#585 From Search Engines to Answer Engines: Aaron Burnett on How AI Is Rewriting Digital Marketing

Mehmet: [00:00:00] Hello and welcome back to the opposite of the CTO Show with Mehmet today. I'm very pleased. Joining me, Aaron Burnett. He's the CEO at Wheelhouse Digital Marketing Group. And of course, the guest can guess from the company name. We're gonna talk about digital marketing. We cover, of course, like AI versus the traditional distribution.

We're gonna talk about data privacy, control and ownership, and you know. Little bit about regulated environments, how this affects also the go-to market for especially startups because I cover a lot about startups and, and the enterprise in general also as well. Um, so without further ado, Aaron, thank you again for being here with me today.

I don't like to steal much from my guest type. Tell us a little more about you, you know, your background, your journey, and a little bit about, uh, wheelhouse Digital Marketing group, and then we can start the conversation from there. So the flow. That's 

Aaron: great. Well, thanks very much for having me. So I'm Aaron Burnett, I'm CEO and founder of Wheelhouse.

Digital marketing group. Uh, we are a performance marketing [00:01:00] agency focused on privacy first industries with a strong concentration in healthcare and medical device manufacturing. Although we do work with a few other clients, uh, that aren't in those industries. Notably, we've worked with NASA for the last seven or eight years.

My professional background is that I began client side in very large corporations, uh, running marketing. So I was a VP of sales and marketing for a division of at t Wireless, had other similar positions, uh, with. Principally tech companies, uh, here in the United States and in creating wheelhouse. 15 years ago, I was creating the kind of digital marketing agency that I wished I'd been able to hire when I was running marketing for other people.

Mehmet: Great, and thank you again, Aaron, for being here with me today. I want to start, you know, with the elephant in the room, the ai, right? So, so everyone is talking about ai. Um, when, you know, before we, we, we recorded this, and of course there was some communication, something stuck with me, which you [00:02:00] mentioned is that 99%.

Of queries now, like health queries and in AI overviews, like what does this fundamentally break into the traditional internet model? So let's start from there and then we can come to the distribution and the marketing part of it. 

Aaron: Yeah. Uh, so not only are. Our queries ending in ai, consumer use of AI and consumer trust of AI is rising at, I, I think the adjective I'd want to use at an alarming rate, given also the prevalence of inaccuracies.

Uh, and so, you know, what that, what that breaks, uh, in terms of. Uh, the traditional digital marketing, uh, approach is, is that of course, this is a, another channel, another distribution, uh, path that you have to take into account when you are looking to promote your brand to, uh, reach and attract your audience.

Um, so, uh. We have found that we need to double down on some fundamental tactics, some technical tactics, uh, that, uh, [00:03:00] perhaps had had become fallow and that we need to think of ai, uh, as a, a core distribution channel, uh, in all of our digital marketing. 

Mehmet: On something, which personally I started to feel it maybe couple of months after, you know, chat GPT and then later on the other tools started to appear is, you know, shifting from relying on traditional search engines, mainly Google, right?

Um, to relying on AI as kind of a search engine. So, and I'm sure like maybe you have seen something similar. Let's say someone is searching for something. Uh, but in the case of Google and other search engines, whether it's Bing or any other, um, big search engine, you know, we have, you know, or let's say we are trained to just see which one is relevant to us, and then we click on that link, we go there and we do what we wanted to do with ai.

So it's a generated [00:04:00] answer, right? And. I noticed it even on myself is that I don't actually click on that link. I would let AI Tell me about the content. So are we seeing a shift from search engine to answer engine or, and, and what's the role of the website in, in this whole formula, in your opinion? 

Aaron: Yeah.

We absolutely are seeing a search from, uh, uh, a shift from search engine to answer engine. You know, we just conducted a, a study of about. 1100 consumers asking about a couple of things. One of the core areas we were asking about was the use of AI versus other, more traditional digital channels and most notably, uh, search engines.

And we found a rapid increase in the use of ai, which is expected, right? Uh, increases year over year of 50 and 60%. But we also found though that was more alarming, uh, is that. Trust in AI without verification [00:05:00] was increasing at almost the same sort of rate. And I think, you know, uh. Thinking about that from a psychological perspective, you are right that the answer is structured almost conversationally.

It is structured as though a, a person, an entity, is giving you the information that you need. And so as we probed, to what extent do you trust the information that you're getting from ai? We saw very high rates of trust, and then we asked follow-up questions. To what extent do you verify that information?

And this is specifically around healthcare information. Uh, we found a surprisingly high percentage of people who accepted the information from AI without verification. And then most notably, and again I'll use the word alarmingly, we found that 12%. Of the people surveyed, and this is a survey of about 1100 people, said that if they were given conflicting information from an AI query versus what their doctor told them, they would trust AI over what their [00:06:00] doctor told them.

Mehmet: Now, let me ask you, Aaron, as a follow up question. So from digital marketing perspective, we know how search engine marketing works. So how this do you think affects, you know, the economics, uh, when it comes to the website versus, you know, like, uh, you know, the click ratios and all these factors, how that is changing in, in the age of ai?

Aaron: Yeah. Uh, we, we certainly do see a fall off in certain types of queries. In particular, uh, very long tail queries. So long conversational queries. Uh, we can see that data, uh, in some of the analytics that we use. We have an enterprise grade search intelligence platform that we developed. We can see those long queries there.

We see the long periods obviously shifting, uh, predominantly to AI and pulling out of. Uh, of, of search engines. Uh, we aren't seeing, at least among our clients a decrease though [00:07:00] in conversion. Or in instances where we're dealing with actual transactions in revenue generation, uh, we're just needing, we're seeing a, a difference in the nature of the query, uh, in context.

And then we're also seeing that we can still learn quite a lot. About user intent by mining data from both of those sources. So, um, we get access to query data from, uh, LLMs. We can marry that with query data from search engines. Uh, if anything, we're getting even more insight into searcher intent from the LLM related queries, and we can, as a consequence then refine and optimize website structure.

Content marketing, even digital advertising strategies, so that we are targeting folks based on that intent, regardless of where we're finding the source of the intent. 

Mehmet: Right now, we all know when it comes to LLM that it's fed by data, right? So mm-hmm. [00:08:00] Uh, and this is how these, uh, uh, AI models are able to spit out, you know, the answers that we see.

Now, um, of course we know the mechanics. We don't want to go too much technical here, but, um. Who ultimately control this layer. Aaron, like, and we saw a lot of, you know, debates, especially, you know, about like these L labs were scraping the, the internet, they're getting this so. Who controlled that. And by the way, you cannot reverse this.

Like if, if an LLM already, you know, escape your, your data, it's with them already, so. That's right. Are they in control or is it like there are some other players you think in, in, in there 

Aaron: you're asking are the LLMs in control? 

Mehmet: Oh, who is in control? Is it the LLM, you know, owners? Like who, who's in control?

Aaron: Oh, 

Mehmet: I'll, I'll, I'll follow up [00:09:00] with the question later on that. Yeah. 

Aaron: That, yeah. The answer to that question is unclear. On any given day and in any given moment, I would say that that, uh, the LLM certainly appear to be in control, uh, that they, they were. Pretty aggressive in going out and scraping data that wasn't theirs and training on data that wasn't theirs.

And you can see the byproduct of that in some of the lawsuits that exist. Uh, I think the New York Times has sued, uh, OpenAI for training on some of their data. Uh, there's an author group that has sued the LLMs as well, sued open AI in particular for having trained on, uh, copyrighted data. Uh, but nonetheless.

OpenAI, uh, in particular was quite aggressive in going out and securing data that wasn't theirs and training their models on it on an ongoing basis. I think this is very much a game of cat and mouse. Uh, you know, the LLMs continue to seek new sources of information to train their models. I think OpenAI.[00:10:00] 

Perhaps most notably and most prominently does that in some instances they've structured contractual relationships with, uh, data sources such as Reddit. Uh, in other instances, they are simply slurping that information in, and it is left to folks like me and others in the search industry and others in the legal industry to try to discern from search results where in fact, uh, they're getting that information.

There ostensibly should be some regulatory body that steps in over the top of this and, and says, this is the governance that we will apply to data sources and the way you're able to use your models. But at least in the us uh, at a federal level, there is absolutely no indication of enthusiasm for that.

Mehmet: Yeah, the time will show us like how, how this will change. But let's talk about the privacy, Aaron, and I'll tell you why I'm asking you about privacy now. Everyone knows, and even people who are not like, let's say, too [00:11:00] much techie, I would say everyone is tech savvy in this age, I would say. But let's say the ones who understand exactly how the model works.

So I've seen this personally, I did multiple times, but I do it with caution because I know, you know. I might be feeding these LLM something which might contain private data, so I try to anonymize as much as possible. Let's talk about these risks, like especially in digital marketing context, when you are trying, you know, to, to enhance maybe the way how the LLM will tell other people about your product.

Or maybe you are just trying to write a. You know, marketing copy for something to your client. Let's talk about the risks, especially related to privacy here. 

Aaron: Yeah. The biggest risks, uh, relate to data, uh, and data privacy. Um, you know, the, the nature of an LLM is that it is a public model and as you just.

Suggested in the the last exchange, any data [00:12:00] that is incorporated into an LLM becomes defacto public and is subject to being surfaced elsewhere. And so, you know, for us, because we are focused on, uh, privacy constrained industries, healthcare and medical device manufacturing, where it is possible that, uh, PHI might become a part of a data set to which we have access, we have a data warehouse that is a.

A HIPAA compliant or A-G-D-P-R compliant data warehouse that enables us to have access through contractual relationships, to, uh, data that might contain PHI that makes our data, uh, uh, that makes our, our digital strategies richer and much more effective. And so for us in particular, we have to have a hard and fast rule that no data will be shared with a public LLM.

The only way that we can use ai. From a data science perspective and a business intelligence perspective, which is core by the way, to the way that we deliver exceptional results for our clients, is that we have to use a private model. We have to be on [00:13:00] infrastructure that is HIPAA compliant and GDPR compliant.

We have to deploy, uh, an LLM that is, uh, effectively private and trained only on that data with no public connection and no public sharing when it comes to more conventional digital marketing. Content creation, copy creation, that sort of thing. Our philosophy and our ethos and the rule for our employees is that we will use AI extensively as a foil for our thinking, uh, as an aid to make us more efficient in, uh, completing certain repetitive tasks or certain forms of analysis.

And we do use it extensively.

Um, what we won't do, because again we work in these highly sensitive industries, is use AI for any form of content creation.

So we are not going to create articles, blog posts, that sort of thing with ai. And there are a couple of reasons for that. One is the accuracy of the information in healthcare and medical device [00:14:00] manufacturing is absolutely critical. So we don't want to introduce any risk of error and regardless of. How extensive and how fantastic your prompt is, uh, and how much you perfected it.

There still is risk of error. There is variation and inconsistency when you're using LLMs. Uh, and the second is that in these industries, perhaps more than others, uh, authorship really matters. The fact that you have a credentialed physician or researcher who has authored content, uh, means something to the reader, and it also means something in terms of the performance of that content.

If you've structured that content well. 

Mehmet: Right. Arun, let me ask you how all, what you describe, you know, especially because you mentioned hipaa, uh, and you know, like for people who doesn't know, like this is the, uh, uh, the standard for securing data and, and you know, assuring privacy for healthcare in, in the United States.

And we have similar or equivalent standards in other countries as well. You [00:15:00] mentioned also like, uh, some other, uh, protocols as well or some standards and, and these are like everywhere and you know, you mentioned you deal also like with, with high profile customers, so some of them they might be, you know, governments, some of them they might be, as you mentioned, healthcare, banking, all this, everyone now.

Knows the importance of this, but how this change, you know, the risk profile around compliance and trust how much this has become, you know, paramount more than even the technology itself to establish, you know, this trust with, with the customer when you, when you are dealing with these LLMs as I'm asking you here from a, a marketer perspective, uh, Aaron, 

Aaron: yeah.

So you're asking how much, uh, the introduction of LLMs in these contexts. Uh, changes the trust paradigm. 

Mehmet: Right, right. And the risk 

Aaron: profile. Yeah. Yeah. So I would say there are, there are a few implications. The first is that, um, the brand [00:16:00] owners need to be explicit about, uh, the ways that they will and will not use, uh, customer data, whether it's PHI, protected health information, uh, or not, how they're going to use data, how they are and are not going to use ai.

So trust, I think, begins with transparency. Um, and. Uh, I, I think that they then need to bake that into their consent protocols as well. So one of the things that, that you need to do, uh, here in the United States under HIPAA and, and really globally, under the other various privacy standards, is ensure that you have a really explicit consent mechanism.

That is true in Europe under GDPR. It is true globally, under all the other standards, and so. Use of, uh, data in the context of AI or LLMs also should be disclosed in that consent mechanism as well. And increasingly, you know, you need to be baking in, uh, not only your position on privacy, but your position on business [00:17:00] practices and AI into the technology and the tech stack that you use to execute your marketing.

Mehmet: Right now, Aaron, like regardless if it's a private model, public model, these LMS can do mistakes. So when. AI summarizes sensitive or regulated information incorrectly. Who should be accountable in your opinion? 

Aaron: Oh, the human being. The, there should always be, there should always be an expert who is the responsible adult in the room, who is, uh, both guiding the LLM, but also reviewing the content.

Uh, there's so much in the zeitgeist now about. Uh, agentic AI and automated workflows. And if you just built the right chain of agents, then you could automate your job and you would only work two hours a week, but get the productivity of a hundred hours. And that's, uh, first of all, I think that's, that's folly.

It's puffy. I've not seen any evidence that anyone's actually being able to [00:18:00] do that. And second, I think that's terrifying because. It. It completely negates the ability of the human in the room to be the expert who guides the action and the expert who is ultimately accountable for the output of what? Of whatever LLM and whether it is accurate.

That's the base state, but also quality. If we automate all of this work, particularly let's say in digital market and content creation, we would rush to mediocrity so quickly. There would be. Massive volumes of undifferentiated and almost unreadable content. Uh, so I think, uh, at least in digital marketing, we're a long way from, uh, true agentic ai.

I think if we take our responsibility seriously, we should never get there. There should always be a human being who is exercising discernment and ensuring that the output is accurate and of high quality. 

Mehmet: I think it's not only in digital marketing, it applies to everything because, you know, we've seen it.

Times and times [00:19:00] again, um, with the wrong information. And we have to have someone, as you said, as a human to check on that. And, you know, we see it, we saw it also a couple of weeks ago with, um, I'm sure like maybe the audience, uh, have, have seen it also, or like they heard about it when. Amazon, you know, the AWS division, they had to call their engineers and said, Hey, you need to stop.

Just keep writing code, uh, using ai. I mean, they're not writing the code. Actually, they are using AI to write the code for them and without like, uh, any. You know, like, uh, some, any qual good quality control, I would say. And they said, no, you have to go back and write code by hand sometimes because you can't give anything to ai.

We've seen again and again, you know, these, uh, guardrails, which are not put in place. And I think here where we call it with some of my guests, like human in the loop. Um, of course, like some of my guests, they said even the human in the loop, we can give it to an [00:20:00] ai. Sometimes, sometimes not. But yeah, like if something critical, especially if you're dealing with sensitive data and if you're dealing like with something critical.

Whether it's a marketing copy, whether it's a piece of code or whatever that is, you need someone to check it. And yeah, I agree with you, Aaron. Like we tend to take shortcuts, what I call pe. Why I tell people like, we like to take shortcuts. We are lazy by nature. And yeah, like, let's, let's get to the AI now.

Aaron: Yeah. 

Mehmet: Let me, let me ask you with, with, as we were talking about digital marketing, of course there's a debate today about. Where SEO or search engine optimization is, is, is heading, is it effectively that, is it evolving into something else? Some people they call it, uh, ai o, which is AI optimization. Some people are calling it like, uh, GAIO, just AI optimization, whatever the name is, so.

Uh, from your perspective, you know, how [00:21:00] does winning look like in this new paradigm? 

Aaron: Yeah, it's a good question. Uh, winning in this new paradigm looks like, um. Paying close attention to the signals that tell us what really works. Uh, and then in investing in those areas where we have proof positive that those signals, those tactics, uh, actually still work.

So for us, that means, uh, going back in some instances to. SEO tactics that worked historically and are now shortcuts for inclusion in LLMs. Uh, things like structured lists and facts and top 10 and, uh, the, the top 10 best or the top five best, or that sort of thing. Uh, those were kind of tired SEO, uh, tactics and haven't really worked very well for the last few years.

All of a sudden they work very well in the context of an LLM and they work very well in the context of an LLM because LLMs are looking for computational efficiency [00:22:00] and they would much rather, uh, visit one source that gives them the answer to give me the best of, or the top 10 than visit 10 different or a hundred different webpages.

But it also means doing other things that always have been part of good SEO, uh, and now. Are even more important in the context of ai. Things like focusing on structured data and really good code. So, pardon me, code hygiene on websites and going back to, uh, some of the fundamentals of content marketing.

Alright, let's not just produce content against the list of ideas that we have as marketers. Let's pay close attention to user intent again. Finding the signal in all of that data that we can get from LLM queries and from search queries in conventional search engines, and ensuring that the content strategies and the content that we create are highly attuned to addressing the implicit intent and the explicit intent in those queries.

And [00:23:00] then. Placing that information not only on our site, but also finding other venues in which we should place that information. And then finally, and actually perhaps most importantly, it's ensuring that our brand is mentioned in all of the places where, uh, you would expect a brand such as ours. Such as ours to be mentioned.

So, uh, forums where people may be discussing our product or service, uh, review sites where they may be discussing our product or service media, but not necessarily mass media. So an article in the Wall Street Journal may be of much less value to me than an article in a niche trade publication that sends a clear signal that I am a notable and prominent and authoritative player in a particular industry.

So. It's not a radical rethinking of the kinds of strategies and tactics that you would use in search marketing. It's a refinement of those strategies and tactics in a way that pays close attention to what really is [00:24:00] working in the context of ai. 

Mehmet: Now let me ask you a question and, uh, excuse my ignorances Aaron, because I'm not the expert in, in, in search engine optimization.

I know a little bit about the concept. So, back in the days to have like the authority of a website for Google, you needed to have, for example, you know, the, the back links. You need to have like the constant refresh of the content of the website, all this. So more or less people knows how it works. Now with ai.

And this is my personal experience with the LLMs and I'm talking, you know, if the information that you're looking for about a company is already in the dataset that the LLM have already taken, or like it's been trained, uh, has been trained on. So it tends to me like it's sometimes random how the LM would reply to you.

For example, a company names, and I'll just give you an example. If I go today and I ask. For example, [00:25:00] any LLM give me for example, a company that does X, Y, Z, right? Just any random thing every time. It'll answer differently now unless you go and force it to go and search. And then I think here we fall back.

On the traditional kind of SEO because again, what they, I mean chat, GPT, Claude Gemini, they're going actually opening a search and then they're trying to, to, to get the data from there. So, but the first step, because again, we are, we, we are lazy. Right. We discussed the slack couple of minutes ago. Can companies do anything to make sure that they, that the LLM is the first company that they will be split from, from the LLM to, to, to their customers?

Or is it something random? 

Aaron: Yeah, I don't know the answer to that definitively. I can tell you what the signal tells us. So you mentioned a couple of ways that SEO worked. Traditionally, you needed back links, you needed to refresh the content on your site. One of those two things still works very, very well.

I. [00:26:00] In the context of AI and LLMs and that is fresh content. And here again, there are, you can think of what's happening with LLMs and, and optimization for AI as almost the early days of SEO, where there were some mechanical things you could do that had an outsized impact. And one of those mechanical things is that you can refresh dates on content.

The freshest content tends to win. And so. For example, you publish, uh, a blog post and you do nothing but change the date of that blog post. You published it last year, uh, it's kind of fallen off. You change the date in the URL, you change the date in the heading of the the site. You change it in the, the metadata and you use structured data to signal the date.

You will find that you get a nice performance pop from that content. 2026 content performs much better than 2025 content, which is so simplistic and dumb. We are at that level. In some ways, LLMs are so sophisticated and yet very unsophisticated in the [00:27:00] signals that work. And so an imperfect and incomplete answer is that if you want to be known for something, start publishing really fresh content with very recent dates that mentions.

Your brand name in combination with the thing you want to be known for. Do that over multiple pieces of content and see if you can see some of those mentions in other forums that are relevant for your industry. And I would vary that by LLM. You know, look at the specific data sources that are, uh, prevalent for a particular LLM or the set of LLMs in the moment, and ensure that you appear there.

Mehmet: Right now, let me discuss with you a little bit about the go-to market, GTM as we call it, uh, an implication, uh, on, on startups. 'cause as, as I was telling you, like we focus on startup, but also like, maybe it'll be applicable to, to other, uh, you know, like established companies as well. So, if AI becomes the primary interface to customers, how [00:28:00] do you think, you know, this will affect the way startups do, or they have to think about their go-to market strategy?

Aaron: It's a good question.

Mehmet: Or is it business as usual? 

Aaron: It's not business as usual. I'm kind of thinking that through. We tend to work with larger and better established, uh, companies, but I'll tell you, uh, based on our experience, what I think. I actually think that,

you know, startups come to market with a focus on differentiation. They should have a different thesis, uh, a different take on a particular market, and to the extent that they can clearly articulate that different thesis, and they can promote that different thesis, not only in their own content, but in the forums and in the niche media.

That are a part of the industry in which they want to penetrate. I think that startups, and this is a kind of a contrarian opinion, I think that startups, [00:29:00] uh, actually have an advantage in a context in which AI is dominant over traditional SEO in a traditional SEO context, you need domain authority, which is really a proxy for time. How old is your domain? Uh, how, how much has your domain been sorted? You need links, which also is kind of a proxy for time if you, in fact, if you get a whole bunch of links very quickly. That's a spam signal and it's more likely to hurt you than it is to help you.

You need a really significant body of content. You probably need diversity of content as well. All of these things take a lot of time, and so there is a, a, a bias or an advantage to a much better established brand, a much better established site that's been around for a while. In the context of of LLMs, you need signals that.

Say clearly, I am an authority, a credible authority in this space. I have a perspective, uh, that perspective is valued and cited by others. Uh, it is in all of these contexts where you would expect [00:30:00] someone who is focused on, uh, I don't know, raincoats for zebras, something like that, right? You're going to have something that is slightly left of center from the main players in the market.

And so to the extent that. You are very thoughtful and you are strategic in the way that you come to market with your content, uh, with what you publish, with the media that you pursue, and, uh, where you are able to either place articles through, uh, pay promotion or hopefully through organically place just PR generated articles.

You have a much better shot at being a part of the mix. Uh, in LLMs than you do in traditional SEO. So I think my thesis, and again, this is different than some of the prominent voices in the SEO who would argue that there is a disproportionate advantage for well established brands. I actually think there is potentially a significant advantage for startups who can clearly articulate their value proposition and go after only the [00:31:00] relevant media and forums, uh, that really matter in their market niche.

Mehmet: I think I will. I will agree with you here, Aaron, and this is something I noticed, like the new mode is the way if you. I'm talking here about like startups. If you know how to choose the right channel or the right distribution access. You got it. I'll give an example. For example, you know, like it's not a startup necessity, actually.

Yeah, it was a project at what acquired, like the open clothing thing, you know, like which Open AI acquired, so the whole noise happened on X, on and on Reddit. Mm-hmm. And I think the guy cracked. Crack the code by knowing well exactly. To, to put his messages. And then he got all the notice, right? So he got, he got noticed, and I've seen this, you know, repeated again and again the past two to three years.

So either it's, it's like they choose one channel or two maximum. Some people they do it on LinkedIn. If it's more B2B, I see. Some people do it on Reddit. I see some people, [00:32:00] they do it on x. Or as known as, uh, Twitter before. So I, I'm feeling that this is kind of a new mode of distribution if they got it.

And then the LLM will pick it up and especially if it's an account of LLM with the access to live, uh, data, like something like Grog or Gemini. So now you have this. You know, advantage, I would say just that it's, it's my own, you know, thought

Aaron: yeah. Yeah. So you raise a really good point. And in fact, I've been talking with other CMOs in healthcare and med tech, and they're noting the same thing. And that is that not only is it targeting the right channel, but that also, uh, caters to changes in, I'll call it media consumption habits.

Digital consumption habits. Increasingly people are, uh, our, our. Focusing on one channel and staying in that channel. So historically, you might promote something to someone in TikTok. You would be driving for, uh, a click. Get them to click [00:33:00] off. They're gonna look at something on your website. You'll feed them more information, and it might be a multi-channel exercise.

Increasingly what marketers are finding is they need to tell the whole story in one channel. It needs to be a series of messages that are in Reddit or in TikTok or all on YouTube. Everything from introduction to uh, A problem and here's my brand and here's a solution for your problem and here's how you might convert, and here's some video testimonials and all sorts of things in that channel.

So I think you're absolutely right, and that's reinforced by consumer or buyer behavior. The other thing that's interesting to me, uh, in what you mentioned is, okay, claw bot. It was terrifying and astounding to me to see Claude Bott hit the zeitgeist and be see so many people so enthusiastically start using it and giving it access to all of their data and their systems, and be delighted by the fact that they were letting Claude Bott run rampant on their machines.

I've had very smart technologists reach out to me in the last couple of [00:34:00] weeks to say, Hey, what do you think Claude bought? I'm starting to use it and I think you should not use it, given what I know is on your machine and in your systems. 

Mehmet: Yeah, it's, it's, it's so super, you know, you, you need to be super careful, I would say, like, I'm not saying don't use it for me as, as a, as someone who come from technical background.

So I can sort things out. Like I put it in, in, you know, in isolated environment. And of course I didn't give. Give it any access to my files or emails, just, you know, I want to see what its capability. It has huge potential. Let me tell you this, but yeah, you need to, to understand how you should set it up in the right way.

But yeah, I absolutely agree with you on this. Um, as we are coming, uh, to, to the end, uh, Aaron, like. Just, you know, maybe final thoughts on, you know, the future ahead when it comes to digital marketing and of course my traditional question, uh, where people can find more about, uh, you and about the company.

Aaron: The Future of Digital marketing will bring together a [00:35:00] few things that we've talked about. First, our thesis is that what will matter, what will continue to be valued in digital marketing's true expertise, like top 10% of your discipline, your skillset, level of expertise.

And combined with that discernment, we were talking about the, the criticality of the human in the loop. It's great to have a human in the loop. That human in the loop has to know the difference between excellence and mediocrity, and I think that would be an increasingly critical role. And then finally, nimbleness of thought.

The ability to, uh, jump among systems, understand how things might be interconnected in a way that is novel and to orchestrate among channels. We have found that we have been able to be exponentially effective by also combining our own technology, uh, with. Effective marketing strategies, so. We didn't talk about this a lot, but one of the things that's happening in digital marketing is that the platforms Google Meta, [00:36:00] other advertising and analytics platforms are increasingly taking data away and taking control away from advertisers and agencies.

And the reason for that is that they want to, uh. They want to tell you that they will algorithmically optimize your campaign. Simply give them your money and they will make the good thing happen that you want to happen. And so as you lose signal, as you lose data, it's important to take control of your own destiny and.

Implement your own instrumentation, your own approach to analytics, your own data store that enables you to integrate and harmonize data, not just among advertising and analytics channels, but also to integrate first party data, data that is in the client's own data store and integrate with their CRM system so that you have really clear signal and persistent signal.

In a data store that, in our instance is HIPAA and GDPR compliant. So allows us to look at full fidelity data while respecting user privacy. And then you can start to do very interesting and sophisticated things on [00:37:00] top of that, like, uh, propensity modeling. So, uh, our ability to algorithmically score every behavior and every lead submission and provide a real time signal back to advertising platform saying, on a scale of.

Zero to a hundred, that was good, bad or indifferent. We want more of that, less of that. All of those things are within our control so that as platforms continue to change in a way that suits their self-interest, we don't have to worry so much. We have controlled our own destiny. We have our own data store, we have our own instrumentation, and we can continue to drive very effective performance regardless of what's changing in the ecosystem around us.

Mehmet: Great. And where people can get in touch, Aaron, to learn more and uh, you know, maybe get in touch. 

Aaron: Sure. They can find, uh, our company website@wheelhousedmg.com. They can reach me at aaron@wheelhousedmg.com and they can find me on LinkedIn. 

Mehmet: [00:38:00] Great. I'll make sure that I'll put the links in the show notes. So for the audience, if you're listening on your favorite podcasting app, you can go to these links by checking the show notes.

If you are watching this on YouTube, you'll find that in the description. Again, Aaron, I can't thank you enough for being here with me today. I know it was early for you in the morning to to, to jump with me on this recording, uh, as you are in, in the US and I'm in Dubai. So thank you again for joining me and this is how I end usually my episodes.

This is for the audience. If you just discovered us, uh, thank you for passing by. I hope you enjoyed If you did, so, give me a favor, subscribe and share it with your friends and colleagues. And if you are one of the, you know, loyal fans who keep listening to the show or watching the show, thank you very much for all the support you are showing.

And I'm repeating myself. I can't thank you enough because without you, we couldn't. Make it in the Apple top 200 podcast chart across multiple countries. And this is a time since last year we keep changing countries, [00:39:00] but every two weeks or three weeks, I'm seeing we are going to a new country, which is perfect.

And I'm seeing the podcast is always, at least in three to four, uh, top 200 chart. So again, this is not my effort, this is because, you know, you listening and coming back. So I really appreciate that. And as I say, always stay tuned for a new episode very soon. Thank you. Bye-bye.