We have to stop treating the absence of a click as proof of failure.

This episode explores the shift from click-based attribution to AI visibility, emphasizing that 'zero-click' searches can still be successful. Monica Tomasso introduces a framework for SMBs to move from being merely mentioned to being recommended and eventually 'chosen' for direct action within AI platforms.

A founder interview with Monic AI Systems, published 2026-11-16 (27 minute interview).

Topics covered

Questions answered

How do I choose someone to help me get visible in AI search?
Ask three questions, because a report alone does not tell you what to do next. First, are you measuring whether AI mentions my business, or whether it would actually recommend me? Second, can you explain what you are doing to improve my visibility beyond monitoring the results? Third, can you show me what has changed over time, including the limits of what you are measuring? In our view, a report tells you where you stand, and a useful strategy tells you what to improve next and why. Some people tell us they don't want another dashboard. They want a provider who does more than show them numbers. A provider who is open about what they can and cannot suggests is usually a good sign.
What is the difference between being mentioned, cited, recommended and chosen by AI?
They are four different stages. Being mentioned means AI knows your business exists. Being cited means an AI references your website or another source associated with your business, which matters because otherwise the buyer has nowhere to go for more information. Being recommended means AI identifies your business as a relevant option for someone's specific need. Being chosen is when the buyer actually takes action. Action is the part now starting to arrive. People are beginning to book or buy from inside tools like ChatGPT without visiting a website. In our experience, recommendation is already happening. Metrics for the action stage for smaller businesses are not really available yet.
How do I know if my AI visibility work is actually working?
Look at whether you are being recommended, not only mentioned, and be careful about what you claim from it. Say a company appeared in none of the 5 buyer questions we tracked and is now recommended in 3 of them. That is measurable progress in recommendation visibility. It does not suggests more website visits, inquiries or sales, and it does not suggests our work caused the change. If the 5 questions were tested only once, you also have not shown the change is consistent. Repeating the test over time helps separate a lasting change from ordinary variation in AI responses. We can show recommendation visibility improved for the questions we tested. We cannot automatically say customer acquisition improved. Those are two different things.
Why don't I see AI referrals in my analytics, and how can I track where customers come from?
A missing click does not tell you what happened, so analytics alone will leave gaps. Someone might ask ChatGPT for a recommendation, remember only your business name, then search for you on Google or call you directly. Your analytics show a Google visit or a phone call, and the AI step is invisible. A missing click could also mean the person got everything they needed, or that they did not trust the answer and gave up. One practical step is to ask every new customer, "How did you first hear about us?" and include AI assistants as an option. It is not perfect, because people forget. In our experience, clients often cannot recall the exact words they typed. It still gives you a signal your website analytics may miss.
How do you track whether a business shows up in ChatGPT, Claude, Perplexity and Gemini?
We run a set of buyer prompts for each business and look for patterns. If a business has 25 or 30 prompts in our platform, we ask each large language model every prompt a meaningful multiple, because the answers can differ from run to run. One prompt might show your business in 3 of 3 runs, another in 1 of 3, and another in none. Over time those results shift. New clients are often surprised by this, but they come to understand it as they watch the pattern. We cannot pull the exact conversations people have inside these models. So we test prompts and look at the mentions and citations that come back. In our testing, a handful of prompts, around 5 to 8, tend to bring a business up most often.
What should a small or mid-sized business do to start getting recommended by AI?
Build corroboration across the web, not just on your own website. Businesses that get recommended have corroborated the web, meaning third parties confirm who they are and what they do. That includes testimonials and profiles on major platforms such as G2, Clutch and GoodFirms. You have to be visible and mentioned before you can be chosen. If you cannot be found and mentioned, you will rarely reach the action stage. It also helps to think in phrases rather than keywords. Buyers ask these models full questions, so understanding the phrases that are actually coming into large language models is a key shift. To see where you stand, you can use an AI visibility checker to see how you appear in large language models such as ChatGPT, Claude, Perplexity and Gemini, and compare against your competitors.

Transcript

Hi Monica Tomasso — welcome to episode 6 of the Say Hi Podcast for Monic AI Systems. Over the next 30 minutes I'll ask you strategic questions to pull out your best insights. Just speak naturally. Ready to begin?

Yes.

Great! Today we'll go through the questions you prepared. Let’s start with the first one: When someone asks ChatGPT a question, gets an answer, and never clicks a website, how can a business know whether that interaction was successful?

Well, we don't know. And that's the exact problem. I think for years, for years we tried clicks as evidence that someone was interested, but now an AI answer can But now AI can answer a question without even requiring someone to visit a website, which is kind of crazy. So a missing click could mean that the person got everything they needed, or it could mean they didn't trust the answer and abandoned the search. We have to stop treating the absence of a click as proof of either outcome.

That’s such a clear distinction, Monica. Can you walk me through a recent situation with a client where you saw this ambiguity play out—where you couldn’t tell if a missing click meant a satisfied user or a lost opportunity? What did you do next?

Yeah, so I personally see it. So what happens is like if you go to a large language model, let's just choose ChatGPT, right? If you go in there, and you are looking for something and ChatGPT can only mention the name and not give you the link to the website, where does somebody go after that, right? It's like, do you have to take that name and go into Google or maybe prompt ChatGPT a little more to get more information? And that's exactly the situation. So you take that name, you put it into Google, Google gives you the URL and boom. So where did the attribution start? Like, did it start from ChatGPT Dear Google, that's the issue.

That’s a great example. In that scenario, did you see any way to track or estimate how often users were taking that extra step—copying the name from ChatGPT and searching it elsewhere? Or is that still a total black box for businesses right now?

I think it's a black box for businesses. I do think it's getting better as more companies are starting to be getting citations and links and mentions, but right now it's evolving.

That makes sense. Let’s move to your second question: Microsoft researchers found that some searches without clicks were successful. What does that tell us about using website traffic to measure AI visibility?

Oh, well, the research challenge is something we've taken for granted in digital marketing. The more clicks automatically mean more better results. Some people leave because their question was answered. Others click and still don't find what they're looking for. So clicks remain useful, but they're incomplete, as I just mentioned in question 1. We need to look at the quality of the answer and whether it helps someone move toward a decision.

Absolutely. Can you share a specific example where you saw website traffic drop, but you had other evidence that customers were still getting what they needed—maybe through direct calls, in-person visits, or another channel? How did you piece together that the AI answer was actually helping, even without the click?

Well, the AI answer is giving the idea right? And from there, the idea gets researched and it evolves to the click because you can have conversations inside these large languages now.

That’s a good point—the conversation can continue inside the AI, even if there’s no website visit. Have you seen any patterns in the types of questions or industries where this happens most often? For example, are there certain SMB clients where you notice more “zero-click” satisfaction, or is it pretty universal?

Um, I think it's pretty universal right now.

Thanks for clarifying that. Let’s move to your third question: Suppose a client's website clicks decline while their AI visibility and recommendations increase. How would you interpret that, and what would you investigate before calling it a success?

Yeah, I wouldn't immediately call that a success or failure. I'd first want to understand what visibility actually increased. Are they simply being mentioned or are they being recommended for the services they provide? And then I'd look at other signals. Are more qualified inquiries coming in? Are people mentioning Chapter P when they contact the company? Are they appearing in more relevant buyer conversations? If those indicators improve, That's encouraging, but I wouldn't claim AI caused the business growth without stronger evidence. In fact, there's data out there that the conversion is still really low, but when it does happen, it's super qualified.

That’s really insightful. Can you give an example of a specific signal you’ve tracked—like a client reporting more direct inquiries mentioning an AI platform, or a noticeable shift in the types of leads they’re getting? How did you connect those dots to AI visibility, even if the website clicks were down?

Yeah, I mean, definitely have had clients report their audience level has dropped, like their— his higher-paying clients has dropped. And I'm not sure I can attribute that to AI. It could be a lot of things, and that's where we really need to be looking at our Google Analytics, looking at a lot of analytics that is coming in about the business and where those clicks are coming in from to get that you know, drop in like audience.

That’s a great example of being careful not to over-attribute. When you see those kinds of drops, do you ever run specific tests or experiments—like changing messaging or tracking new referral sources—to try to isolate whether AI visibility is playing a role? Or is it mostly about monitoring and waiting for clearer patterns?

I think it's monitoring, but right now we have full monitoring on like if we have 25 or 30 prompts that are coming in on a business inside our platform, we prompt each of the large language models 3 times with that prompt because you can get different answers. So we look for patterns in the data.

That’s a really practical approach—prompting multiple times and looking for patterns. Can you share an example where that pattern analysis revealed something surprising about a client’s AI visibility or recommendations? Maybe a case where the results were different than what you expected?

Um, I wouldn't say, I wouldn't say so. What's interesting to clients when they see prompts coming in and they're in that, that one prompt is in like 3 of the 3 and another prompt is in one of the 3 and another one is in 0 to 3. And over time it changes, which is why we see the patterns. But I think as a new client that comes in and sees that, it's surprising. but over time they get to see the pattern and understand it better.

That’s helpful context. Let’s move to your next question: What’s the difference between an AI system mentioning a company, citing its website, and actively recommending that business?

Oh, that distinction is central to what we do here at Monic AI Systems. Being mentioned means that AI knows your business exists, which is fantastic. Being cited means an AI references your website site or another source associated with your business, which is sort of what we talked about earlier, that citations are really important because if not, then where do you go to get that information? And then being recommended is different So that's when AI identifies your business as a relevant option for someone's specific need. And then there's being chosen, and that's when the buyer actually takes action. And one of my last episodes, we talked about the action, right? Like, think about, can somebody now book a hotel room inside a large language model like ChatGPT without ever going to the website? Or Can somebody book on your calendar as a business, create a meeting right from ChatGPT? And that's all coming out right now. So it's being mentioned, being cited, recommended, and chosen and then actioned on. That's what we're looking for these days.

That’s a really clear framework. For your SMB clients, do you find that most are focused on moving from “mentioned” to “recommended,” or are some already seeing that “chosen” or “actioned” stage—like bookings or direct contacts—happen inside AI platforms? Can you share a recent

I

example?

think the whole action piece is just coming. I don't think that we're— we can find metrics for the smaller SMBs at this moment, but being recommended is absolutely happening. And that happens when you see a business who has corroborated the web, who has literally trained the web on their business, not just their website. So it's again, getting third-party corroboration, testimonials and major platforms like G2 and Clutch and GoodFirms is a big deal. Yeah, that's, that's exactly where it's at.

That’s a great point about third-party corroboration. For a client who’s just starting to see those recommendations, what’s the first indicator you look for that tells you their AI recommendation visibility is actually improving? Is it the frequency of recommendations, the quality of the prompts, or something else?

Um, definitely the frequency of mentions.

What's

When you

interesting is now that ChatGPT has launched their Business Hub, you can now start to track your mentions. It tells you how many times you're mentioned inside ChatGPT, which is truly amazing.

That’s a big development—being able to track mentions directly in ChatGPT’s Business Hub really changes the game for SMBs. Are you seeing clients start to use that data yet? And when you look at those mention counts, how do you help a business interpret what’s meaningful versus just noise?

Yeah, so, um, I think that announcement with the Business Hub is just getting announced and people are getting in there. Um, gosh, tell me the second question you were asking me again. It's right on the tip of my tongue.

Of course! The second part was: When you look at those mention counts in the Business Hub, how do you help a business interpret what’s actually meaningful versus what might just be noise or irrelevant mentions?

Right, since we can't pull directly out of the large language models the exact phrases and conversations that are happening inside these large language models, we have created an opportunity to find the right prompts for that business, and we take and look at exactly the citations or even the mentions that are coming in on those prompts from the large language models, and that's the details that we can pull just to find out what prompts for that business are valuable or not.

That’s a smart way to focus on relevance—by testing specific prompts and tracking which ones actually generate meaningful mentions or citations. Can you share an example where you discovered that a certain prompt was driving valuable visibility for a client, even if overall mention counts were lower than expected? What did you learn from that?

Could you repeat that? That was quick.

Absolutely! Here’s what I asked: Can you share an example where you discovered that a specific prompt was driving valuable visibility for a client, even if their overall mention counts were lower than expected? What did you learn from that experience?

There's specific prompts driving overall. Yeah, there's definitely like the top like 5, 8 prompts that are getting higher percentages and being mentioned when we prompt the large language models with those phrases, right? Um, that's, that's how we learn what are like the main prompts for that business.

That’s really actionable—focusing on those top-performing prompts. Have you ever had a client who was surprised by which prompts actually worked best for their

Yes.

business? Maybe they expected one service or keyword to dominate, but something else ended up driving more AI mentions or recommendations?

Yeah, so a lot of businesses are so focused on keywords because that's where we came from, SEO, right? But it's not keywords anymore, it's phrases, and that's the paradigm shift. So first understanding the phrases that are actually coming into these large language models is the first aha for them. And then really understanding, wow, like, that's amazing that I have content that closes that gap and the content comes from our system, like Monic AI Systems system. So that is a revelation.

That’s a powerful shift—from keywords to real conversational phrases. Let’s move to your next question: What’s something you’ve deliberately refused to promise or report to a client because the available data couldn’t prove it?

I won't tell a client that an increase in AI mentions automatically means more customers because right now that is not the case. I also won't claim that a citation caused a sale simply because the company appeared in an AI answer because that's not the case. We can measure visibility, we can observe recommendations, we can sometimes connect those to customer actions, but there are gaps between those stages. And I think being transparent about those gaps is part of doing this work responsibly.

That’s a really honest stance. Have you ever had a client push back—maybe asking for proof that AI visibility was driving sales or leads? How do you handle those conversations and set expectations around what’s actually measurable

I love

right now?

that question. You know why? Because several of my clients found me inside ChatGPT, recommended me inside ChatGPT. And when that happens, they become my clients because they see that I can do that for myself. She should be able to do that for me.

That’s a great real-world example—your own visibility serving as proof of concept. When those clients mention they found you through ChatGPT, do you ask them for details about their journey? For instance, do you try to find out what prompt or question led them to you, or do you just note that the recommendation happened?

Yeah, no, that's a really good question. I have definitely tried. And if you ask the client the words that they use, they don't remember. One of the clients One of the clients had said, who are some good AI visibility companies that work with 15, 20 people company? Because they didn't want like an enterprise system, I'm guessing. But yeah, that's a fun question. But yeah, I do try to understand it, but a lot of times the clients don't recall how it backed me into the recommendation out of all the AI visibility companies out there.

That’s so true—people rarely remember the exact phrase they used, especially in a conversational AI setting. Let’s move to your next question: If AI recommends three companies and the customer contacts one directly without clicking an AI referral link, what can the business realistically measure?

When there's an attribution that becomes difficult. Someone might ask ChatGPT for a recommendation, remember the business name, and then call directly or search for the company on Google, like I mentioned. The business may never see an AI referral in analytics. One practical thing I encourage businesses to do is ask new customers how did you first hear about us and include AI assistance as an option. It's not perfect. People forget where they first encountered a company, but it provides another signal that traditional website analytics might miss.

That’s a practical workaround—adding “AI assistant” to the “how did you hear about us” question. Have you seen any clients actually start to get responses that mention AI platforms or assistants? If so, what percentage of new leads or customers typically mention that channel?

Um,

Take your time—if you have a ballpark or even just a recent example, that would be great. Have you seen any clients surprised by how often AI gets mentioned as a referral source?

Um, I don't know how to answer that question, to be honest. I think that we're trying to answer the question, are they surprised when they get referred?

Exactly, yes—that’s what I was getting at. Are your clients surprised when they discover that new customers found them through an AI recommendation, especially since it’s such a new

Oh.

channel? Or is it something they’re starting to expect now?

Yeah, you know what is so cool is when people for the very first time hear like that they were found in ChatGPT, like their business was found in ChatGPT and that they got a client from that, they're like, oh my gosh, I didn't know this was a thing. How do I get more visible? How do I get more clients? And that's happening. Those are the people that that also become my clients. So the people that just got mentioned or found in ChatGPT, it's pretty incredible.

That’s such a great moment—when a business realizes they’re being discovered through AI for the first time. Let’s move to your next question: What should a business owner ask an AI visibility provider to determine whether they're improving meaningful recommendation visibility or simply generating another dashboard report?

I like that. I'd ask 3 questions. First, are you measuring whether AI mentions my business or would it actually recommend me? Second, can you explain what you're doing to improve my visibility beyond monitoring the results? And third, Can you show me what has changed over time, including the limitations of what you're measuring? A report tells you where you stand. A useful strategy tells you what to improve next and why.

That’s a really actionable set of questions. Have you ever had a client come to you after working with another provider who only gave them dashboards or reports, but didn’t actually help them improve? What did you do differently to move them from just measurement to real visibility gains?

Well, that's a good question. I haven't had somebody come from another company yet. What I have had was people coming to me saying, I don't need another dashboard. So that might be the fact that they have seen that a lot of companies are just dashboards, but I haven't had one specifically say to me, we just came from another company and it was just dashboards. They've only said I need something more than just a dashboard.

That’s helpful context. Let’s go to your final prepared question: If a company goes from being invisible in AI answers to being recommended in three out of five relevant buyer questions, what have we actually proven, and what haven’t we proven?

Say that again.

Of course! Here’s your question: If a company goes from being invisible in AI answers to being recommended in three out of five relevant buyer questions, what have we actually proven, and what haven’t we proven?

That's a great example. Um, if a company wasn't appearing in any of the 5 relevant buyers we were tracking and now it's being recommended in 3 of them, we demonstrated measurable progress and recommendation visibility. But here's what we haven't proven. We haven't proven that 60% of all people asking AI about the industry will see that company. We haven't proven that those recommendations resulted in website visits, inquiries, or sales. And we haven't necessarily proven that our work caused the improvement. Other factors could have influenced those results. What we have evidence is that the business is becoming more recommendable within the questions the AI is measuring. So the next step is to understand whether those recommendations are consistent, whether they appear across different AI platforms, and whether we're seeing corresponding changes in business outcomes. That's why I believe AI visibility measurement needs to be beyond counting mentions. It's not just about whether AI knows you exist. It's about whether AI recommends you for the right reasons and ultimately whether customers choose you. One distinction I'd like to add: if those 5 questions were tested only once, you observed an improvement but haven't yet established it's consistent. Repeating the test over time helps distinguish a sustained sustainable change from ordinary variation in AI responses. We can prove that recommendation visibility improved in the questions we tested. We cannot automatically claim that customer acquisition improved. Those are 2 different things.

That’s a really nuanced answer, Monica. For your clients at Monic AI Systems, how do you communicate that distinction—between proving recommendation visibility and proving actual business outcomes? Do you have a way of showing them both the progress and the limits, so they know what’s real and what’s still unproven?

Yes, we have a ton of data that they have access to. They have the ability to look at what's trending, where their business is trending, um, we can, we can demonstrate the visibility metric, we can show the patterns, we can show it over time, and we can show that they're demonstrating good business value.

That’s fantastic—giving clients access to those trends and patterns really helps them see the bigger picture. Since we still have a few minutes, I’d love to dig a bit deeper. You mentioned earlier that the “action” stage—where a customer books or contacts directly through an AI platform—is just starting to emerge for SMBs. What do you think will be the biggest shift for small businesses once those in-assistant transactions become more common? How should they prepare now?

Oh, that's a big question, and it's just starting to unravel. I think I brought this up in one of my last podcasts. One of the biggest things that we are seeing is that— I think it's the company is called Bytes. They came out and started taking orders inside ChatGPT. out in the California area and going against DoorDash. And they were giving lower cost of food and delivery. And so they were averting revenue from DoorDash So it's things like that that is going to start to happen. I personally— Monic AI Systems has a booking calendar inside ChatGPT, so somebody doesn't need to go to my website. They can book right inside ChatGPT. And that's, that's what's happening. I think that there's now announcement that came out from ChatGPT just yesterday that they can— an agent And I think it's their dots. An agent can— you can give— you, you can go out and book. You can tell it like, go do my Christmas shopping, and it doesn't need your credit card, and it just will order it, pay for it, and ship it. That's what's happening. And it's, it's coming quickly.

That’s a huge shift—especially for SMBs who’ve always relied on website visits or phone calls to drive business. For your clients who want to get ready for this new world of in-assistant transactions, what’s the first practical step you recommend they take right now, before it becomes mainstream?

Oh, that's— I think the very first practical step for SMBs and mid-level would be to mid-level SMBs is to get the Business Hub on ChatGPT. You can start to build some insights there. Corroborating the web is huge. That's still very huge. And that will help you like get mentioned. You have to first get visible, get mentioned, get chosen in order to get that action. Like if you can't be visible and can't be found and mentioned, you'll never get to that action.

That’s a clear roadmap—visibility, mentions, recommendations, and then action. For an SMB that’s just starting out and maybe feels overwhelmed by all these changes, what’s one small, concrete thing they can do this week to start improving their AI visibility?

Well, I have an opportunity to give a free AI visibility. You can go to Monic AI Systems at the very top. You could check your AI visibility. You can see where you're coming into many of the large language models, ChatGPT, Claude, Perplexity, and Gemini. And you can also, after you get your results, you can see your competition. You can see if you're getting mentions. You can see like if you're visible. There's a lot of data that comes from that free visibility checker. And then you can also work to get a 90-day roadmap right there from the platform, and it'll give you I think, 90-day roadmap. It's an incredible offering.

That’s a fantastic, actionable starting point—using the free AI visibility checker at Monic AI Systems to see where you stand and get a roadmap. Monica Tomasso, this has been so valuable. Thank you for sharing your insights on behalf of Monic AI Systems. Our team will transform this into content assets for you. Have a great day!

Thank you.

Thank