Why AI Isn’t Killing Search (And What to Focus on Instead) with Tony Pataky
Why AI Isn’t Killing Search (And What to Focus on Instead) with Tony Pataky
In this podcast, Kevin Kerner sits down with Tony Pataky, Director of SEO and Marketing Performance at Procore, to discuss why the "death of SEO" is a myth fueled by anxiety and how marketers can navigate the new reality of AI-driven search.
Tony provides a grounded, operator-level perspective on maintaining search dominance, leveraging data from the $500 million acquisition of Levelset to his current role leading search at a global construction tech leader.
The Strategy: Why Brand Authority Wins in 2026
While the industry is currently fixated on AI Overviews and LLMs, Tony shares a startling reality check: for most mid-to-large companies, less than 1% of total traffic currently comes from LLMs like ChatGPT or Perplexity.
Meanwhile, organic search remains a massive engine, driving 30-40% of total traffic for most businesses.
How to Optimize for SEO and AEO Simultaneously
1. Implement the BLUF Method (Bottom Line Up Front)
AI retrieval systems (RAG) and human readers both prioritize speed.
- The Tactic: Start every section with a direct answer to the user's primary question.
- AEO Benefit: Placing your answer in the first sentence makes it easy for AI models to "scrape" and attribute it to your brand as the consensus answer.
- SEO Benefit: This improves "time to value" for human users, a signal Google rewards with higher rankings.
2. Focus on Content "Freshness" and Decay
Both Google and AI models favor recent, verified data.
- The Tactic: Look at content published over a year ago that is currently decaying in traffic and find specific content gaps to refresh.
- The Result: Refreshing these gaps and republishing with new info helps you win in both Google rankings and AI performance.
3. Prioritize Site Speed and Technical Health
Speed is not just for users; it's a retrieval requirement for machines.
- The Tactic: Focus on Core Web Vitals to decrease "confusion" for bots and search engines.
- The Result: A fast-loading site allows LLMs to retrieve information efficiently and accurately without timing out or hallucinating.
Episode 31 Key Takeaways
- The 1% Trap: Don't overhaul your 40% organic engine to chase a 1% LLM traffic share.
- Elevator Down, Stairs Up: SEO mistakes act like an elevator to the bottom; recovery requires taking the long stairs back up.
- Research vs. Buy Intent: Users often research in LLMs but return to Google when they are ready to transact or buy.
- Zero-Click Strategy: Treat AI Overviews as a digital billboard; even without a click, your brand presence there builds essential trust.
Updated February 19, 2026
Tony Pataky: 0:00
And the biggest fear I think that the industry was having, had and also has right now, is that the LLMs are going to steal traffic, right, from Google. The data does not show that. In fact, Google usage barely barely changed, I think 1%, right? LMs are increasing, but uh LM usage is increasing, but Google usage, search usage is not decreasing, right? So maybe like 1%. So LMs are not killing search. It's just part of that search experience. It's just an added experience uh to how we look for things online, how we educate ourselves and everything else.
Kevin Kerner: 0:42
Hey everyone, this is Kevin Kerner with Tech Marketing Rewired. I just got off a podcast session with uh Tony Pataki, who runs uh SEO for a company called Propor. They make some uh, I guess it's automation around the construction industry. It's a fascinating conversation. Tony runs their SEO in digital. And I had gotten a hold of uh Tony on LinkedIn. I didn't know him, but I was in I was uh impressed with the work that he's done on the site and some of the things he posts in his newsletter. And so I got a hold of him. He was kind enough to give me some time. I was really curious about the hype that's in the SEO space right now. And is it real or not real? I have my own opinions, but um I wanted to get at actual operators' opinions on what's actually happening, um, what he's doing to uh maintain their uh credibility and ranking in Google and um just generally talk about best practices as we head into 2026. It was a great conversation. Honey uh is is has been doing this a long time. I think he said 15 years, and um just had the really good insight. So if you're a CMO or marketer who is thinking about their SEO strategy and your broader digital strategy, we get into some of that too. This would be a great one to listen to. So um, before we get started, this podcast is sponsored by my company, Mighty and True. Mighty and True is a growth agency that helps technology companies grow faster and also be more efficient. We have an automation and an AI part of our business that not only builds great marketing programs and strategies for customers, but we focus on also implementing AI and automation into their into our what we do and also what they do. So if you're interested in that at all, come visit us at www.mightyintry.com. We will do a strategic uh blueprint or gap audit for you for free and show you where we think we can improve your results as well as automate your business. So that's it. Um really excited about this one. Let's get to it. This is Tech Marketing Rewired. Tony, welcome to the podcast.
Tony Pataky: 2:45
Thanks for having me, Kevin.
Kevin Kerner: 2:46
Yeah, it's great to have you on. I I can't imagine all the topics we could be talking about. This is like one of the most timely, crazy topics that we could possibly be talking about. So I was we're very excited to get you on the podcast because I wanted an operator that actually sees us every day. So thanks so much for joining. Um I maybe I wonder if maybe I'm gonna I did a little intro at the start of this thing, but I wonder if you could just give a quick intro to your your role there. I think it'll give us good context. Uh, a little bit about Procore too.
Speaker 1: 3:15
Yeah, sure.
Tony Pataky: 3:16
So I work for ProCore, which is a construction construction SaaS tech company. Um, we build project management software tools and actually an entire platform for um for construction, basically. You can manage your entire project, your financials, uh, all through the platform. And I I'm there as the head of really the director of SEO and marketing performance. So I head up SEO, CRO, as well as analytics. Um and yeah, basically I joined the company through an acquisition. I used to work at level set. Procor acquired that for $500 million and have been working at ProQR ever since.
Kevin Kerner: 3:52
Uh yeah, wow, you're definitely in the right category. And isn't it interesting how like all these industries like construction and you know all these uh older school industries are now like getting AIFI'd and platforms and all the automations that are happening is really pretty incredible. Yeah, absolutely, absolutely. And it's interesting where in construction, if you if you take all the industries and you list them from top to bottom, and you you list them according to which ones adopted technology and software, construction would be the one before last. The last one is agriculture, farming, and hunting, and construction is right below right before that. So construction industry in general is kind of like late to the game when it comes to using software and technology, but even here, you know, like folks really leading the the the charge, so to speak, and bringing uh a platform, and it's we we also have a lot of AI within our in our software. Oh, yeah, bet. Yeah, it's great. It's like um we work with a lot of supply chain companies, and they're all the blue yonders and forkites and you know, redwoods and all that there it's incredible what's happening in that industry. So it's just really cool. I love your site experience, by the way. So you guys have done a great job on the site. It looks, it looks great. It's it looks like it would perform very well. Um and that's kind of one of the reasons I reached out to you was I was I wanted to get an operator on the podcast to talk about the end of SEO, which I know is not happening. Uh or that's if you if you were on LinkedIn earlier today, you would think that all this stuff is is going, you know, it's all going to AI and there's no Mark Evans at all. Um but I wonder your perspective on on uh what's actually happening on the ground. So I think this is gonna be a really good uh conversation and very useful for a lot of people. Maybe we could start with without going into the answer how to fix it, let's or what's you know, what what to do. Maybe we could start about the cur your uh insights on the current state of affairs, like what's going on with with uh web SEO search. Let's just level set for a second.
Tony Pataky: 5:56
Yeah, I I think I think to be able to talk about today, we're gonna have to go a little bit back into history because of what happened really 2025. I think most companies experience, most, if not every single company, experience traffic loss in organic, right? And that sparked this fear, this fear and anxiety that a lot of people are are feeling right now. And as you were mentioned before, like the death of SEO, LLM's coming in, you know, there's a lot of people jumping on this. And what really happened was 2022, 2023, Chat GPT exploded, right? It became super popular, everyone's using it. And everyone looked at Google, you know, basically asking, like, what are you gonna do? How are you going to answer this, right? And they came out with Bard and Gemini, but that's not what people were looking at. They were looking at their biggest product, search, right? How is search going to compete with Chat GPT? And they're already then everyone was talking about how LLMs are going to take over everything, and and then that's when AI overviews came in. So Q4 of 2024, November, December of 2024, was the biggest change in search and SEO in all of SEO history. And we've had a lot of big changes throughout. Uh, you know, I've been in SEO for 15 years, so I've seen a lot of a lot of stuff that's happened, a lot of panic and everything else that's going on. But Q4 of 2024 is when AI overviews came back, right? Because it it was launched earlier in May. That was and that was like Google's answer to Java GPT, but that was an epic failure, right? Like when AIOs first came out, it was hallucinations, bizarre answers, a lot of dangerous answers. So at the end of 2024, when Google brought it back, it's spread globally. Uh, and basically a lot of people started seeing a drop in traffic, 30 to 40 percent, organic traffic, right? And this caused panic, anxiety, and the the conversations around oh, everyone's moving to LLMs. Everyone's all these things started started popping up, and the the the industry, I'll call it that, um, through social media and everything else, have been have been like have been really um building it up, building up that fear. It's kind of freaked out, yeah. Freaked out, right? And I mean, if you think about it, it's like, yeah, people are seeing decl decreasing traffic, they don't know exactly where it's coming from, why, where is it going, right? So people are selling you know services such as you know, LLM services, how to increase your visibility, how to do the a lot of that's going on right now in industry jumping on this anxiety and this this fear uh that's being felt throughout the entire industry.
Kevin Kerner: 8:46
Yeah, no doubt. And um I think the the current state now, with uh uh ChatGPT saying that they're they're gonna add ads, and it's just kind of interesting. And then you have AI overviews becoming taking up a little a lot more real estate. Kind of uh continues to push the uh nervousness, I think, in the whole industry about where this is all gonna end up. I'm really worried about ChatGPT because I don't know how they compete with Google on the search side. I did try to, I did uh I can't remember what their browser was. Was it Atlas or something? I did download it at one point and tried to use it, and it wasn't as good as as uh Chrome, so I just kept using Chrome. But um and it seems like now all of the uh LLMs, I know Claude does this, has a sidebar inside of Chrome. So you're gonna have a plug-in and you can have click Chrome actually or Claude actually look at the Chrome page, but none of it really seems to have taken off yet. It's nothing is better than the search like bar and window, at least for me as a consumer.
Tony Pataky: 9:50
That's exactly right. And the data backs that up. So and what I mean by that is like LMs, they're great, and I love using them too, of course. I mean, I use them all the time. And the biggest fear I think that the industry was having, had, and also has right now, is that the LLMs are going to steal traffic from Google. The data does not show that. In fact, Google usage barely barely changed, I think 1%, right? LMs are increasing, but uh LM usage is increasing, but Google usage, search usage is not decreasing, right? So maybe like 1%. So LMs are not killing search, right? It's just part of that search experience. It's just an added experience uh to how we look for things online, how we educate ourselves and everything else. So yeah, it's it's been kind of interesting seeing of like really search is not dying, it's just expanding.
Kevin Kerner: 10:49
Yeah. And the other big hype moment, maybe for that I see on LinkedIn is a lot of um search experts, you know, kind of uh gurus around search, a lot of applications being built around search. Um, what's your perspective on anyone really having this figured out at this point?
Tony Pataky: 11:06
Yeah, yeah. Let's let's let's go into numbers because I think there's some hard truths here. Uh some hard truths that people don't want to talk about in social media where people are selling their services, right? And and the one is uh kind of like what we're just touching out right now. People are not leaving Google, right? They're not they're using Google sources, but they're also using LLMs. So so there are people moving to that, but they're using it for different things, right? If we look at if we look at the if anyone looks at the traffic, I would say mid to large size companies, on average of all their total traffic, less than one percent of it comes from LLMs. Chat GPT, perplexity, cloud, everything combined, less than one percent. Right. So these most of what's being sold right now is this unknown. And what I mean by that is they're trying to sell people on hey, we're gonna increase your visibility and your performance in LLMs. Right? This is what's being sold everywhere. Everyone's an expert. The fact is, there's no experts. There are no experts. We're all trying to figure out right now. If someone's claiming to be an expert, they're trying to sell you something. The the data that we have right now on what what moves the needle in LLMs is mostly observational, right? So far, there hasn't been anyone that I've seen, and if someone does have this, I'd love to see it. There hasn't been any actual causal impact analysis, actual statistical significant research that's been provided of like if I do this, this will happen. If I do X, this this will happen. And I've spoken to a few of these agencies, right, that that have been approached, uh, that have approached me with how they're you know helping different businesses and improving their visibility in LMs. Yeah. And because a lot of companies are shifting a lot of budgets, of everyone's falling for the hype. Gotta, gotta, you gotta go after Chat GPT, gotta go after Gemini, right? Like I said, it's less than 1% of traffic. So a lot of these companies who have approached me, not one has shown research, show statistical significance of what they're doing actually works.
Kevin Kerner: 13:16
Golly, that's it. That's so interesting. That's so interesting. I had on Jason White from uh who was an SEO, I can't remember what he's with. He was he was on the podcast. I'll have to look it back at a large insurance, I think, or financial institution last year. And he said the same thing. He's like, I can't tell you what actually is happening. He was looking at the law um server logs to see, you know, what at that time, this was a year ago, to see how many uh what the traffic looked like. I think now you can find the traffic easier. But same sort of thing. I just think it's um it's a black box. Uh these things are a black box. The other thing that I think it would be helpful to hear from your perspective is it seems like it's it's quite dangerous to recommend, or maybe an exact to hear an executive say or or someone say, let's say, let's pivot because of this big change that's happening. From your perspective, like what is the danger of some of the hype in the industry right now? What what's what's the harmful, what's harmful that you might need to know about instead of like making a hard pivot?
Tony Pataky: 14:26
Great question. Because I've been approached by I don't know, about a half a dozen people who I know in working at different companies and asking me, should I do this? I was approached by this uh company and they're telling me to do this so I can rank better, to so I could get more performance out of ChatGPT perplexity and all these other things.
Kevin Kerner: 14:46
Yeah, it's like the new keyword stuffing, basically.
unknown: 14:49
Right?
Tony Pataky: 14:49
So I can not just start. Well, the thing is that a lot of these recommendations that are coming out are not just expensive, they're huge overhauls. Um I'll give you an example. One of the one of the one of the most dangerous recommendations I've been seeing around is this extreme content chunking. Now, let me kind of go take a step back of like where where this is coming from and why it kind of makes sense, but also doesn't. The the content chunking is is just another, it's just a word actually being used. I'm seeing it now. It's just a way to rephrase um of having good content structure of answering the question as soon as possible, which has always been an important technique. Uh really it just good content writing is answering the question as soon as possible. So if you have a header, what does a general contractor do? Your friend's first sentence really should be answering that question, right? It shouldn't be a long story about your time at the cabin with your father, and then you know what I mean? Like you should get into it and then have supporting data right below it. Now, what's happening now is people are saying that you should have a header and then maybe two sentences, another header, two sentences, basically turning your article, your educational content, into a FAQ. And that has a lot of danger to it, right? If you do that, there's a good chance that your rankings in Google are going to go down. Right? If you redo your website and and uh you redo all your content, the traffic that you have now coming from Google is most likely going to be negatively impacted. Interesting. So this is where I tell people don't forget where your customers are coming from. Don't forget where a big portion of your business is coming from. 30, most likely for most of the companies or people who are listening to this, if you look at analytics, 30, 40% is coming from organic, if not more, right? Uh so are you really going to impact that to go chase after that less than 1% traffic?
Speaker 2: 16:49
Yeah.
Tony Pataky: 16:50
Right. And the problem is when you make a mistake in Google, right? If you make a mistake in SEO, you take the elevator down. And when you fix that mistake, you take the long stairs all the way back up. It takes a long time before Google gives you credit.
Kevin Kerner: 17:04
It takes a long time. No matter how many times you hit re-index, it takes a long time to come back up. There's no way. Um the other I'm curious about this because we you uh you um You know, there is a lot of site restructuring going on and stuff. Are there other things that people may be trying that aren't a good idea? Like I've heard about like the one thing I've heard is a lot of agent-to-agent talk. So our website will eventually become something that agents talk to. And so you'll have agents that are buyer-led agents, and you'll have the agents that are seller-led agents, and those will talk, and eventually they'll get to know each other, and that will be the way that we transact on the web. Is it too early, or is it that a bad idea to even be thinking that way?
Tony Pataky: 17:51
It's so good question. Like the I'll kind of go through what we know now because it does tie into this. Um, so the I would say a lot of the stuff that's happening right now, and this kind of ties into this, uh, comes down to uh what we know of how how the LMs are going and then retrieving data from websites, right? So RAG is uh is a big term. Retrieval augmented generation. It's the framework of how LMs go out. When you ask when you ask a question and you put in a prompt into ChatGPT, if it doesn't have any training data, it's gonna go out and it's gonna technically Google it for you, right? Or go to being as well, and then retrieve that information. Now it's trying to do that as efficiently and as fast as possible. And this is why a lot of this stuff is coming out of like content chunking and everything else, simplifying your content so that it can actually do that. Comes to your website, grabs it, the answer's there, doesn't have to do anything, doesn't have to do major vector analysis and stuff, pulls the information, goes back and gives you the answer. So the I would say that we're it makes a lot of sense uh of of the directions that that are gonna be positive for you, for for your for your current customers as well as for LLMs, are gonna be like your site speed, right? Your site speed should be fast for your users, you know. Let's start there. You're cut don't don't forget your customers, right? But having a fast site so that when LLMs come and they try to retrieve an answer uh or information, that's gonna be important. They want to be able to do it quickly. Your content being structured in a way so that it can look at the content, get the answer, go back is also important, right? This is this comes that's that bluff method of bottom line, up front for organizing content. These things I would say definitely, those those are those are good approaches to do. Uh and we know they work in Google as well, right? Having your answer up front does help um with your overall visibility. So there are these there are these things that that um I would and there's other things we'll talk about as well of like of recommendations to go to, but I wouldn't go to any kind of big overhauls, right? Yeah. Even though LM is a very small party traffic, I'm not saying to ignore it, don't ignore it. It's there, right? It's it's important. But if you're going to tackle it, do small iterations. Tackle it smart. Don't do big overhauls. Don't try to predict the future. So much has changed in the last year. And that's why, like when there's people who are experts and hey, you know, we have all these case studies. That's it's it's simply simply not true. It's really not you can't really be an expert in something that's changed somewhere.
Kevin Kerner: 20:29
So far. Yeah, so it really has changed in the last year a lot, but it really hasn't changed in the last year a lot, too, because because the you would think the with all the hype on this stuff that LLMs would be be now driving a significant part of of traffic. Now, now I use an LLM as product discovery and I use it for research and those type of things, but I will always go to the web and search for things. So yeah. It's like it is changing. Like the buyer behavior is changing. It's I'm probably way far ahead than than most people in their use of you know the uh AI, but it's it really is uh it is changing fast, but in a lot of ways it isn't because you know, search still dominates, it's just the thing.
Tony Pataky: 21:14
You actually touched on a really important point here, is like with LLMs, right? I I love using it too. I use it, like I said, nonstop. But usually when we use LLMs, we're not using it the same way we we use Google, right? For LLMs, we usually go there to do coding or refresh an email or whatever it might be. Um, but when we're looking to buy something, right? If you're doing the Go stuff, if you're gonna buy, go somewhere, whatever. Generally people go to Google, right? Yeah, it's Google in your industry. For sure.
Kevin Kerner: 21:41
In your industry, too. Think about the buyer, construction buyer.
unknown: 21:44
No.
Kevin Kerner: 21:46
Exactly. Well, what I liked about the conversation we had before, the pre call stuff, is you we were you were getting into fundamentals and things that you actually should focus on that are that are going to have some effect on your Ability to attract BLMs, but really the things you should be doing. I wonder if you could run down the four or five things that you definitely need to be doing now. Um, a lot of them are fundamentals, but um you know, it'd be helpful to hear from your perspective, like what you're doing. Um and uh if it has any impact on the AI and LLM stuff, talk about that too, but I'd like to hear your best practices.
Tony Pataky: 22:26
Yeah. And it I want to say, in a way, like if I was sitting down with the company who are like, what should we be doing right now? Because, like I mentioned, there's a few people I spoke to in the past, and they were gonna do some pretty serious overhauls until um, you know, put a pause on that basically. The fact is, what I'm gonna start with this. What we know about AI now, right? Because like I said, don't ignore it, you can still work towards improving your performance there. Um, but there are things that you could do, and you could do it safely, right? But what we know about LMs right now of of um the the correlation analysis that have come out is, for example, brand mentions are important, right? Fresh content is important, um content structure is important, right? These things we know, it's correlation, but we don't know if you do this, how much is it gonna move the needle? We don't know actually, even measuring how much it moves the needle is a tough one. So going back to your question and just answering straight of like what should people be focusing on now? Um as I mentioned, LLMs, what we know about it is really a combination of really good SEO mixed with really good marketing, right? Brian's good marketing and the rest of it, like the SEO stuff. So site speed, take a look at your site speed. That's an easy one. I would definitely focus on that of just just you know, fix those things, look at your core web vitals, go through that. Um, you know, I that's that's gonna be beneficial to your users as well, right? You're hitting two things at once, both Google users and also you know AI. Um I'll take a look at your old content, right? Stale content, fresh content does quote unquote win in LLMs and AI. And this is this is actually just good practice too, anyway. Take a look at your content. Anything that you published over a year ago. Look at the ones that you haven't touched over a year ago, and look at the ones that are actually decaying. Now revisit those, find some content gaps in there, right? The search landscape has changed. Your competitors have built better content and everything else. Take a look at some of the gaps. Uh, take a look at some things that you could brush up on, republish it. Now you got fresh content out there with new information that's gonna be helpful. You're gonna see both Google rankings go up as well as most likely AI performance uh with some of that. Um, and we'll talk a little bit about AI performance measurement because I keep tossing that around. Look at your content, look at the content structure, use the bluff method, right? Don't don't bury to leave. Answer the things up front. These are things that these are this is just good SEO best practices, right? That's also gonna be helpful in for AI and LLMs. Um what else? Oh see, these are big ones. Strategic internal linking, helping Google find your content, users find your content, right? Building those educational funnels within your website, that experience, that's important. It's gonna help, like I was mentioning RAG before. Make it easy for them, but honestly, for your user, think of your customer first. Um, that's what's really important here. And then last but not least, if you have time, do some like schema markup, these kind of things, right? Schema, it's not gonna hurt you, right? There is I'll say that the research or the data on how important schema is for your for website is and and AI is kind of shaky. I haven't really seen anything great on there, but technically schema markup tells bots, search engines, really robots, what your content is about. It's easy for them to be able to go through your content and know, okay, this is an FAQ, this is a product, whatever it might be. So make it easy for them to know what it is. Because really, when it comes to SEO, a big part of our work in SEO is decreasing confusion to bots. That's a big part of it of what we do. We decrease the confusion for bots, but most importantly, connecting your business solutions to users who are looking for a solution, right? Like that's the primary. But we do that through just decreasing confusion, right? Don't don't let them don't let them kind of make a decision. Make make it for them of like this is what this is.
Kevin Kerner: 26:30
Yeah, such a man, it's just such a no-brainer to what it to what the traditional marketing, to what marketing looks like, right? Making it obvious to the person, the human, but then now we've got to make it sort of obvious to the bot as well. I want to get into measurement because I think that's a super interesting area. But before I do that, I wanted to ask you about there's a lot of activity that needs to happen as an SEO sort of marketing professionals or analytics professional. Have you found any, are you playing with any automations that might make some of the work that you're doing easier? Uh it can either be um synthesizing data or creating new content or um the new way to analyze something faster. What what are you using, if anything, in terms of just automating the whole SEO process?
Tony Pataky: 27:21
Absolutely. Yeah. So the the SEO team, um see, my SEO team has been doing amazing work around leveraging AI to help you know, help with our workflows and just all the work that we do. Uh Gemini, for example, creating gems for specific tasks, looking through content, what you know, looking through content, coming up with content ideas, um, analyzing data, right? A big part of our work is really number crunching, looking at what's happening, what's moving the needle, what's not moving the needle, creating forecasts. Um, but AI has helped with a lot of these repetitive tasks that take up a lot of time. So reviewing content, coming up with content ideas, even with outlines, right? When you give it the right data, it can come out with some really good outlines. Um, one thing I would tell people not to do, and there's a lot of like, there's a lot of research around this as well, is using AI to create your content. It's I'm saying this because Glenn Gabe, if Glenn Gabe, who's on LinkedIn, he's one of the leading um, I would say SEO thought leaders out there. He's put up some really good examples of um what he calls Mount AI. And what Mount AI is, is he'll show you a graph of rankings and traffic um pulled from SEO tools of different websites. And the traffic looks like a mountain. It goes up on one side and comes back down the other. And that fast growth and that fast drop generally came from companies who kind of felt fell to the temptation of scaling out a bunch of content using AI. When you scale out a lot of content using AI, we've all run into hallucinations, right? Hallucinations, bad answers, whatever it might be. Imagine creating 10,000 content pieces just using AI and publishing it. You're gonna have a bad time. So, and that's what these companies were were running into. So I would not use it for that, but a lot of these other things of strategy, building out outlines and plans, AI has been amazing. Number crunching, fantastic. Doing analysis of content gaps between competitors, um, using gems to just do that analysis for you saves hours of work.
Kevin Kerner: 29:34
Yeah, custom gems are great, especially when you extend them to a team. It's just amazing because you can train them. Have you moved to have you moved anything, or do you have any plans to move anything to be a little more agentic where it can actually make changes for you? Or are you guys not there yet, or would you trust them with an AI to do that at this point?
Tony Pataky: 29:53
I would not. So I would. Great question, by the way. I love that question. So the the I would not allow an AI agent to go make changes on the website. Um, I I do believe right now AI still needs a lot of critical thinking. It still needs someone with critical thinking to look through it and be like, that that probably doesn't look right. Um but what we have, what we are working towards is building AI agents to help with, for example, technical audits, to identify issues, right? Identification that takes up a lot of time. As I mentioned before, site speed, this AI agent I just built not long ago. That's been that's that's something that's that's um that's gonna be very helpful moving forward to make sure, because the fact is, is like companies out there, when you do rebranding, make changes to the website, whatever, your site speed is going to have it's gonna take an impact, right? And it's hard to stay on top of that. So site speed optimization is always an opportunity. So things like this, where it takes up a lot of time and that we want to do over and over and over again, we are building AI agents for, and there's a long list of things that I have on my plans for.
Kevin Kerner: 31:02
So that's cool. Yeah, that's super cool. I do a lot of work in a tool called N8N, which is an automation tool. Um, and uh I've hooked it up with my GitHub and my um clock code. And it's just amazing what you can build in in terms of your own internal apps for all kinds of stuff. And it but I have not let it go. Uh, the other thing I've I've been doing a lot of is um having the agent work, but then it'll notify me in Slack for you know whatever. Check this out, look at this, approve this, and it'll just send me lots of Slack notes for the stuff. But I would never let it change anything without looking at it because it could go haywire and I would be like just be over. And a lot of these people that are using uh CloudBot or those tools right now are that those things scare me because it just it's just not ready yet. I love the idea of autonomous agents, but I haven't seen it work in any business case, right? Maybe personal.
Tony Pataky: 32:04
But I I'd rather wait and let someone else.
Kevin Kerner: 32:07
Yeah, right. It's your industry, right? You don't want to let it go. Um, measurement seems like a uh it can be both um, there might be some changes, but some of the measurement um challenges and all this stuff might, you know, m make it more confusing for people. How do you see measurement? Uh you know, good solid SEO, digital measurement right now, and then also if you put the AI lens on top of it, how are you thinking about things?
Tony Pataky: 32:34
Yeah, I mean, it's one of the biggest, one of the biggest, I would say, mysteries or biggest data points that's missing within the whole LLM optimization thing is you know, we don't get we don't get the search traffic, right? We don't know what um, sorry, what I meant to say is keyword volume or prompt volume. We don't know exactly what prompts people are using, how often. Companies are backing into that using SEO tools, and they're like, okay, well, if on Google someone's searching this keyword 10,000 times, we know that people are using ChatGPT at like 5%. So we're gonna take 5%. You know, they're doing this math backing into uh into prompts.
Kevin Kerner: 33:15
I wonder if Goop, I wonder if Google has that data and they just don't it's too too creepy to reveal it. But yeah, that's a that is that is difficult. We don't you don't know what that is.
Speaker 1: 33:27
Yeah, you don't have that. You don't know what people are searching for and how and everything else. I mean we're uh people are getting better at it, and there are people who are collecting some of this data, but fact is like that's that's a big missing piece. But the other one is there's no rankings technically, right? If you ask your LLM the same question over and over, you're gonna get a different answer every single time.
Tony Pataky: 33:47
Right? Is it gonna give you a different answer because it's not static, it's dynamic. So there's no like, oh, you ranked here and you're gonna rank there tomorrow. Or if you search for it and I search for it, it's gonna use the memory to kind of also influence how it's you know, what kind of answer it's gonna give you, especially if you gave it some kind of um, you know, basically guidance or directive of like make sure to answer questions a certain way. But memory, so your results and my results are gonna be different, and it's gonna be different for for for pretty much everyone. And like I mentioned, one thing the LLMs will not do is tell you, I don't know. It's gonna give you an answer, right? So you can keep asking the same question, you're gonna get different results, so there's no rankings, you can't really. So what companies are selling right now is visibility, impressions, quote unquote, of hey, for this prompt, you show up, right? And you show up, and then here's your brand sentiment. It's it's positive or negative, right? So when someone's typing in best construction software tool is pro corbeing, yeah. So that's the current measurement, and there's a lot of issues with that because well, okay, what if I show up for this prompt that you have selected, right? Like if I would be speaking to an agency, you have selected this prompt, I'm showing up for it, right? How many, how much money am I gonna make? You can't tie it down to it to the dollar, right? And you you can see visibility almost like you're you're basically making up the prompts of what you know what you want to track. You can see the traffic coming from just Chat GPT, but it you can't really say, when I did this, this is what happened. You can't not like an ISCO. I know this moved up three positions, we're getting more traffic to that page. You know, you can connect the dots. Right now, connecting dots and LMs is very tricky. And here, if the thing is, like these people who are selling their LLM services to optimization services are kind of winning in that, meaning that, you know, hey, I'm gonna optimize you for these prompts. The customer doesn't know like if this is a good, no one knows if that's good or bad, right? It's and so that's the dangerous part of the whole thing.
Kevin Kerner: 36:00
Is so many spoken like a true um marketer that you can't, you know, like, oh yeah, we get this there. People are asking that question about you. So good you're showing up, but you're you're comment about did it actually sell anything? Did it actually result in anything? It was good. I mean, and the funnel is so crazy right now because it's you you may show up there for that at some point, but then you know, six months later, you might show up somewhere else. And it's it's just I think for brands, it's a it's a good thing to be showing up there, but the just the buyer, the buyer's journey, how you get from an initial hearing about a brand all the way to buying something is just super complex right now.
Tony Pataky: 36:43
So you you bring up a great point, that pre-click journey, right? That's just such a huge unknown. And and look, fact is people are going to LMs to research companies and then they go to Google to buy. Connecting that journey right now, we can't, right? We can't see that journey. So I understand of like how important it is to be in LMs. It is, it's important to be there, people are using it and everything else. Um, and a lot of these LMs are using things like Reddit, looking Quora and other sources to get consensus when you ask for things like best CRM software or whatever it might be. So being in those places are important. Um and I don't want to say that it's not, you know, it's it is important. Measuring it is is really tough. And that's that's gonna be once we're able to measure that, that's gonna be a next next big thing, I would say, within within AI search and just user journey from LLM to sale, that will be a real game.
Kevin Kerner: 37:37
I don't think it'll probably happen at some point. I don't know, someone will figure that. It's really the data, it's just behind such a black box right now. You know, who knows how that's gets solved for? Um, I'm wondering uh finally about you since you had a purview that's across a lot of digital channels, do you have any um thoughts about other channels? You know, you've got search and then you've got let's say you've got uh SEO, but then you have search. And what's the implications for all this stuff on search and promoted posts? And just how are you thinking about all the other channels that might surround the funnel beyond SEO?
Tony Pataky: 38:15
Yeah. I mean, when you asked that question, the first thing that where my mind went is, for example, ads coming to ChatGPT. I haven't seen them yet, but yeah, but how paid actually it's gonna be interesting.
Kevin Kerner: 38:29
I don't know how they'll show up, and who knows?
Tony Pataky: 38:31
But yeah. I I really liked Claude's response to that during the Super Bowl. Those ads, I that was pretty good. It's there's a lot of changes happening, and and and so AI with some of the other channels. I think that AI is fantastic at helping with copywriting and some emails and obviously giving direction and not just copy-pasting it. But there's a lot of opportunity um for using AI and then other, I'll just say, marketing motions, so to speak, like whether you're A-B testing, doing paid activities, where you can test out a lot of things very quickly, right? I think that's one of the that's one of the biggest things I'm seeing being used a lot through different companies of of being able to just test a lot, coming up with new ideas very quickly, and just you know, could go crazy there and just get on. Black marketing, yeah. Yeah. Um, so yeah, it's it's just being infused with everything, which is awesome. I love the fact that AI is everywhere. Yeah, but kind of like what you were asking before of would you let it change anything? I think that's all that's the part where I'm always like telling people to be a little cautious about.
Kevin Kerner: 39:43
Yeah. I've heard some people say, like I've had conversations with people that just say cut cut our search budget entirely, because you've got all the AEO stuff, and you know, who you know, let's say branded search is a topic, just cut it because you're already getting to us somehow. I'm like, man, I don't know, I don't know if I'd do that. It seems like search is even more of a useful tactic now because if you do have the AEO showing up with that much real estate, you got to show up somewhere, and your your actual organic site might be pretty far down.
Tony Pataky: 40:19
Let's stay here. Yeah, actually, that's so great topic, yeah. So yeah, with AI overviews coming in, a lot has changed there, right? And you're exactly right. In in in Q4 2024, when AI overviews showed up, what we also noticed is like, hmm, these ads seem this ad real estate seems a little bigger than what it used to be. Uh, it might not be right, but I'm like, you know what? If you go now to Google and you search something, above the fold are just ads. And then you'll get AI overview, and then below that is organic. So people who are stepping away from uh so what this is causing is that people, if they know where they want to go, if they know where they want to go, if they're typing in a branded search, they're more likely to click on the ad. Because people don't like to scroll. Interaction cost, it people are super sensitive to interaction cost. It blows my mind when we do A-B tests and we just move something up just a little bit, and it's like, oh, it's getting a lot more engagement because people weren't didn't want to scroll the extra amount. It so don't if if I were to give a friendly advice to people, uh, I would not give up my my uh my paid ad space right now, even though it's showing up in organic. I would I would be very, very cautious of doing that um because it has changed how people are interacting with the search engine results page.
Kevin Kerner: 41:38
Yeah, super interesting. Yeah, because you you don't know how many times you have to scroll. I I would encourage every CEO and you know, senior executive to Google their category and see where you end up, and then Google your brand and see where you end up. And you'll be very happy to see that there's a promoted post with your brand name at the very top above the AEO. You you might you because you might have to scroll down, and there might be a bunch of other competitor ads, and then you got the AEO, and then you have the your brand, uh or your organic brand. And it's just it's the territory is so competitive, and the space is smaller now to be able to be able to rank. And so it's like certain. And maybe Google did it that way. Maybe they did it though that way because they it allows them to sell more uh ads, you know, because they you know, there's so much AEO space. I don't know. They're pretty good at making money. I wouldn't be surprised.
Tony Pataky: 42:37
You know, if I put my Timfo hat on, I'd be like, ooh, there's they did complain about not hitting so smart.
Kevin Kerner: 42:43
They're just they're so smart. Not be they're not evil smart, but they're just so smart in how they make money. It's just incredible. And you know, I uh all hats off to you. They're pretty incredible.
Tony Pataky: 42:55
I mean, even with AI overviews, and we were touching on this earlier, AI overviews is so huge now. So when you type in a search term, um, you get an entire article right there, right? So this increased that zero-click search uh tendency that people have where they search, and they don't have to click on a blue link. They're just like, I have my answer, I'm gonna go back to doing what I was doing. And that article is huge now. It used to be 2,000 characters back in May of 2024. Now it's over six, seven thousand. I don't know, like it's there's YouTube videos embedded as well. So that real estate has become extremely important. It's it's a huge billboard, technically, because if you go there, right, you open that up, you see the little like logos, little companies being mentioned. You want to make sure that you're there. Um, but that's where the bulk of the traffic that people saw decreasing in 2025, that's where it went. People getting their answer, quote unquote tire kickers, right? They got their answer there and then they went back. But what a lot of companies also saw in 2025, traffic went down, but conversions engagement went up because that valuable traffic still came through. The people who wanted to do something. something still came through. And I I would tell anyone who's listening to this, go look at your data. Most likely you're seeing that exact same thing. Yeah. So yeah, AIO has has really changed how the SU industry is doing forecasting and looking at traffic moving forward because it's I mean it changed the game. It changed the game and and how you can actually rank, right? There was there's a lot of I could go on for hours on on how they changed what you can and cannot rank for ever since Q4 of 2024. There's a lot of big changes. But yeah, don't sleep on AI overviews. That's a really important part to be in.
Kevin Kerner: 44:37
Yeah I completely agree. Okay, so I wanted to um I wanted to do one thing before I let you go uh we do an a uh AI roulette question that um and it might be it might this AI roulette question I do it through Gemini now. So hopefully Google hadn't been listening to this conversation and just torch us both on the on the uh AI roulette question. We'll see here. So let me so I put it I put your profile the outline of what we're going to be talking about and then I just hit send. So let me hit send and see what it gives me. This is an AI question for you. Oh this is interesting. If you had to bet the entire ProCore organic budget on a single gut feeling tactic that lacks any MSV data or traditional attribution what is the one unquantifiable marketing bet you would take today so you don't have any data you got to make a bet the whole budget entire budget what would it be?
Tony Pataky: 45:37
Gosh no pressure yeah no pressure right oh man unquantifiable well there's a lot of um man that's a tough one yeah uh entire budget look I I'll I will go back to this is it's just content like right the the problem is it is quantifiable but con building good content helpful content and looking through your website and adjusting your content to be helpful I think is isn't always it's a winning strategy right you're thinking of the customers first but you will be rewarded in Google as well as AIs um because you know Google will give Google ranks those that that do well from a UX perspective right they Google ranks those uh that that people want to see is really what it comes down to. So that's right.
Kevin Kerner: 46:36
I would focus everything on just just content and content structuring and um yeah that's just I would encourage everyone to go look at the ProCor site because it's it's really the uh information architecture and how you go through the site I looked at it earlier today it's just it's just really well done. So you can tell you guys have put a lot of time into this and you have a professional team that runs it all because it just doesn't happen on accident. So well done to you and the team there. It's really good stuff.
Tony Pataky: 47:05
Thank you. Yeah a lot of teams are involved in that it's it's it's a never ending process right every time you make a big change you gotta you gotta update everything.
Kevin Kerner: 47:12
So yeah a lot of teams are involved in that excited to see how it all I'll have to try to do this in a year from now and see what happens. I noticed that you had um when I was on your LinkedIn I noticed you have a newsletter if I'm not mistaken. Yeah so I just subscribed for it so I'd encourage everyone to go and subscribe to it. It's I am definitely gonna to read it because there's there's this space is so crazy. So maybe the new the name of the newsletter how do people get a hold of you if they want more information.
Tony Pataky: 47:40
Yeah thanks uh so the newsletter is called search velocity you can if you visit my LinkedIn page it's it's up there um but basically what I'm I'm just basically writing about what's going on right now how to how to move forward in this search space right I'm basically just sharing my thoughts for free you know I'm not I don't I'm not I'm not a consultant I'm not I don't work for an agency so I'm just getting into the conversation and sharing it from my perspective uh and it's good to see that there's a lot of people who are benefiting from it. So I like I like the fact that people actually like it and I love writing about it. I love talking about SEO I could talk about it for hours. So I'm gonna continue doing that.
Kevin Kerner: 48:18
So yeah you're very generous on on your time here and and it's and it really is all about community at this point because we're we're all trying to figure this stuff out. This is the biggest rewire of of my career and I've been around a lot for a long time. So it's great that you're sharing stuff and and uh are willing to do that. So I would encourage everyone to reach out I'll put your name in the newsletter in the show notes and everything but I really appreciate you being on Tony it's been great to uh to talk to you.
Tony Pataky: 48:44
Thanks Kevin I really appreciate you having me on cool talk soon
Guest Bio
About Tony Pataky
Tony Pataky is the Director of SEO and Marketing Performance at Procore. He specializes in decreasing bot confusion and connecting business solutions to high-intent users. Tony joined Procore through the acquisition of Levelset and shares his unfiltered search insights through his newsletter, Search Velocity.
Connect with Tony:
Conversations That Rewire Marketing
Explore more unfiltered insights from the CMOs, founders, and disruptors reshaping how tech brands grow.

%20-%20Web%20Tiles%20Small%20(616%20x%20308%20px).jpg)