How to Be Visible in AI Search

Right, I think. I think we are live, everyone. I am a CEO and co-founder of PromptWatch. Thank you for joining our webinar today. How to Be Visible, in AI search. We'll be going over a couple of topics. We'll give you insights on the latest trends in generative engine optimization, learn how AI is shifting buyer behavior, and help you to clearly understand GEO and how it's different from SEO. And also, you get some expert prediction about where GEO is heading in the future. And today I'm joined by Jimmy Park. He's

A partner at Opascope and a very experienced SEO professional. Can you give a short introduction of yourself? Jimmy? Yes.

Thank you, Gijs. Hi. My name is Jimmy Park. I'm a partner at Opascope. I've been in the search industry for over 17 years, and I've built a couple of agencies and now at Opascope, and we offer digital marketing and services to companies. B2B to e-commerce. And so thank you for having.

Cool. And how long have you been working on generative engine optimization?

Well, for generative GEO. Probably been very deep working in it for the past eight months. We've been monitoring it for since May 2024 and seeing what the behaviors are and search and industry trends, but very deep into actual optimization over the last eight months.

Cool. So lots of knowledge to share today. Yeah. So I'll just, kick off by going into some latest insights into the AI search space, and then we'll switch over to you, Jimmy, with some more, some more insights as well. So maybe a little bit about PromptWatch first. So we are an AI search optimization platform. We help brands and also agencies to manage visibility in AI search. And we do this in a couple of ways. So we help to track prompts, get insights from it and form a content strategy from there. But we also focus very much on website level analytics.

So crawler analytics and website analytics, and combining all this data together will help you to get insights on how to improve and also help you to create the right content. So that's the platform and why we bring you this webinar today, some latest insights in AI search trends and why you should be working on this topic. A recent study by McKinsey has showed that about 50% of consumers already use AI powered search, and that in the coming year, 2026, it's nearer than you might think. 20 to 50% of traditional

Search traffic is at risk of of shifting to to AI search. And this will mean that by 2028. This is of course, prediction 750 billion of consumer spend will flow through AI powered search. So if you are as a brand or agency active in e-commerce. Online visibility AI search is definitely something you should be paying attention to and working on, rather sooner than later. The latest introduction in AI search is ChatGPT shopping. So yesterday, OpenAI, or the day before yesterday, OpenAI

Announced the introduction of a shopping assistant in ChatGPT, which will be available in all the plans of free Go Plus and Pro plans. And it will help consumers and also business professionals to find the products and services that they'll be looking for. And this marks a clear shift in how people find new products, and also how they'll be using AI in their shopping experience. Quick view of what this will look like. So this is a prompt I ran in the new feature today, running shoes for mid to long runs. And you see that the AI will start to show products and also already give insights and an opinion about what the

Best running shoes is and also what you see right here. What sparked my attention is the localization in the results. So I sent this prompt from a Dutch IP address, and you see that the shopping results are also localized for the Netherlands with this NL website. So this is a very interesting development and something to keep an eye on. As someone who is involved in online visibility and at PromptWatch, we try to give you as many insights as possible on this. So we introduced a new data source which is which you can find promptwatch.com/data. And what we see in our data is that ChatGPT

Is showing more and more these shopping features. So for some prompt types, 20% of the time these shopping features will be shown. So I think it's a very important data point to be looking at from now on. Also with the introduction of the latest open AI feature. So this was a quick look into to what is happening in AI search at the moment. I would now like to give the floor to Jimmy and to start sharing his insights on how to work with generative engine optimization.

Great. Thank you very much. I am going to go and start with a presentation here. Okay. Thank you for having me. One of the things that I thought to talk about is that my name is Jimmy Park. I'm a partner at Opascope, and I've been doing search for about 17 years. And one of the things that I can show today that's of values, that we have a portfolio of clients that we're able to see, sort of a bird's eye view, where many brands and companies sort of get to see one company. And so we get to see our portfolio and see what's going on across multiple brands and what we're seeing in

Google Search Console, and you've seen this probably many times, is that and this trend you may see is that impressions are going up, meaning the search engine traffic is ranking and everything is good, but the click traffic is sometimes just showing equal or and, or maybe even dropping. And one of the reasons is that there is going, there are going to be less clicks to websites. And that's due to the fact that people are using LLMs to figure out their problems and understand what they need to see back in

2025, early 2024, 2023. You would have a problem or something that you were trying to find out, and then you would go to the search engines and it would give you a destination that you could go find the information to solve your problem. Now, in this new buyer journey, the LLMs are actually solving the problem for you. And so what is happening is from what we're seeing, is that what before you would have a blog, you would have a feature page, you'd have listicle. All of this content where someone is going in, clicking in, coming back out, clicking in and going back out, and then finally getting to a brand search

Or a pricing page for the ultimate conversion in 2026 and beyond. What's going to happen is you're going to see some Google, but a lot of LLMs searches that are giving a lot of this information through the process and understanding how to make a decision. And at the end you're going, the websites are going to get the last click or the last few clicks, meaning that overall clicks are going to decrease because in the time that the LLM is giving you the response, they're actually going out there and spending time in the clicks and assembling that information

To give you what you want. And so the process that LLMs is, the user asks a question. AI understands the issue and they understand linguistics. They go and they source information and going out and hitting websites. Top ten, top 20 and various searches for the problem. And then they start to understand the priorities and details of the decision making process. And that identifies options for solutions. And importantly, it seeks out various persona types and variable variability of each of their needs. Then it presents this information and sometimes more qualifying

Questions really quick. Let's go take a look at how this sort of works. Look, I did a B2B search and a B2C search a best scent diffuser to make your house smell nice. E-commerce or a CRM and you can see what the response that comes back is when you're asking a question. First, it teaches you what to look for and how to solve your problem. It's it's teaching people to be better decision makers and what's important in the decision making process for CRM. It would be all of these for large room diffuser, talking about technology

And placement. And then it gets to the second area of the top options to consider. And so it's going through and giving those options. And this is where companies want and brands want to show up. And our clients want to show up as an option here. And then further down the prompt response, you can see that it starts to talk about and how you choose the right one for your company or based on your space, and it tells you some other information and can lead you to a question that says, okay, if you tell me a little bit more, I can even

Get you closer to the solution. And so understanding how these work, it helps them to understand what's the same and what's the difference in GEO and SEO. First page gets the traffic. You're going to optimize for blog articles and long tail high content production. You're making a lot of blogs. You're putting content on pages, backlinks, and of course technical SEO. Some of the differences in GEO is the URL formatting. They're looking at URLs a little bit more meta descriptions that's not on this page is something that's used

More recency comparison. Listicles answer first content and it's important to understand why. Why do comparison listicles? Why do they do better? So when you're thinking about some of the content, or when you're using PromptWatch to be able to understand what articles to write, then the reason why is because the prompts will only have a few seconds to go out and find this information of the decision factors and some options. And so if they only have enough time, if they see an article that they can get 4 to 5 different brands of information in a single click, when it's going out to search engines

To find that it saves at time. So summary information comparisons, these types of information, if you think about it, saves time for the Lem to provide an answer to the user. You can see why this is important. And so another thing because they're only looking at the search engine results page to understand. And which one should I go click? That's why you are all formatting is a little bit more important, as well as the meta description which is not used in SEO, but is something that the LLMs will use to be able to sort of steal a click, or to try to get the MLM

Agent to come and crawl your article or your piece of content so that you are one of those options. Now, when I first started using PromptWatch and you're here because you're probably interested in some tools to use. We went out and vetted everything from SEMrush tool to Ahrefs, a bunch of them out there, and then we landed on PromptWatch for a couple of different reasons. One is that it used information for us to track monitors and track various prompts and using monitors to track that, but it gave us access to things like how

The agents are calling the site versus how visitors are calling a site, which is very important if you think about that bar graph that I showed, there's a transition point of where the agents are getting their information and then turn it over to a human to be able to say, okay, here are your options, and that's where you're getting the clicks, and it starts to give you information of content that you can create for the agents, as well as pages for humans when they start to enter. Now, when I first started to use PromptWatch, I would take a full keyword list because my mind was still thinking in keywords

And it took me about 4 to 5 months to sort of train my brain to move out from a keyword type search to more of a concept type search. And so what I would do is I would prompt, I would put a list of keywords, and some of them would be very focused and some would be ones that I wanted. But what I found is putting in a very small amount of very focused keywords to help me generate prompts, and then using prompt, which would give me a list of various prompts, but still very focused on the keyword. And so most of the time when I'm using and I'm using,

I'm generating my prompts from this keyword, this keyword tool. Now let's go really quick about how to go find the keyword, let's say for HubSpot. And you're ranking on 100, 800, 13,000 keywords. And this is a tool using SEMrush that you all probably know. But let's get down to the keywords. And let's say, for instance, now I'm thinking about email and email marketing. I get down to 6400 words with by searching only for the word email quickly. Then I start to go and I'm going for the top three. So you go to your website, you pick a specific

Service or a brand or a product. Get to the top three. I'm taking out brand and then I want to go for the slash products page, because that's the core service of what I'm offering. And then you can see I drop down all the way from 813,000 keywords to 23 keywords. And the ones I want to pick are the ones that when you configure PromptWatch and you're going for your prompt monitors. The question I ask myself is what is the keyword? If I rank number one and they click into the site and then they go to it and they say, this is exactly what I'm looking for CRM with email

Tracking, sales email tracking, email content management software. This is very high. And these are the keywords I would put two, 3 or 4 into PromptWatch, not 50 of them, 4 or 5 into the generate by keyword prompt monitor. And then you get a set of. I usually start off with about 10 to 15 prompts. I start to go through those. Get an idea? Am I right or is this close? I'll usually take 2 or 3 out prompt, which is very good. If I put those very solid keywords in, I can get to these monitors.

Here's an example about ABA software. Now I can go and see how can they be a therapist. Software to data collection day to day tasks. And These are variants. If I take a look at these are right exactly the prompts that I want. And now I'm able to use these prompt monitors to see volume difficulty type. And then how to then start to target the content that I will. Now many times when I first started using PromptWatch, I would go in and I would start to think, all right, what am I going to do with the prompts and these monitors?

And I would see I would go out and try to create some content to try to see if I get a little bit more visibility and you can do that. But after months of using the tool, what we have found very effective, if we go back and try to find this area right here, which is understand the priorities and details of the decision making process and then become an option. So what I'm trying to do is provide content that shows how we do, based on the priorities and the details of the decision

Making process. Now, how do you go and find that? Well, in PromptWatch, what we can go do is now we set up those monitors and you can click into the tool under the responses. And when you go into the response details, what we do is we go into here and we find out what are the factors, what are the decision making factors in the prompt details, number of users for this software platform. Number of users. Additional features. Customization is important. Telehealth options. Other key features. These are responses

That we're looking inside a PromptWatch because PromptWatch goes out and hits those monitors, get the responses, comes back, and starts to assemble. Various platforms and LLM platforms because the Google responses are a little bit different than ChatGPT than perplexity. And what we want to see is what the responses are saying. Oh, automated data capture is important for our for this brand's clients when making a decision. They want an all in one solution. Geolocation support, signature capture, all of these types of features. Now what I can do with this,

This information is assemble this and you can actually pull this down if you want to. You can export this information. PromptWatch uses this for other tools. But sometimes as someone who's managing this, I want to see how the ICP how the customer is thinking. And how they're making a decision. And now I can go make content if I'm going to make a listicle, if I'm going to do a comparison table, I can put my brand in a couple of other brands as comparison. And what am I going to compare? I'm going to compare additional feature customization.

I'm going to compare geolocation features. Does it have automated data capture? Is it HIPAA compliant? Does it have a user friendly interface? This is the information that I already know that LLMs are starting to look for or what they want to see. And if I'm able to provide to them holistically with my brand as an option, I have a much higher chance to be listed as an option in LLM responses. So that's one of the very targeted ways that we're actually using this tool is to start to go through and not just analyze the visibility, but

Analyze the results to help you build some of the content. One other area that I think is very important, and I talked about this a little bit is that there are tools that are out there that can help track, and you can use things like Google Search Console and GA4 to see what traffic is coming in from humans and what are clicking. But what's very important to see is how the LLMs and how the agents are crawling the site. What it allows you to see is where you're being cited, where you're being used as a citation, where it's coming in to learn.

You know, there's two aspects that the LLMs are coming in. They're coming in to learn, and they're coming into on demand prompts. So someone's out there and they put in some sort of question or problem, and within three seconds the LLMs are going out and hitting search engine and hitting various websites and starting to come back and assemble that information to provide the answer. And so to see where the LLMs are crawling, for example, one insight that we've seen is that LLMs we talked about recency LLMs are weighted towards content that's built in the last

Three months. So where do they check? Well, one of the things that we discovered is it goes and crawls the new sections of websites. If you have a slash news and you have information that's posting it there, if we think about it, why is that important? Because the LLMs wants to know what you as a company are doing in recent months, what have you changed? And so yes, they're taking a look at the site map. Yes, they're taking a look at the most recent content. But what they're also going to do is they're going into the news section and seeing do you have press releases, did you make a product release? Did you release a new category

Or whatever? But that news section, which is funny because back in the SEO days, there was a lot of press releases that were going out, but it was for backlinks, but the LLMs are looking at in a slightly different way. And so to understand where the crawlers are coming in and again, where that transition from crawler traffic is seeing these pages, these pages, articles, news, feature releases, release notes, that type of information. And then when the humans start to come in, you will start to see that they're coming in a little bit a lower funnel. They come in from information

That they've already gotten through, those high level questions that the agents have been giving them the information, and then the transformation into human traffic. When you start getting into the user and the visitor analytics coming in. And so to know where that transition is very important and can help you target some of your content.

And maybe also just to interject a little bit, but also maybe also good to have in mind here that in PromptWatch that you can also see the different types of crawlers that are out there. So the ChatGPT citation crawler, the search bot, etc. They serve different functions and give you information exactly what Jimmy was mentioning, what you're what you're looking for. So that's also something you can see for the different LLMs of what kind of bots are hitting what kind of pages. And it helps you to see what the focus is on basically.

Exactly. And visibility is different. Obviously with ChatGPT you're getting a great deal of information or you're getting a lot of users. Many users are using ChatGPT in terms of a prompt question type answer. But there's been a big push lately, especially in the last three weeks by Google. And Google has made some changes in terms of being able to serve ads and so forth. So to be able to see what Google is giving from a response, as well as perplexity or Claude or different depending on what your industry, there are industry weights on

What they're using in terms of the LLM platforms. And so when we're going and analyzing some of the decision spaces or some of the returns and some of the information that's being presented in the prompt responses, doing that by LLM platform is important. And so again, we're learning a lot from ChatGPT because we're getting a fair amount of traffic for our clients there. But at the same time, Google's making a recent push in something that we're going to analyze more over the next year. So I think that's what we have for time. And I went

Through fairly quick on the presentation. But if you have any questions, I think there's a chat to be able to ask any questions or maybe there's some coming from LinkedIn.

Yeah. Let's see if there are any questions coming in right now. Well I have some a question for you, Jimmy. So you've been working on this for the last couple of months. How do you combine the geo practices and the SEO practices at the same time? So how do you make sure you're visible in Google, but also do visibility in AI search at the same time?

Yeah. So one of the things to do is to and we talked a little bit about this to split up some of the content type. And so if you take a look at a marketing funnel and some of the information that used to provide that, we're going from blogs and that type of thing. I would just sort of move down that line where some of the top of the funnel content information now is being pushed a little bit more towards GEO than it is towards SEO. And when we get a little bit lower down the funnel and starting to get down into human traffic in that transformation,

We push again back sort of towards SEO and making sure that we're there. Now, it's hard to say is that you want to do one or the other. It's just a little bit of a sort of a mixing weight of that. So let's say, for instance, if I have some top of the funnel traffic that I feel is a little bit older, that has gotten some traffic but has started to drop down in terms of clicks coming from SEO. And you go into Google Search Console, you had a blog that you used to perform very well. Now it's sort of dropping in terms of clicks, but you're

Still getting that impression. That's an opportunity to be able to start to optimize. That's where I would maybe go in and change the meta description. I may refresh the title, add some new information in there that's recent to 2025 or coming up to 2026, and then start to be able to try to push that higher into the LLMs. Again, changing the meta description might even change the title of the article and put some recent content in there. And that's one of the ways that we can target SEO versus SEO.

Yeah. That's also what we see in our data. Right? So we aggregate a lot of data on AI search. And we see that the pages with the right titles get a lot more traffic. So with that recency bias, it's really something that is proving itself to be right a couple of times already. Question that's coming in. So what are your thoughts and experiences on sites facing traffic drops after LLM releases? Do you see conversions or do the number of conversions actually stay the same?

If you're if you're doing it right, they should stay the same or actually increase. One of the things that I that I've spoken about many times when I'm asked this question is, let's say, for instance, you are a software platform or CRM platform like we had in the example, or you're selling sunglasses, the same number of sunglasses are going to be sold next year. There are the same number of eyeballs and the sun is still the same. So the same number of sunglasses are going to be sold. The question is where? Where are they going to get them. And so the number of conversions are the same

Worldwide in terms of what exists in the world. And so if you're doing geo correct, that should be able to go up higher. So you're starting you still are doing the SEO and you're still getting being an option. But when you get to that conversion at the end, you want to be very geo focused. In 2026, I think there's going to be a major shift into LLMs for a couple of different reasons. One, we're seeing integration happening with ChatGPT. We're seeing ads coming in from Google right now. People aren't using LLMs to shop, and that is a majority of the population

That's out there. And I personally believe that once people like my wife are going to go out there and start to shop and using it to help buy things or to shop around, I think that transition is going to start to LLMs because it's good. It teaches you to be a better shopper and a person to make a better decision. And then I ask this other question for those who are using LLMs and using LLMs a lot, how many out of 100 searches have you? Do you now go to ChatGPT or Google AI tools versus an old classic search like SEO search most

Of the people who are using LLMs. Once you made that transition 25 to 90% of your interactions, starting with an LLM versus google.com search engine. And that to me is very telling. And so once people make that shift, I think the shift is going to be actually faster than people think in 2026, once the shopping starts to occur and the general population starts to move over.

Yeah I can I agree completely. I think 2025 has been the years of early adopters in AI search. And in 2026 the general population will will follow. Indeed. Yeah.

Once they switch it's it people don't really switch back. So.

Yeah. Yeah, exactly. Let me see another question. Can a local expert shoe retailer now win back some of the revenue from their from their global competitors? How do you look at that? Jimmy?

Yeah. So what? Okay, so if you think back on September 8th or 9th, Google removed the num=100 search, meaning that you could go and plug into the URL, do a search and do num=100. You get 100 listings instead of just the top ten. They took that away. Why did they take that away? Well, if you read between the tea leaves a little bit, it's likely due to the fact that LLMs were using that num=100 search to go fast to be able to scan through the top 100. What we have found is back in the day in 2023, 2024, before long

Popularity, you would have to rank at the top ten really to top four to get SEO traffic. But I believe one from that num=100 search. Also, understanding how our LLMs are pulling in citations, and you could see about 15% of the searches that are using sources are on the first page. Most of them are not in the first page. And what does that mean to a small retailer is now you don't have to rank in the top ten, which is very difficult for small business going against some of these bigger, larger clients. If you can get into the top 20, 30, 40 now and write a very

Good description, a very good URL slug and start to do, you may get selected as a source of citation as an option. I actually think in the world of LLMs, yes, you can definitely win back some and you are able now to battle a little bit easier, more on a level scale if you do things right and being able to understand how they think, create good content, and now be listed as an option against some of the big players, I think that you can win back some in 2026 and in 2026 is the year to do it because there

Are some efficiencies that need to be gained. Some of the big players are a little bit slow on geo optimization, and if you're aggressive on 2026 and you might be able to win some of that back.

Cool. Well, I think that's a that's good to hear for these these smaller retailers out there that are looking to to win some extra market share. I think that wraps it up, for today. Thanks, Jimmy, for your time and presenting us the information, the webinar is recorded. So, when you just joined in 15 minutes ago or want to watch some things back, you can watch the recording. Thanks for your time, everyone who joined. And, have a great day.