How Data and AI Are Shaping PPC and SEO

Speaker A: It's. Mhm.

Speaker B: Welcome to the Deconstructing Data podcast and live stream. I'm Jesse Lisak, fractional CMO at BDEX. Along with David Finkelstein, we've got BDEX's founder and CEO himself. Um, what's new this week? David?

Speaker C: Hey Jesse. It's been a busy week I have to say. I'm going to give a shout out to Bordy because anyone that's listening or watching, if you haven't heard of Bordy, look up 40 on LinkedIn because he's been filling up my schedule lately with a lot of introductions and really valuable introductions. So um, it's been great, it's been really interesting and the new version Boardy Pro has been and like managing my calendar and adding some really interesting introductions. Uh, so um, shout out to Bordy and uh, I don't know, it's been, been a busy week.

Speaker B: That's good. Well we like Bor. Um, and we like being busy too so uh, great to hear. Um, yeah, I'm going to Tag Bordy on LinkedIn and here just so other people check it out because I agree it's super awesome. Um, but speaking of awesome, we have another pretty awesome guest on Deconstructing Data today. Um, today we are welcoming on the co founder of site Smarts. Welcome to the, I'm sorry, smart site. Welcome to the show Alex.

Speaker A: No worries. Thank you, thank you for having me.

Speaker C: Yeah Alex, we're excited to get this conversation going but uh, appreciate you joining us on the show today and we'd love for you to kick it off for us. Let our viewers and listeners get to know you, a little bit, a little bit of background about yourself and what led you to co found smart sites and what you guys do for your clients today.

Speaker A: Yeah, yeah. Thank you, thank you for handing it off. Um, yeah so I co founded Smart Sites with my younger brother 2011. So it's been, I don't even want to do the math, more than 15 years now. So it's been, it's been a while. Um, I've personally been involved in digital space much longer than that. I actually started my first company in the mid-90s, incorporated UH in 1997. I started one of the first web hosting companies uh, which uh, for those of that were around back then that that was the age of uh, Tripod, geocities, Hypermart. I don't know if any of those names ring, ring any bells to people but I've been involved in digital space for a very, very long time. Um, been Doing the smart site thing. The smart sites thing uh, for 15 plus years now, where now I think we are the largest private digital marketing company in the U.S. um, a couple of our competitors took on equity or sold. So I think, I think uh, last time I fact checked myself it was still true. Uh we have over 400 employees, um, work with thousands of clients, uh, and everything digital, make websites, SEO, paid search, pretty much anything that touches anything digitally. We're there to help small and medium sized businesses be successful with it online.

Speaker C: That's awesome. I love that journey. And yes I am familiar with all of those companies you mentioned. Um, I was, I uh, started my first Internet company in 94 as an Internet service provider. So uh, I uh, was definitely uh, in a parallel space with you back in those days as well. Um, let's uh, jump right in um, and kick off uh our first topic. Using data to optimize pay per click performance. And so being a data guy myself, um, and this being deconstructing data, would um, love to have you sort of kick this off and give us some insights on like how companies can be using data to optimize their, their performance.

Speaker A: Yeah, that's, that's a good one. So uh, paid search, what's interesting. So digital marketing, what I really like about digital marketing and I briefly before I done a million different things but briefly before uh, starting smart sites I worked at Publicis, which is one of the biggest advertising agents. I think now they are the biggest, um, I think back then they were one of the biggest. Um, and they did all kinds of media, right? They did digital, uh, was always the, at the bottom of the totem pole. Um, I was in charge of digital for Walmart for a little bit Samsung before that. And um, this, this wasn't the 90s, this was 15 years ago. Uh, but digital is always the lowest on tone pole. And the biggest media was always print, tv, radio, right out of home, those kind of things. And as powerful as those are, they are really not data based. And what I really liked about digital and when we started smart sites, the reason we focused on digital separate from my background digital is that it's very database and very measurable in a very good way. If you think about the other mediums, uh, I don't know, tv, tv, uh, uh, TV advertising. Right. Which is still very, nothing to say about tv. There's still value to it. It's still a huge, huge channel. You think about TV advertising, um, as much as it tries to be data driven, it really isn't uh, the Data for it is, I don't know, Nielsen estimates viewership by a couple of people having boxes and then they project that out. Um, and whether they actually saw the ad, didn't see the ad, uh, did perform an action after the ad. It's really not trackable at all. Um, so paid search is the opposite. I think it's one of the most powerful advertising channels because of the data angle. What's interesting is as we are in 2026 now, um, the data component I think could be split in almost two different ways. One side of it is the data that is natively feeding into a platform and the platform uses itself. So this is the machine learning that Google Ads, Microsoft ads does. So that's 100% very data driven. Still needs to be set up correctly and still needs to be running correctly and it still needs someone overseeing it. But this is the data the campaign itself collects. This is the data from your website feeding into the campaign. This is setting up actually values on your conversions. It's setting up. Some of these things seem simple, but it's really the only way the data will actually do something useful. So that's I think half of the Data equation in 2026 and moving forward, the data that's natively uh, uh, feeding into the machine learning engines, right, it's your data but you're feeding it to Google, you're giving it to Google, you're giving it to Microsoft and you're supervising it and making tweaks, but you're giving it to make a lot of on the fly decisions. Um, Google calls it auction time, bidding, um, for those who are not familiar with it, when you search something on Google in the fractions of a second after you press submit and before the ad learns, uh, the ad loads. Google uses the data to determine what ad to show you what's relevant to you. There's so many things that happen. I think that's one piece of the data equation for paid search and probably the most common one that people think about. The second one is actually more interesting to me. What's been happening now is there's this whole second data component of the data that you have not only of yourself but that you could have from others. That becomes very relevant for us as an agency. And I'll give you guys an example. Uh, we've been very AI forward. We have AI enabled for everyone. We're cloud centric. So uh, we use Claude a lot and Claude is hooked up to all of our data sources. So for us as an agency, let's Use uh, dentists as an example. Uh, we don't just manage one dentist, we manage thousands of dentists. Um, and we're hooked up to the data sources for all of this. Google Analytics, Google Search Console, all these data sources. And what's funny is I mentioned some of these things, uh, at conferences in Europe sometimes and people like freak out. You have all this data that you're allowed to access us. We are. And we have uh, even clients who canceled, who never unlinked us. So we have this huge, huge data set. So how this is different from the data you're giving of um, your own website to Google is this data that agencies and other companies can now access. And the way it works is, for example you're a dentist and you're feeding all of this data to Google Ads. Everything's optimized really, really well. Google Ads is doing machine learning, it's bidding really well, but it's just not performing well for you this month. Right? It's a slow month for a small business. That's probably the most common complaint they'll ever have. It feels slow. They'll say, right, anyone who works on the agency side or works with small businesses, they probably feel slow a, uh, million times a day. Um, uh, with the data that we have we could actually, the PPC analyst who's managing that dentist could literally go into Claude and say my client is having a slow month. Is it slow for everyone? Is it geographically slow for those that's not slow for what are they doing differently? Right. So that becomes a second part of the data that perhaps not businesses uh, might not have themselves. But I think that's what gives agencies like us a lot of uh, a lot of advantages and a lot of more uh, insights. So I think data for paid search is really going to come down to those two things and I think they're very separate. I think it's the data that you own that you feed into Google and the data that you are able to get, agencies able to get or third party data or things like that that could help you make more educated decisions about your marketing.

Speaker C: That's a great point. And that's a really interesting data set that aggregated data across an industry is super valuable. Um, I, I spoke to somebody this week actually that's doing something similar in the E commerce space where they have a um, I think it's like a Shopify, um, app and it's doing something and it has one purpose but the sort of secondary purpose of it was to totally like get this aggregate data about what people are doing. You know, how people like the journey that people are making through their e commerce journey when they make a purchase. And that, uh, aggregated data is really interesting when you really start digging into it and understanding what, you know, people are doing as a whole. You know, whether it be you can look at it from the entire data set or then to your point, you can sort of separate it and say, okay, what's happening in this industry, um, versus this industry, or what's happening in this geographic, uh, area with one industry versus another geographic industry area in that same industry.

Speaker A: What's interesting. Yeah, and what's interesting is we kind of had this data forever, right? It's not all like all of a sudden we have more Google Analytics data, right? We always get, we're always growing, we have more clients, but relatively speaking, we always had tons of this data. But there's no way, no easy way to analyze it, right? I don't know. We have Google Analytics access to 100 dentists. How would you figure out if it's slow for everyone or just a specific region? How would you figure out if a certain channel is working better for some than a different channel? How would you figure out if all of a sudden organic search is slower for them because AI overview is taking over organic search volume? You'd have to spend like months, right? Uh, I think the big push in being able to use more data is AI and Claude, literally within two minutes can analyze all of it. And uh, it could even tell you that you could ask it, you could say, for these 100 dentists, are they getting less organic search traffic because of AI overviews? And it would be able to tell you that because we have Google search console access. So it could tell you the impressions, clicks, click through rates, things like that. It could tell you search information before they click to your website for Google Analytics to capture. And it could marry that data with Google Analytics and could give you such rich insights in like seconds where it would take months before. So I think that's the big, uh, that's, that's the big change that's happened and I think that's, that makes data so much more valid. It's one thing to have tons of data, right? But if you can't really use it. And yes, there are dashboards and reports and there are companies that help you combine it and show it to you in different ways, but to able to ask questions against your data and have, uh, a logical response after an analysis within seconds, I don't know. Definitely not minutes. I Don't know, fractions of a minute. Um, I think that makes all the difference. And, uh, it's funny, I don't see it being utilized, uh, as much as I thought it would for us. Um, I was on another podcast and the question was, what's your competitive advantage? I think our competitive advantage is literally having this data and being able to ask questions against it. Um, a lot of people have this data, though. It's not like only us. It's proprietary. Not proprietary at all. Uh, it's not like, um, I think there was another question of did you build proprietary AI stuff or whatever? I'm like, no, not at all. Uh, we use CLAUDE for everything. So it's all out there for everyone. It's just people aren't creating the processes and connecting it and all of that to use it. Even for a small business, let's say you don't have the access to all the data we do. Let's say you're a small business. Right. Um, connecting Google Analytics and Google Search Console to your Claude, uh, could get you such rich answers, uh, about your search activity. Um, literally based on that data, it could suggest what content you should be writing. Um, it could show you the impact of AI overviews. Um, it could analyze competitors in your market and what they're doing correctly. It's all this really cool stuff that even a business with smaller data set could be doing.

Speaker C: Yeah, but I think you might say that that's kind of one of the reasons why people work with companies like yours, right? Because not just because you have the unique data set necessarily. Um, um, but because you have the unique insight, you know, from an industry perspective to know what to do with that data. And, like, there's no dentist that's gonna be able to figure that out. Right.

Speaker A: And so, um, yeah, it's interest. That's. I think that's why. Yeah, I think that's why we're gonna stay in business. Uh, because there's a lot of, like, a lot, a lot of our competitors sold. Uh, and they sold that. It's not like they was a fire sale. They, they're selling a crazy multiples. But a lot of competitors sold because they're like, like crap. Like, is there a place for us anymore if the dentist could do this all themselves? But yeah, I think, I think that's the value. Uh, I think, uh, the way it should work is the dentist should focus on being an amazing dentist and not be learning about chat GPT ads in the middle of the night. So I definitely think, I agree with you 100%. Um, but I think it's still really cool that um, uh, small businesses are able to access some of this and um, really have intelligent conversations about their data. With AI, I think that's just um, such a big change. Not again, not that they should spend hours doing it versus becoming good dentists. Right. It's not. You should, you should become a good dentist, uh, and have someone help you with marketing versus get, Become. Become medio, Mediocre dentist because you're learning marketing. Uh, but it's very interesting about uh, how AI is giving you the ability to work with your data a lot easier than it was before.

Speaker B: Yes. And what we consider data because there's so much that they can like, such as content. Moving into our next topic, SEO strategy powered by Data Insights. This is a big conversation. Everyone's talking about geo, uh, AI, SEO, all the different things. So um, yes, walk us through this topic, Alex.

Speaker A: Yeah, so this has been a hard ones. I do a lot of public speaking. This has been the most requested. This has been the most requested topic by far. SEO, GEO and AEO. GEO is generative engine optimization. AEO's answers engine optimization. Um, I actually use geo as a word more often, but slightly different, but very similar. Um, yeah, whole new frontier coming now with AI search. Right. Uh, Chat GPT already has ads. Anyone can do it. It was beta for a little bit. You needed to have a lot of money invested. It performed terribly. But now it's beta for everyone. It's a little bit better. Anyone could go and do ChatGPT ads and then similarly, um, anyone could start optimizing their website to um, to rank higher in AI search. And if we look down the road a year or two, uh, a search for sure will continue to grow. Um, it's a little bit different than traditional search. And what's interesting is traditional search at the same time also continues to grow. It's not that uh, ChatGPT is eating from Google search activity. Google search activity is growing at the same time. What's interesting is that we're seeing um, them use slightly in different ways. So AI search continues to be more upper funnel. So it's where people go to have conversations, uh, really top in the funnel. So if you're looking for a car, you might go on AI search and ask it what the safest cars are or have it review the safety features on this car. Or for example, you're looking at two window stickers. You upload those two window stickers into AI search and tell it to compare the features on both cars. And guess what, you could even ask it to recommend which car is better for you. Um, and then ultimately when people kind of make a decision whether it's the car they want, the car dealership, the car brand they want, then the, they need a dentist. AI. AI told them that they need a dentist. They wind up actually on traditional search finding those businesses. For now, that obviously might, might change. But for the time being, people are still trusting traditional search to find a dentist versus, versus, uh, getting, getting AI to recommend them a dentist. Um, but uh, I guess to your question, um, as this whole new market opens up, um, it's definitely, uh, becoming very, very important to figure out, um, how to come up in AI search, similar to how you would focus on SEO and coming up in Google Search. And uh, I would say all of this would be very, very difficult a couple of years ago without AI. Um, but AI now gives us, um, so much assistance that for sure you could use AI to help you rank on AI, if that makes sense. Um, you could literally have conversations with AI about the kind of content and have it guide you, um, with how to rank on, uh, AI search, Google Search. Again, uh, the data thing comes into play if you have your, um, if you have your AI, uh, we use Claude. So if you have your Claude, if you have your Claude, uh, hooked up to your Google Search console, you could literally have conversations with Claude about what content is working well for you, what content it recommends you work on, what content gaps uh you have, which content you should be working on. And um, so far, SEO and Geo ao, whatever you want to call it, have been working pretty similarly. It's still, it's still the same things that make them work, right? It's still about content. Uh, content gives you the opportunity to come up whether it's on AI Search or Google Search. So if you're a car dealership, um, and you make a page comparing the Honda Accord versus the Toyota Camry and someone goes, and whether it's on AI or whether it's on Google goes and searches Honda Accord versus Toyota Camry. By having that page, that, having that content, you have the opportunity to show up in both of those. So same logic continues. Creating content is very important. Um, some people have been using AI to create very shitty content. People call it AI slop. Um, AI is definitely very useful in assisting in creating content. I wouldn't go mass create crappy content that's similar to a lot of other crappy content others are creating. Um, but it's for sure very could assist with that uh, in finding the kind of content you should be creating. Then there's a couple other small pieces, link building, things like that, they're still valuable. That's a little bit outside of AI. Uh you could for sure use AI to help you do link building but a lot of the most valuable link building is still done a little bit manually. Uh but for sure a new frontier. Uh, what's interesting is um, and, and again the people, the people have been around for a long time. I appreciate this. Uh, with the new frontier we're back like in the wild wild west like we were with the, with the Internet in 90s where there's so many scams going around, so many people like selling snake oil. Um, I had a client forward me a spam email they received from a company. I'm not gonna call them out but it's not a small company, it's uh, dozens ah of employees that found a one man shop, a company that emailed him, it was a spam email but emailed him and offered him to get him to rank number one on chat GPT. And uh, he literally forwarded this me this email and he's like do you think I should do it? And I, I must have reread the email like eight times. It literally it says we'll rank you number one in chat GPT. And that doesn't even make sense. You can't rank number one on chat GPT. There's no ranks. Uh, so all this craziness going around with the Geo aeo, I warn everyone to be very careful. I think there's more scams around it than actual companies doing it. Um, you have companies like Google who have come out and publicly Google I think last week published a uh, blog on Geo ao. They've publicly multiple times now both Google and um, their Google search liaisons who are probably a little bit more trustworthy than even Google saying it but they've come out and said that AEO Geo is the same as SEO. It's the same things to focus on and be careful not to get scammed. But definitely a new frontier. Definitely reminds me of uh, the 90s where there's like, I don't know, scams going around, there's companies, new companies getting started every day, companies going out of business every day as well. Uh, definitely with that I guess comes a lot of opportunities with these kind of changes.

Speaker B: Yes.

Speaker C: Yeah, I mean I remember that for years everybody was sending emails out saying we'll rank you number one you know in Google or whatever. And it's like, yeah, okay, maybe you'll rank me number one in Google for, you know, best dentist in, you know, whatever this city that.

Speaker A: Yeah.

Speaker C: You know, can do braces. You know what I mean? Like, I'm number one for that entire string. But, you know, it's, it's, uh.

Speaker A: But it's crazy that they're carrying that to say that.

Speaker C: Yeah, it's crazy, right?

Speaker A: It's crazy they're carrying that sales pitch into Chat GPT though, because you can't bring. There's no number. There's no number one in Chat GPT. It's not possible. It doesn't make sense. But people are buying into. They're like, I think I should try.

Speaker C: Like what?

Speaker A: Like, you can't be number one in Chat GPT. That's not how this works.

Speaker C: Yeah, yeah. Or back in the days when you could just like sort of keyword stuff,

Speaker A: you know, pages, when you made that you put in your footer and you would make the. The text the same color as your background.

Speaker C: Yeah. Um, I'm curious though. Like, you know, everybody talks about the AI slop content out there. Um, and I'm, I'm curious, like, you know, people. So many people are creating these websites filled with AI slob. I mean, you're just like, you know, it's just tons of, you know, AI generated articles that are clearly, you know, um, you know, on sites that are probably. MFA is made for advertising sites. Right. And all that stuff, and they're just trying to get traffic to it. I'm curious though, like. And I know Google has gotten obviously much better over the years at detecting all of this kind of things. All the keyword stuffing and now the AI slop. Um, how is that working with like DEO? Um, you know, are the, are the AI platforms, are the LLMs detecting their own slop and, and, you know. Right. And determining that, okay, this is just AI generated content. So I'm not going to rank it, I'm not going to, you know, quote it, whatever.

Speaker A: Yeah, good. Oh, man, you had so many good questions I'll try to tackle. There was a couple of questions in there. First. First one is detecting AI. So what's interesting is. So I, I wrote a book a couple years ago. I actually wrote it. That's automotive. I was going to do automotive search marketing. I was gonna do one for every industry, but that I had such a bad experience with a publishing company that I took a little bit of a break, but maybe I'll still do it. Um, so I did this before ChatGPT, I think two months before. So I did this old fashioned way, like one keystroke at a time. And I was familiar with OpenAI. OpenAI has been around, was around for a long time before they released ChatGPT. Um, so in the end of the book I was like, there's this open AI stuff coming, um, that's gonna really allow AI to create tons of content for you but don't do it because ah, Google's gonna know. And my thought was Google and for those that don't know, Google bought DeepMind, which was probably one of the first, first AI conglomerate companies, probably the, some of the most uh, brilliant people in AI work there. And that was one of the reasons Google bought them. And that was a long time ago. It wasn't, wasn't recent. So Google owns that. It's like a, it's, they still have it, it's people, it's still brilliant people there. I'm like, if anyone's gonna know that the contents AI generated, it's going to be Google. And um, I knew, I thought for sure that Google was going to label it. I think I thought in like, in search results it was going to say like AI generated. And then all this came out and Google came out and they said AI, ah, content's fine. Their guidance was um, uh, quality contents, quality content, regardless of how it's created. So that's kind of sort of been Google stance and they've cleaned it up a little bit and now they say that quality uh, content uh, is still good as long as it's created for people or something. I don't think it says by people, I think literally just says four people. But um, what's interesting is that we ran a lot of tests, um, and a lot of times AI content will still rank on Google. It's hit or miss. Sometimes it does, sometimes it doesn't. Um, it's been so little logic around it and Google adjusts something and then stops ranking. So um, it's been all over the place on the Google side. So you could technically create a lot of AI slop and still have it rank and it's a crapshoot. Sometimes it'll work, especially if it's an authoritative domain. Um, with AI it's actually been a little different. It tends to not reference AI slop. It tends to be a little bit smarter than traditional search. And the way it gets around that is with AI it values two things much higher than traditional search. So it values uh, domain authority you could call it, which Google does As well. But it values domain authority a little bit differently. Instead of backlinks and all of that, it prefers certain uh, domains websites over others obviously Reddit for example is one of them. Reddit stock went up like 30% when Google said that. They're citing it a lot. Uh, so value certain uh, domains more than like, it's harder to game it. The other piece is the domains that does cite and by the way smart sites. There was a lot of studies done. We're actually one of the most cited brands in AI so or whatever an AI search. So whatever we're doing is clearly working. Uh, but what helps is you want lots of authoritative websites to also um, consider you authoritative linking you, mentioning you. It needs to be like a correct mix, just like links, spammy links, not gonna work anymore. So that, that's a big piece of it. So you wanna have a very authoritative both websites and brands. Uh, so it's harder to get away with AI slop. You have to be like, and by the way, uh, I was going to say you have to be like Forbes. Forbes actually does this. Forbes creates tons and tons of AI content because of the Forbes.com domain. Right. So if you are a very recognizable domain and brand you could then do it and then the second piece which is again a little bit unique to AI. So on traditional search, if you go, if you search, I don't know, you'd um, be surprised at the kind of AI slot Forbes is creating. And they actually started a sub company to do all of this and the sub company that does it is now worth more than the parent company and I think try to buy the parent company. So it's all they literally like license the Forbes.com name. So if you search, I don't know, Forbes.com go on Google search Forbes, uh, Apple pie recipe. They actually have uh, 67 apple pie recipes recipes on Forbes. And they're all different recipes that they like AI ah create. You know, I don't want to point fingers. I don't know how they create them. Maybe they manually came up with 60 plus different recipes. Um, so on uh, on AI, along with valuing domain authority, the brand AI really values topical authority. So that's kind of where they, they, they take away uh, places like for Forbes, uh, creating billions of AI content pages where they want you to be a topical authority on a certain piece of content. So uh, I don't know if you're a car dealership, if you create a lot of, let's say your Honda car dealership, create a lot of content about how to repair Hondas, you AI will now see you as a topical authority on um, servicing Hondas and they will more likely reference your dealership. When people are asking questions related to servicing Hondas, um, they want you to really become a topical authority as a, as a brand or for some people it even makes sense as an individual. If you're a small business, uh, you could probably more easily become a topical authority yourself. You are the topic expert on topic A. Um, so that, that's, that's the way it's been with a, with AI search. It's, it's, it definitely does a better job of not fighting AI slop. Uh, Google still, you could. AI overviews is still a lot of AI slop and there's a lot of uh, a lot of uh, apple pie recipes from Forbes.

Speaker B: Hopefully not AI slop apple pie recipes, but I guess you won't know until you bake it. But I can't believe it's already 4:48. We got um, topic three and tech stack to get to still. So we've talked a lot about AI slot, but what about actually applying AI in modern marketing? Like how are you using AI these days?

Speaker A: That's a good one. Yeah. And I, sorry, I go up on very long tangents so I could, I could be doing this for like 3 hours and not notice. Um, yeah. So I mean uh, I think the, the, the way to get the most value from AI, um as a company is, is there's, there's a couple of pieces and they'll have to align. I know a lot of companies who literally just buy chat GPT member paid accounts for their employees and they're like that said we did AI, we did it, we're there. I, I think there, there's a lot of pieces. Number one, you do have to give employees access to AI, uh and you have to encourage them to use it. Um, some employees are more likely to try new things, others you really need to encourage to uh, because a lot of employees don't even know how much they could make their jobs easier by using it and they're not going to actively try to find it out. Some will someone. Uh, so you need to give them access, you need to give them training, you uh, need to have processes in place. Um, and then I think it's connecting all your data. I think AI becomes powerful with the data. Right. Uh, if all of our employees have Claude, uh, we pay for it. They all have Claude. It costs us a ton of money. Um, if we don't have the data connect to it, they're not going to get anything valuable. I mean they could have conversations with Claude like they would with ChatGPT about, I don't know, it's uh, slow for my dentist. Do you think it's weather related? And ChatGPT will say let me check the weather in your area. It's been very warm so maybe people are at the beach. Like you can have those kind of conversations. But the whole value comes from the data. Right. So we have literally all the data sources possible, LinkedIn there and with the data and you could build skills for it. And um, once you literally build around it, um, I think that's a game changer. I think that's, that's where it could do crazy things for you and people, people don't, I don't know, some people are afraid, right? Oh, I don't want to look up data. But of course you need, you need guardrails, you need processes in place. You can't just have it like let's link it up to everything. Um, but I'll give you a very crazy example and this is the kind of stuff that people don't even realize AI could do. So I have it hooked up to my email to my Gmail, um, in a, what I think is a safe way. It doesn't necessarily, can I say there's guardrails in place. I don't want to go into too much, I don't want to encourage people to try to hack my email by again too much details but guardrails in place. Um, I'm a big car person. That's why I have the book on automobile search marketing. I do a lot of things in automotive marketing so I'm a big car person. I expensive hobby but I change cars often and that's. I, I like to say that I don't have a lot of other hobbies. Like I'm not a big sports person so I save money on the sports tickets but I spend on cars. Um, so I'm always like upgrading cars, getting a new car. So I was like with this whole AI stuff let me try to use Claude. And this is completely, this is completely different from Data, ppc, SEO. This is just like a cool thing that uh, I just want people to know the kind of crazy things they could already do. And it's a regular Claude. It's not like we're not like I spent a million dollars on it. It's the same thing that you guys got. I had to build a website from, for my existing car. So it literally built out a whole website with pictures, information that I think looks better than most car dealership websites. It really looks professional sharp. Um, it did that in about a minute and a half. We have hosting hooked up to it. So literally uploaded to Internet, it's now live. Then I, then I told it to go and find uh, the um, um, companies that buy um, luxury cars and email them the listing of my car and ask them for trade in values which it literally connected to my email. It literally found. It went out, it found uh, all these websites, it found the emails, it submitted the forms, it went to them, submitted the forms. So it sent like 20 different uh, places these emails. They started replying back with treat and values. And I told her to use these emails to negotiate the best price. So literally started quoting prices that one guy gave me to another guy, uh, like including screenshots saying that this, this actually company gave this price. Would you be able to beat it? Having all these conversations behind the scenes and it was doing. And I forgot I even told it to do this. I was doing it in the evening. I turn on my computer in the morning. I have like, I think like three people send me contracts of like we're ready to pick up your car tomorrow. Here's our contract to sign it bid people up by like $15,000. Is that crazy? And uh, I don't think it's sometimes uh, I told the story to a car dealership and they were worried that their industry is being disruptive. I don't think it's a disruption thing. Uh, the reason I give that example is there's so many outside of the box ways to use it. And um, that's not even like a company example that just literally some really cool things. But for it to do that need to be hooked up to my email. So um, you really have to hook it up to, to all the things you do. You can hook it up to your calendar. Could literally have it move your meetings around. You really have to have to have it hooked up to everything and have to be, have to have I guess some level of trust. But it's, it's so powerful and has gotten AIs been around for a couple of years now. But the kind of stuff we're doing now with Claude and MCP connections, um, we couldn't do three, four, five, five months ago. This is all like. And I know I'm going off on a tangent again but I'll give another, another really cool example really quick. Uh, will be 60 seconds. 60 seconds or less. Uh With Google, With Google Ads, for those that are in Google Ads all the time, um, you could go into your campaign, you could go to ad groups, you could go look, uh, at keywords and ads. Right? Uh, each ad group has keywords and ads. So you have an ad group, let's say, has 20 keywords and then you have 10 ads within Google Ads. You can't know which ad does the best for a specific keyword. So you have a lot of keywords and you have five ads. Let's say one of those ads could be terrible and one could be really, really good for one specific keyword. And there's no way to get that information. The Google Ads UI interface gives you no ability to look at ad performance at keyword level. With the Claude mcp, you could literally ask it that and it will tell you right away that's the power that you have with AI by hooking it up to different things you could. And that's why, by the way, Facebook was banning, uh, a lot of ad accounts for using the MCP before they finally, uh, changed their mind and came out with their own. Because you could do a lot of things with it that you can't even do it in the interfaces of Google Ads, Facebook ads. It's really, really powerful.

Speaker C: Yeah, that's a, that's a perfect sort of, um, way, uh, to top off this topic. Applying AI and modern digital marketing. There's just so many things that you can do and, and we've only touched the tip of the iceberg. Like, like you said, this is something that's just begun. Um, and it's only going to get better. Yeah.

Speaker A: Yeah.

Speaker C: Awesome.

Speaker B: 100%. Well, should we switch gears to the last topic, Tech Stack? Something we ask all the guests on all 200 plus episodes now of deconstructing data, which are, what are your favorite tools in your tech stack that empower you and your team?

Speaker A: Yeah, it's a very, very good segue from the conversation we just had. Um, honestly, for, for me and for most of our company, it's literally clawed with MCP connectors. It's, uh, it's become, it's not even close on the most valuable tool. Um, I can't say it replaced a ton. Um, like it's a lot of what it does is connecting to other tools. Like, it doesn't replace HubSpot for us. It connects to HubSpot and lets us be more efficient with the HubSpot data and do stuff that we couldn't before. But without a doubt, it's Claude and The MCP connectors. Um, and again, a lot of people use ChatGPT. Uh, they use Claude and they use it without the data connectors. It's literally they just have conversations and they think they're using AI and it's really not the same. And again, it's come such a long way in the last couple of months. Uh, two months ago, Facebook was literally banning any advertising account we had, accounts spending a million a month that were banned for using Claude MCP connectors because of how powerful they were. They would literally let you, with Claude, see what audiences are working well and aren't and not spend money on the bad audiences. And, uh, if you think about it, for Facebook, if all of a sudden everyone is not spending money on certain audiences, they can't monetize them. So that's why they were banning people. But, um, it's literally all in the last couple months. Um, and what you could do with it just blows anything else out of the water. And you could see that in, uh, your Switch tool, if you look at SaaS valuations, um, it's dropped a lot because it's, again, for us, it's not necessarily replacing a ton. It's working with the existing tools and using the data in a better way, but certainly could be replacing, and will be replacing a lot of. A lot of tools.

Speaker C: Yeah, I agree 100%. I. I'm a cloud user as well. I use a bunch of different AIs, but Cloud is really the one I use the most. And I've been, I'd say each week this, in the past few weeks, I've been connecting it to something else, and I just keep adding more and more connections. And I keep realizing then I spend a week just playing with that and being like, okay, how many more ways can I use this? You know?

Speaker A: Yes, yes, yes. And what's crazy that a lot of people don't. When I talk to a lot of people, uh, who are using AI, um, they're using claude, they're using Chat GPT, but not connecting it to their data sources for various reasons. And a lot of them have really cool data. Um, and they're just not. So I think that's for sure the next. The next frontier in AI, um, connecting it to data sources. And all of a sudden you're now able to actually use all the data that you've been collecting and that you have that you really haven't been using.

Speaker C: Yeah, um, we built an MCP server. We haven't officially announced it yet, but, uh, I'll talk About it. We built an MCP server for our platform, which basically gives anyone access to our entire identity graph, all of our consumer audiences. You build audiences off of it. You can map identities, you can.

Speaker A: You know, it's amazing.

Speaker C: So, um, and it's, it's, uh, it's. I'm amazed at how powerful it is and, and how many things we've been able to do with it that previously we used to have to, like, you know, somebody in our account management team would have to ask an engineer to go, you know, get counts for how many people would be in this audience, you know. You know, you know, just to give it, like a real application, you know, a dealership, you know, in this area. Uh, they want to know everybody that's within this radius that owns a BMW, whatever. It's like, we used to have to go and ask engineer to get counts for these things, and now it's just, you ask Claude. Claude tells you.

Speaker A: Yeah, it's crazy. And what's crazy is that Claude itself could build a lot of these things for you to use your own data. It's, um, uh, Claude code, cloud, cowork. It literally could help you build all of these tools, uh, itself.

Speaker C: Yep.

Speaker A: Yep.

Speaker C: Awesome.

Speaker B: It's a new day.

Speaker C: Yeah.

Speaker B: Well, in closing, Alex, would you like to let listeners how they can reach you after the show?

Speaker A: Oh, yeah, absolutely. So, um, if you go to alexmellon.com I have all my social media things at the bottom. You could connect with me. I'm most active on LinkedIn, but I'm pretty much everywhere. So. A L E X M E L dot com. Of course, it could go smartsites dot com. Check us out there. All the cool work we do. Or search for us on, um, Google. I used to say you could search for me on Google or search for me on AI.

Speaker B: Awesome.

Speaker C: Awesome. Alex, thank you so much for joining us, uh, today on the podcast. This has been a great conversation. I think our viewers and listeners got a wealth of information thanks to you. So thank you so much.

Speaker A: Perfect. Thank you for having me.

Speaker B: Yes, absolutely. I'll close this out with the video. So, audience members, we want to invite you to get Powered by bdx, build authentic connections with your audience, and drive unprecedented engagement. So just imagine harnessing the power of AI to craft your audiences that truly resonate with your brand, ensuring that your marketing is hitting the mark every time. We want to bring you closer to those marketing goals, and we're eager to guide you. So just visit BDX.com, you can click on the Talk to an expert button and get followed by bdex. Deconstructing Data is the show that makes data make sense. Every Thursday, David Finkelstein and myself, Jesse Liszak, sit down with industry experts to unpack how data really drives marketing. We do this in plain language anyone can understand. If you want real insights, real conversations and zero jargon, tune into Deconstructing data live on LinkedIn every Thursday at 4:15pm Eastern Time.