a floor in a small bathroom, and you're dropping down vinyl plank and there's nothing much else that you're really doing. It's a simple job. And it ranges to the most complicated layouts with thoughts and some actual plans and some spit balling. You can write all of that out in a note as you're going. As we've discussed, people are not good at multitasking. So, as you're writing, you're losing the momentum of the project, especially if it's a more creative project where you're doing inlaid hardwood and you're doing elaborate designs and you're trying to create a certain atmosphere. 

Instead of writing, you can talk to your phone essentially as a recorder, but much more useful than the recorder, [AI] will understand what is a thought, what is a plan and organize that for you. Then, when you look back, you can clearly see the direction you're going without having to guess what you meant by something. It can organize those things for you. 

Additionally, it's essentially a consultant. Having recorded that [information], you can talk to it, and you can say, "Okay, what are some other things that I've missed that I didn't consider? Can you look up the standard for the National Wood Flooring Association or the tile standards?

Is this in compliance with that?” Instead of doing hours of research, it can quickly find that information for you. And you can even skip ahead and say, “Look up these standards and correct the plan accordingly.” Within seconds, you have that information. So, it expedites the process, and I suppose creates a much faster world, which I know can be scary for some people, but the alternative is we become hopelessly lost. There's so much information, so much data out there. 

You can sit for hours on the Internet and find a bunch of articles, and how do you know how many of them are real? Just trying to find the relevant information, you have a bag of needles and you're trying to find a specific one is what the Internet has become. With a lot of useless data, a lot of useful data, a lot of fake information, how do you sort through all of it? And how do you compare and contrast between different articles, different data points? It's become hopeless for most people, unless you're a big corporation and you have hundreds of staff. But most of us don't have that, right? This [offers the] ability to level the playing field with a single person being able to do a task of 20 or 30 people in a matter of minutes. 

It's not at the level to compete with elaborate human out-of-the-box thinking. But if all you need is a simple logo or a website layout and you don't even know where to start, but you have this business idea and you don't have any money, it's a fantastic start.

Floor Trends & Installation: What are some of the disadvantages? 

Nazarchuk: As with everything, there are tradeoffs, and AI is no exception, especially in the stages we’re at. A lot of it stems from the clarity of the prompt or the question and the data you're giving it. But one of the disadvantages that we see is called hallucination. Where it's giving you an answer and the answer seems right, but when you research that answer or you ask it how it came to that answer, why is this, where is it getting this fact or data point, it turns out it's made up. But what we see with that is over 80% percent [of] it is how the question is phrased. I think as people get more used to working with AI, they will learn how to ask it questions. 

The difficulty that most people are having is that they don't understand how AI works. [Being ambiguous] or implying things without really saying it, oftentimes will cause these hallucinations. With AI, you need to be clear what you want and direct. It's a bit of a challenge to a lot of us out there where we're used to people picking up on social cues and facial expressions, hand gestures. AI doesn't have that ability. It's reading your prompt, and it's trying to guess what

you mean by that and what answer you're looking for. It's essentially predicting what you want to hear. Obviously, you want to hear a correct answer, but it's using your word choice based on that. When you give it a prompt, you may need to play with it a couple of times before it gets to the correct answer. 

One thing that I would suggest for people [to do] is ask for sources, especially if you're doing research or you're writing some sort of article or if in your own business, you're trying to do finances or anything like that. Ask it what sources it used. Always put the prompt “include sources,” especially when you're looking for factual data. It will cite every article for you, what it used to get that data. There's still some legwork involved. You do have to go in and read those articles and judge for yourself if [the info] is accurate. 

Just like with any technology, there's a learning curve for the speed that it brings to completing tasks, writing an article or sometimes writing an email. As with most technologies, there's a starting point to where they're hoping it will get to, which is, mass acceptance. But there's a certain point where technology takes on a life of its own, and it just does its thing. And AI is one of those things. 

Floor Trends & Installation: Can you explain how AI-powered estimating systems process customer photos while protecting their privacy? 

Nazarchuk: There are a lot of government requirements and compliances that are indirect. Governments are still trying to catch up to specific AI regulations; some are needed and anybody in the industry will tell you that some aren't needed because the market kind of self-regulates in what businesses are trying to avoid. One of those things is privacy—geolocations and personal data. 

A lot of people don't know that when you take a photo on your phone, there is metadata. Essentially, it's data in the background. What the photo stores is your precise location in most cases. Even when you have geolocation off on your phone, your pictures still have that. And if somebody knows what they're doing, if you just upload a photo, share via email, without scrubbing that data, they can pull that photo and take a look and see where the photo was taken, time of day, elevation, and depending on your device, what type of device it was. It's very extensive. Because there are regulations around precise locations and data privacy, companies want to avoid that liability by creating processes—and Modern Estimates and Measure is no exception to that—by scrubbing that data.

Every photo that person takes, it takes a couple seconds for it to upload. The reason that a couple seconds is there is not just because it takes time to upload or [to make a] connection, but also there are processes that are happening in the background. Those processes are scrubbing what's called personal identifiable information. That would be registration of the device, the actual ID of the device, the location, the geolocation of where the photo was taken, all of that. So, the only thing that remains is the image itself, and the thing that the image is comprised of. That's what's actually fed into a model. 

In our case, when there's a photo with a human being [in it], it literally can't see that because every single space of the photo is converted to a number. When all those numbers are plugged into a big calculator—that's the purpose of the model—it gives you an answer. The answer will always be purpose-driven because my face and me in the photo has nothing to do with the purpose of the AI. It does not perceive me as part of the data set. Instead, it actually ignores me and looks at the things behind me, which actually solves some other problems when it comes to flooring estimating where if you have a rug on a hard surface, that's not a carpet, it's a rug. 

How do you calculate underneath it? How do you have it account for the space? Most people would think, well, you're just using the walls as your perimeter. No, that's called LIDAR. AI actually uses the space itself to determine its size and having a rug or having a big cabinet against the wall can throw it off if it's not trained to ignore it. 

The way [it thinks] isn't actually how most people will think. You don't tell it to ignore human faces or ignore area rugs. What happens is you give it thousands of examples of flooring and some of those areas will have rugs and some of them will not. Then, it looks at the answers you gave it. You gave it a data set, which would be comprised of photos and the explanation of the relevant photos. It looks at the data set. It looks at it in aggregate and of the thousands of photos that I've seen, what is the common theme here? Some had area rugs. All of them have floors and all of them ignore this rug. When it's trained and you give it examples, it knows that the rug was always ignored; that it is irrelevant. I ignore it. I pretend that it's continuous. Same thing with information that AI shouldn't use. 

Floor Trends & Installation: What is important for users to know when it comes to utilizing AI tools? 

Nazarchuk: Asking for sources is a shortcut to verifying the information. But the number one thing is being clear with what you're asking. Boil down your question to what is it that you actually want to know. Don't talk about your question. Unfortunately, we put a lot of fillers in. So, just being clear with your question is number one. 

If you're clear, then you'll save yourself a lot of trouble. I suppose that's true though for any conversation, right? If we talk to human beings and we're not clear about our intentions, our desire, our purpose, then the response we get from a human being is akin to AI; it is a hallucination, except with real consequences. 

I look at it as good. It's challenging us to be clearer. We have no choice but to be clear. It's here to stay. There's so much good and so much potential that it brings to human beings, business owners and our professional and personal lives that is unimaginable. Evil will always exist, unfortunately. We shouldn't let evil stop us from progressing and making tomorrow better and easier than it was yesterday.