Mike Downer: Hello, everybody. Welcome to The Unlearning Lab. I’ve got Kevin over here adjusting his camera. I am your host, Mike Downer. See, he’s got to do everything he can to make himself look good. Keep going, Kevin. You’ve got a lot more work to do.
Again, I’m Mike Downer, your host. This is Kevin Wosmansky, founder, CEO, and head dog groomer at JAR Consulting Group. How are we doing today, Kevin?
Kevin Wosmansky: Good. I’m glad you threw in “head dog groomer.”
Actually, we just did a big social media post. I guess last week was National Dog Day, so our social media manager did a post yesterday where everybody in our company had to submit pictures of their dogs. It was kind of cool. I’d never seen that before. I thought it was a pretty neat little thing she did.
Mike Downer: That’s pretty awesome. I guess my timing is impeccable.
Actually, Jonas, our podcast production manager who does all the work behind the scenes, has four dogs.
Kevin Wosmansky: Oh my gosh.
Mike Downer: So, it’s kind of a cool little thing to see. Do you have a dog?
Kevin Wosmansky: Yes, I have one, and that’s enough.
Mike Downer: Jonas has four dogs, three cats, a turtle, and a fiancée.
All right, let’s jump into it, bud. What do you have for me today?
Kevin Wosmansky: Let’s talk about what ChatGPT thinks about your business.
I know it sounds strange because people think, “It’s a computer. What can a computer think about me?”
Mike Downer: My question for you is this: Can businesses actually measure their visibility in ChatGPT, Gemini, Perplexity, and other AI platforms?
Kevin Wosmansky: The answer is yes. Businesses can measure their visibility in ChatGPT, Gemini, and other AI platforms, but they shouldn’t treat it exactly like traditional keyword rankings.
SEMrush, Ahrefs, and several other great software platforms allow you to track keywords and similar information. When you’re measuring your visibility in AI, however, you can measure things such as mentions, recommendations, citations, and competitor visibility.
One of the most important things you can measure is sentiment. What is the sentiment surrounding your business? What does that large language model think about your business?
You also need multiple prompts because one question doesn’t necessarily represent everything your business does throughout the entire customer journey.
You have to think beyond a keyword search such as “best plumber near me.” People are prompting their AI tools and having conversations with them. They’re saying things like, “I’m looking for a great plumber near me who is family-owned and has been in business for 20 years.”
It becomes an interactive conversation.
What I tell clients is that AI visibility is measurable, but you have to measure the questions or prompts your customers are actually asking.
Mike Downer: I completely agree. You’ve completely changed the way I interact with these tools. Everything I’ve done since becoming involved with JAR Consulting Group and the AI learning process has caused me to use them differently.
When I’m out and about, I hear people talking to AI as if it’s a person or, as you like to say, their own personal concierge.
If you’re trying to become the person, business, or entity that AI recommends—the business someone’s concierge recommends—you probably have to examine several things about your company.
Can you tell me exactly what an AI visibility audit is?
Kevin Wosmansky: An AI visibility audit is essentially a way to establish a baseline for your business.
There are different levels of AI visibility audits, and any business owner can perform a simple version.
The most basic starting point is to go to ChatGPT or Gemini and ask, “What do you think about my business?” or “Would you recommend my business?” Ask what the AI can see.
You can also enter your website URL and say, “I want you to look at my website and tell me what you think about my business.”
That’s a very simple audit that any business owner can perform.
There are more advanced levels, of course. A whole new cottage industry of software has been developed to track AI visibility.
At JAR Consulting Group, we use several sophisticated software platforms. When we bring on a client and conduct an AI audit, we use three different large language models.
We put the information together and have the models audit the business. We identify the most commonly asked prompts related to that industry, vertical, or business.
We then take the three independent sets of results and synthesize them using a fourth large language model. That gives us a strong combination of insights.
Once we identify the top 25 to 50 prompts, we separate them into branded and non-branded prompts.
Branded prompts include the business’s name. Non-branded prompts are general questions people ask without mentioning a specific company.
We then use sophisticated software to track all of those prompts across multiple large language models.
This is where it gets interesting, Mike. On any given day, we can see what each of those large language models said in response to a particular prompt.
We can track and audit how deeply a business appears in those responses and see what the large language model told someone who entered that prompt.
A tremendous amount of information comes from that process. We use it to begin helping clients get cited and recommended.
Those are the two sides of the AI visibility audit coin. One is very simple and can be performed by any business owner right now. The second is significantly more detailed.
The main takeaway is that an AI visibility audit gives you a baseline for how AI currently understands, describes, cites, and recommends your business.
Mike Downer: Very cool. You got going on that one.
Let’s say I’m not doing anything yet, but I’m thinking about hiring you. I’m considering making the smart move and saying, “I need an audit of my business.”
Before doing that, I’m going to take Kevin’s advice and perform a simple self-audit. Here I am, sitting in my office. Should I just search for my company’s name in ChatGPT?
Kevin Wosmansky: Great question.
I just discussed one side of this, so let me give you the other side of the coin.
No, simply searching for your name in ChatGPT is a vanity test.
There’s a significant difference between searching for your company’s name and asking what ChatGPT thinks about your business.
Someone who is already searching for your name knows you exist. That’s called a branded prompt.
The more valuable questions are discovery questions, such as:
“Who should I hire?”
“What company specializes in this service?”
“I’m currently considering three companies. Can you compare Company X, Company Y, and Company Z?”
“Which company is best for this particular problem?”
Does searching for your own name provide some value? Perhaps, but it’s primarily a vanity test.
What matters more is asking different AI models questions such as, “What do you think about my company?” or “What do you know about my company?”
That is a basic audit, and it will provide significantly more value.
Mike Downer: That makes a lot of sense.
Kevin Wosmansky: The money isn’t in asking AI whether it knows your name. The money is in whether AI recommends you when the customer doesn’t know your name.
Mike Downer: You’ve introduced several terms, so clarify something for me. This is more of a personal question.
What’s the difference between being mentioned, cited, and recommended?
This is for me more than the listeners because I’m not very smart.
Kevin Wosmansky: We’re having to educate customers about this new world. We keep discussing the need to unlearn much of what we’ve done in business marketing over the past several years. That’s the whole purpose of The Unlearning Lab.
I’m glad you brought this up.
AI doesn’t simply refer someone to a business. That’s not how the process works.
First, you have to be mentioned. A mention means your business is included somewhere in an answer.
Someone asks a question, and your business is named somewhere in the AI-generated response.
The next step is being cited.
A citation in this context is different from a traditional SEO citation. It means AI is using your content or website as supporting information.
For example, someone might ask, “What is the best mousetrap to use for this problem?”
The AI might provide an answer based partly on information from your website and cite your company as a source.
That’s a significant step because it indicates that AI is beginning to recognize you as an expert.
The final stage is a recommendation. That’s when AI specifically identifies your company as an option for the user.
Recommendations are particularly powerful because they influence someone’s consideration before that person ever visits Google.
When AI makes a recommendation, the user may have already received the answer to their question.
The more people use AI, the more they trust their AI tools. The more they trust those tools, the more likely they are to accept the recommendations AI gives them.
A mention represents visibility. A citation represents authority. A recommendation represents an opportunity for new business.
That recommendation is what every business owner wants.
AI is becoming similar to a sales representative. When it makes a recommendation, that can turn into a significant business opportunity.
Mike Downer: One hundred percent. I agree with that.
I have two sons, and they’re both young, so they’ve been using this newfangled AI technology for a while.
There are three different people in this scenario, and we use three different platforms.
That brings me to my next question. I already have a general idea of the answer because of my children, but I want you to explain why and how.
Can different AI platforms have completely different opinions about a company?
Kevin Wosmansky: It’s not a question of whether they can. They all do.
This is one of the biggest fundamental shifts in marketing.
For approximately 20 years, people have gone to the internet to look for information. Google emerged as the 900-pound gorilla in terms of search volume.
There are other great search engines, such as Bing and Yahoo, but when it comes to pure search volume, Google largely won that race.
AI large language models are entirely different.
What Gemini recommends may be different from what ChatGPT recommends. Those recommendations may differ from Perplexity, Copilot, Claude, Meta AI, Llama, Kimi, Moonshot, and many of the other models available.
Here’s what you need to understand: AI learns, and AI is trained.
What training data was the model built upon? That’s extremely important.
There have been interesting case studies where researchers asked Grok, ChatGPT, and Claude the same question. The surprising part wasn’t necessarily the answer. It was the methodology each platform used to reach that answer.
You and your two sons are using different AI models and receiving different answers because the models are different.
They have access to different information. They retrieve different sources. They interpret questions differently. They update at different times, and they may use different search or retrieval systems.
That’s why it’s so important for your business to establish itself as an expert and build a strong brand.
You want to be mentioned, cited, and recommended—not only by ChatGPT, but across all of the major AI models.
Mike Downer: That sounds like an entire episode about how to do that.
Kevin Wosmansky: There isn’t one AI search engine.
Businesses need to understand their visibility across the entire AI ecosystem.
With Google, the strategy was often, “Let’s do everything possible to get found on Google.”
Now, there is an ecosystem of many different AI platforms. You need to understand what your visibility looks like across all of them.
Mike Downer: Let’s move to the last question for this episode.
Suppose I’m a business owner and I perform my own self-audit. What happens when AI has incorrect information about me or my business?
Kevin Wosmansky: I was going to make a joke, but I’ll stick to the question.
Mike Downer: Come on. It’s a family show.
Kevin Wosmansky: When AI has incorrect information about your business, it becomes a reputation-management problem.
Let me give you a few examples.
AI might find old services, incorrect locations, outdated leadership information on your website, inaccurate positioning, or old reviews.
It may also find negative reviews or unresolved reputation issues that you never addressed.
That’s something I don’t understand. Every business will receive both good and bad reviews.
The biggest problem with a bad review is often that the business never responded. The company didn’t present its side of the story, provide feedback, or demonstrate that it wanted to resolve the customer’s concern.
Those are all things AI examines.
Another major issue is confusion with another business. We see that repeatedly.
You need to identify misinformation and strengthen your authoritative sources.
You also shouldn’t promise that you can directly tell ChatGPT to fix incorrect information. That’s not how it works.
Instead, you need to strengthen the informational environment that AI systems rely upon.
The simplest explanation is this: If AI has incorrect information about your business, then somewhere in the digital world, you have either provided incorrect information or allowed someone else to provide incorrect information about you.
AI doesn’t only look at your website at one particular moment. It may examine your website, social media profiles, independent forums such as Reddit, third-party references, and other online sources before producing an answer.
That goes back to several of the other episodes we’ve recorded. It’s extremely important to audit your business and the information that exists about it online.
Mike Downer: Perfect.
I lied. I have one more final question.
What happens if I discover that AI is recommending three of my competitors but not me?
You’ve discussed reputation, but what if all of my competitors are being recommended and I’m not?
Kevin Wosmansky: Don’t panic. It’s intelligence, my friend.
AI may recommend your competitors for many different reasons.
The starting question is: Does AI even know that you exist?
The next question is: What do those competitors have that you don’t?
Can AI find your business? That’s the first thing you need to understand.
Simply having a website doesn’t mean AI can read it or that your company is being found.
Once you determine whether AI can find you, examine what your competitors have that you don’t.
Do they have better content, stronger reviews, clearer entity information, more authoritative third-party mentions, or more complete service information?
When AI recommends your competitor instead of you, don’t guess why. Find the evidence gap and start closing it.
Mike Downer: Very good.
That brings us back to the same process: Perform the audit, engage in self-reflection, do your research, complete your homework, and then do the work.
Kevin Wosmansky: Let me end with this.
There is no silver bullet or magic pill that will instantly get your business recommended or found online.
I’m seeing so many marketing claims that say, “For $399 a month, we can get your business found online.”
The reality is that you have to build a foundation. You have to put in the repetitions, and you have to do the work.
It is not an overnight process. Anyone who tells you otherwise is lying to you.
Some businesses have already invested significantly more effort into their marketing. They’ve built content and completed much of the foundational work, so they’re going to be leaps and bounds ahead of a business that hasn’t done those things.
Again, find the evidence gap and start closing it.
Mike Downer: I agree. You can’t go to the gym, lift one weight one day, and immediately look as good as you do.
Kevin Wosmansky: They say you didn’t get fat overnight, and you’re not going to get skinny overnight.
Mike Downer: Absolutely. I’m never going to get a raise if I don’t keep giving you compliments.
All right, buddy. I appreciate your time today.
That was a lot of information packed into a short period.
Essentially, businesses need to get out there, self-reflect, look in the mirror, and perform an audit.
From there, JAR Consulting Group can help determine the right strategy and get them on the right track.
Kevin Wosmansky: Absolutely.
Have a good one, Mr. Michael.
Mike Downer: Take care, buddy. Thank you, sir, and have a great afternoon.
Kevin Wosmansky: All right.
