[The Marketing AI Show: Episode 5] Marketing AI Innovation That Could Change Everything

The Marketing AI Show—the podcast that helps businesses grow smarter by making artificial intelligence approachable and actionable—is back with another episode. 

Click here to listen to the first five episodes of the podcast on your favorite podcast app. Keep reading for highlights and video from our latest release.

Episode 5: Harry Syed, Executive Director of Innovation, Sub Rosa 

Harry has an insatiable drive to make things better. Right now, he’s launching an AI and Natural Language Processing engine at Sub Rosa—designed to understand human trends through user-generated content to support creative decision making in the moment. 

He’s previously held roles in leading marketing technology efficiencies for Pepsi, McDonald’s, Unilever, Mercedez, and Booking.com globally. Harry has also spent time in Hong Kong, Sydney, Dubai, Amsterdam, Madrid, London, and New York building, deploying, designing, and testing strategies to drive incremental revenue.

During his interview with show host Paul Roetzer, Harry discusses practical applications of AI that he has worked on, and forecasts game-changing innovations coming in the next three to five years. 

Watch the whole interview to hear more about:

• How AI will change our jobs as marketers. 
• How machines can make us more human. 
• Why McDonald’s is one of the most innovative brands he’s worked with. 
• Tips for marketers looking to become innovators. 

Harry left our listeners with this advice: “What would really be amazing for marketers moving forward is: find new ways to learn the true value of how you can innovate with automation, with intelligence, and educate yourself. It's all about education at this point. The more you learn, the more you understand and how it's relevant to your specific use cases.”

[Video] Watch the Full Interview 

Read the Full Transcript 

Disclaimer: This transcription was written by AI, thanks to Descript. 

Paul Roetzer: Welcome to The Marketing AI Show I'm joined today by Harry Syed, Executive Director of Innovation at Sub Rosa. Welcome Harry.

[00:00:14] Harry Syed:  Thank you very much.

[00:00:16]Paul Roetzer: Yeah. And you and I got talking that was about two weeks ago. I think we got introduced by Sweezey. If I'm not mistaken. Matt Sweezey was one of our guests on the show early on. Um, And Harry and I, we didn't even, I don't even know why we were talking. Honestly. I think Mat just thought it would be cool for us to have a conversation.

[00:00:32] And, you know, it kind of started off like, you know, just learn a little bit about each other, talking a little bit about some interesting use cases of AI, but then pretty quickly we both realized we had a, really shared interest in the transformation of marketing through like, the big ideas, the leaps forward.

[00:00:49] And I rarely come across people in the industry who have that vision, but in your case actually have the capability to build the things. And I was like, dude we got to do this again? But we're [00:01:00] recording it this time. And so here we are talking about how we can change everything with like the big ideas for AI and marketing.

[00:01:08]Harry Syed: Yeah. Going back to how we met is actually. Interesting. It feels into some of the thoughts I've had about the way we actually think about marketing and say, for example, Mathew Sweezey great guy. I was introduced to him by a colleague of mine and we had a great conversation and he turned around and said, you know what?

[00:01:31] I know a great guy that you should totally call. And I said, set it up. Let's have a chat with him. And we jumped on and here we are. And at some point. A little later on today, I want to segue into that particular method of human to human connection and how machines currently find that difficult.

[00:01:50]Paul Roetzer: I love it. All right. So let's, I want to get into that too. Let's talk about your origin story first, because it was fascinating. It's always, when you think about [00:02:00] AI and, you know, I came up back 2011, 2012, learning about AI by looking at other industries, like, what are we doing on wall street? What are we doing in logistics?

[00:02:08] What are we doing in healthcare? How is AI being used there and how may it eventually be used in marketing? And one of the things I took away back in like 2014, 15 was, it's only a matter of time. There's only so many people capable of building AI solutions in the world. And right at that time, they were focused on other industries.

[00:02:29] For the most part, the VC money hadn't come to marketing and sales yet to apply AI and the talent hadn't come to the marketing industry yet. And you kind of follow that path because you didn't start applying AI in marketing. So why don't you just kind of give us your origin story from the PhDs to like how you applied AI and how you fell into kind of the marketing space?

[00:02:53] Harry Syed: I've always had my, my eye on, on the notion of artificial [00:03:00] intelligence. I think it goes back to probably when I was about 12 years old. But, but back then, it was, there was a, a fascination with artificial intelligence and biomechanical engineering. Now that's still young to be thinking about those kinds of things.

[00:03:16] Paul Roetzer : Uh, unpack for me for a second biomechanical engineering. What, what is that discipline?

[00:03:20] Harry Syed:  BioMechanical engineering is one of the ways in which you would use it is how would you combine machine and biological tissue? Okay. Organic Tissue. And how would you combine them together? An example of that is a new form of prosthetic limb that can receive information from temperature or from touch.

[00:03:42] Paul Roetzer: So sense has giving, giving it senses through intelligence basically.

[00:03:47] Harry Syed: So, so you'll basically cognifying. Organic tissue. Okay. You're taking X you're, cognifying it with Y for a biomechanical engineering is potentially an exoskeleton organic, [00:04:00] organic tissue over the top of that. But it's running with sensors that are, that are cognifying by a form of intelligence that then links up to your neurological, uh, pathways.

[00:04:10] Right? And then you receive that information is translated. And I had a fascination about that when I was 12. And I don't know why it's interesting. It kind of just came out of nowhere. I mean, I had a lot of interesting influence as a young child. My father talked about engineering all the time and we often sat and discussed simple things like why light travels, the way it does as you know the reflection of light of a billions of billions, of light years and how that changes time and how you should think about time differently. But when you go down those roads, you start to think about other things and just like today, when you try to solve a problem. And when we look at creative thinkers and I've been working with creative thinkers at Sub Rosa, [00:05:00] they get information from everywhere. They don't right. They don't just get it from one place. They could be walking to a coffee shop and they will be inspired by something.

[00:05:09] And that really separates the creative thinkers from, I guess, people like me. I feel like I'm a logical, I've been told I'm a logical thinker as well. As time went by. I continue to follow the discipline of artificial intelligence, stayed consistent with it. The academics that I did were primarily focused on a lot of image, a lot of image recognition, and a lot of understanding how we can process images very quickly and extract information from images.

[00:05:38] Um, and what use cases of that would be in the real world. So, for example, I had a lot of conversations with, uh, a representative of Lockheed Martin who was introduced to me about image recognition for missile guidance system. Um, but for other kinds of disciplines, uh, using image processing, it was for medical [00:06:00] purposes, which was very interesting at the time.

[00:06:01] So I spent a fair bit of time working in the research field of wireless endoscopy and I loved it. You know, we took 20,000 images of the GI tract of the colon, and we identified 10,000 normal cases and 20,000 abnormal cases, which were pictures. And we trained the machine using neural networks to understand the difference.

[00:06:28] I mean, classified the actual, the actual diseases with a diagnosis, and then. We had a camera that was built with small little small, a little pill with small little cameras, you swallow it and we'll take pictures as it goes down your throat down to through your fecal tract, into your stomach and through your intestines and then passes through your system.

[00:06:51] But those pictures will be uploaded wirelessly. And if you have a stomach ache, or if you're on vacation, your doctors here in New York, but you're in Australia, you know, having a great time. [00:07:00] And you seem to have some kind of issue, you swallow this pill and it will upload the images and it will analyze them and send the results to your doctor in America.

[00:07:09] And we were testing that anything between 88 and 97% accuracy on diagnosis. Without a doctor, without any kind of endoscopy without anything just following a pill.

[00:07:17] Paul Roetzer: Yeah. And what year, like what years are we looking at here?

[00:07:20] Harry Syed: That was, that was published in 2000. And.

[00:07:24] Paul Roetzer: Yeah, I think that's a perfect example.

[00:07:26] Like I always tell people you're like, when they think AI is crazy, it's abstract and Oh, okay. I get it. I can write email, subject lines. It's like, no, like 14 years ago they were doing things like this with AI, like the marketing industry is decades behind the most advanced things AI is capable of doing.

[00:07:46]Harry Syed: It is and why, and that's, and I guess that actually, I guess that segues into how I got into marketing and I, I wanted to continue studying, and it was, it was [00:08:00] extremely exhausting to, to work in the medical field research, knowing that your ideas would sit in approval for decades. Right.

[00:08:10] So although the innovation was there 14, 15 years ago, 20 years ago, So get something approved will take forever. And I, I kind of loved the idea of coming up with an idea and seeing some kind of tangible result, um, and then playing with it and testing it and then sharing it and then having people play with it and, you know, creating that kind of environment, that community with good ideas.

[00:08:31]  And I stumbled upon  another friend of mine who, who said that he had just launched his own website. Now we're talking about back 2007, 2008. And he said, you know, what would be freaking cool if my website could just change because it reacted to somebody's personality,

[00:08:56] Paul Roetzer: like knew who they were and could profile [00:09:00] them from a personality perspective and adapt.

[00:09:02] Harry Syed: He was a dentist. He was a dentist that just had an idea and he stopped. And he said, what if my whole website could just, my dentistry website could just change if it reacts to somebody's personality that came to my, my dentistry website, venues, teeth whitening this, that, the other. Why do I have to show them everything?

[00:09:20] Paul Roetzer: Why are they here for vanity? Are they here?

[00:09:23] Harry Syed: I don't want to show them all the things. Yes. I will not sell them, but I don't want to, to the mall. I want to grab them first and I want to get their attention. And I think that's where it just hit me. Well, I don't think that that's difficult to do. I mean, you can just start with some basic JavaScript and you can mock something up and put it into your, uh, into your host file.

[00:09:39] And then we can do a reaction test script, and then, and then ideas started crawling out and that worked. You would just recognize them an inbound cookie you'd measure. The time that inbound cookie spends on 'em. On a particular piece of content on the website, and then you would classify that story if that cookie returned within a certain time period, it [00:10:00] would just show them content that they spent the most time on.

[00:10:02] That's it? 2008, you said 2008, 2007. So that's what I started doing right around the time I was for a, uh, a data storage company, um, that were basically just taking paper and bring it to the cloud, but then analyzing that information as well. So, so that personalization concept really peaked my interest and then it kind of went from there.

[00:10:30] And then I met another guy and there was a startup on attribution. I didn't know, actually she was, had no clue whack vision was. And, uh, but when they explained it to me, the logic and the math sounded really interesting. And then they came up with this idea of this last click, this first click on tracking media and tracking how you should attribute dollars to a specific channel. But then they were like, well, what about best click? How can we work out? What best click is the average? [00:11:00] The mean, the sum, where do we come up with these ideas? And then that's where you start bringing the forms of intelligence.

[00:11:05] And you start bringing things like random forest or game theory and, and you just start playing with it. You start playing, you create a test environment with all of these ideas. And then you see the results and then you create the, what if scenario and you go.

[00:11:24] Paul Roetzer: I think what you're giving the example of is like what this is going to take for the industry to truly change and realize what it's capable of doing with AI is this combination of the vision to solve something differently, to look at a problem and say, okay, Well, there's gotta be a better way to do this.

[00:11:40] Right. And often that takes the business mind and the engineering mind that the person who can actually build the thing, because what I have found, and again, you, and I kind of just getting to know each other, but there's a lot of the technical crowd that really struggles to find solutions for their expertise.

[00:11:58] It's like, I can build stuff [00:12:00] like I can classify things. I can predict things like. Tell me what you need to do. Whereas me as a business person, I just look at everything and say, well, that can be done more efficiently. That can be done more efficiently, but I don't know how to actually build it. I just now look at things as there's inefficiency in a process or there's under-performance and I know there's smarter ways to do it.

[00:12:21] I just don't know how to build it. And that's why I was so fascinated to talk to you is because you actually seem to blend the two worlds a little bit. While you may not consider yourself a creative, you step into an industry, you understand some fundamentals and you immediately start looking at ways that you could make it more intelligent.

[00:12:41]Harry Syed: I love the, the, the, the play on finding patterns between things that just would never seem to have a pattern within itself. Um, and I think that there is, or there are so many wheels that don't need to be reinvented, but just need to be slightly more polished. And I think that there is a whole universe out there for marketing where low hanging fruit might seem complicated, but it actually isn't that complicated.

[00:13:10] Yeah. It all depends on the use case that you're trying to solve. Right. So, so. I've I worked, I worked in Dubai for three years for a Omnicom Media Group. And, um, I love, I love the Dubai is amazing. It's a, it's a place where thinkers and people with ideas really kind of descend upon in that region to create and build.

[00:13:34] And, and I, I remember speaking to a very large automotive company. Out there. And they said, we need an idea. We don't know if it's ethical from a privacy standpoint, but in this region we don't have any laws for privacy. So what can we [00:14:00] do? And it gets a bit interesting, you know, it's, it's interesting.

[00:14:04] So, so privacy policies were coming up all over the place. GDPR is kicking off in Europe, um, and other project policies coming up in the USA and that time they were like, we, we don't have that. There's no law here really is still a thing. You're not going to track it. Someone's actual first name and last name and their date of birth.

[00:14:22] You're not gonna do that, but you, you can get away with quite a fair amount of, of their use of behavior. So, so the guy was sitting with me and he said dealership, here's the brief dealership, um, social, email website, and signing up for test drives and someone walking in and actually like taking a car for test drive.

[00:14:47] How can you tie all of these together using intelligence, using machine, using some kind of automation and. My first reaction was you're a marketer. And he said, yes. And I said, I don't [00:15:00] see a lot of marketers that think like this, this is kind of outside the box. It's interesting. So unique. And, um, he said that I wasn't a marketer. I was an engineer. I turned marketer and I did a keynote speech about this, the CTMO, the chief technical marketing officer, the actual hybrid of the new form of resource, the new form of talent that is no longer just always marketing all the time and their career has grown, but that is now slowly shifting and changing where you might start seeing data scientists start becoming marketing managers, marketing directors, marketing executives, CMOs.

[00:15:39] I was once a data scientist because I, I really only tuned myself to understand the problem and how I can resolve it with some form of algorithmic logic, maybe. Right, but speed and efficiency scale. Um, But anyway, the idea for that automotive company was to create a heat score, a heat map, um, and give it a temperature.

[00:15:59] So if someone came to the automotive dealer's website, um, and they spent a certain amount of time on a particular model car, uh, the heat ranking would go up and depending on how they interacted with the car, uh, specifically what parts of the car they hovered over and the different, um, aesthetics, they want to change the car to all of these things were taken consideration and were immediately beamed over to the dealership. So when that person signed up for a test drive that the sales person at the dealership would say, Mr. Joe Blogs has a heat rating of 89 degrees likelihood to buy this particular model. You may not have it in stock, but talk about the aesthetics of the model like this.

[00:16:42] And a script will be delivered to the sales person that signed up.

[00:16:46] Paul Roetzer:  How many years ago was this?

[00:16:48] Harry Syed: 2013

[00:16:52] Paul Roetzer: because for our listeners. Cause we do have, uh, you know, we have some beginner and intermediate for sure. I mean, most of the industry I would still say is begin, are trying to [00:17:00] understand AI and just like wrap their head around this whole idea of intelligent automating tasks and making predictions new these things like.

[00:17:06]This to me is such a great example of the opportunity that exists. I didn't know, stuff like this was happening in those years. Like I was writing my second book in 2014 and. Wrote a section about what happens if AI comes to marketing and sales, like it's going to happen, but like what happens to industry when it does?

[00:17:27] And it was impossible to find people doing it at that time, the big tech companies, billions of dollars in IPO is for marketing tech platforms, not an ounce of AI in them, the major companies, not there didn't have a single machine learning engineer. Like these are tech companies with thousands of employees.

[00:17:44] And so like, the stuff you're doing my guess is a lot of listeners and viewers would be like, well, you can do that. It's like, no, you could do that half a decade ago. Like imagine what you can do now.

[00:17:57] Harry Syed:  Right? Yeah. Absolutely. I, I mean, you tell me, what do you think with your experience? And all the things that you've been looking into researching and people you've been speaking to is.

[00:18:08] Do you think there's some underlying fear? So when big data came in in 2010 and everyone said, big data, you need as much data as you can get. And then it's slowly kind of fizzled away and then said, well, you know, garbage in, garbage out. And then that became a thing. But then when, when the revolution of AI came around it quickly, I felt, got stunted because well, I don't want to lose my job or.

[00:18:37] This is so complicated. It's beyond my even understanding. I don't even know how to begin to understand this. And then, so the initiative to wake up in the morning and say, today, I'm going to come up with some cool initiatives for automation to scale, uh, and, and, you know, grow the company, looking at it at a different light.

[00:18:57]I felt people weren't doing that because of multiple reasons. I don't know if it was fear or just concern. No.

[00:19:05] Paul Roetzer: So, I mean, I've done over a hundred talks on AI around the world and you do. Almost inevitably always get, Oh my gosh, is this going to take our job? Like that question gets asked almost every time.

[00:19:17] So I have been under the assumption it's it's fear that it's fear of the unknown it's fear of job loss. Um, our recent research that we did with drift on the state of marketing, I. Doesn't agree. I think only like 14% of people said it was fear. What I think it is now the lack of education and training, by the way, was the number one answer by far, it was like 67% or something of people said that was why.

[00:19:40]My belief is it's too abstract. And like, this is why so much, like I'll often lead my talks with your life as AI assisted in your marketing will be too. And then I will explain Netflix, Spotify, Google Maps. I will show them a Tesla driving itself. Like. Bring it down to like you're using it [00:20:00] dozens of times, it is not scary. It's actually not that abstract. When you understand what it's doing, it's just math, like enabling these incredible things to be done more efficiently and giving the machine some human-like abilities, but it's not taking everybody's jobs. And it's not like if you just know what it's capable of doing. And you can find use cases. It's not scary at all, but I do think people just think it's a buzzword it's it's abstract. And like, eh, I don't know, like I'm not that concerned with it yet because I don't think they realize the leap forward it can give their company and themselves, if they're the one only one in the room, which is going to be the case in most instances, now that actually can look at something differently and say, you know what?

[00:20:45] I listen to this podcast, I took this course. I'm pretty sure we could automate this whole thing and have it get better on its own over time with the right person in the room who knows the right way to apply AI. [00:21:00] Just to be able to say that it's like, what are you, what are you talking about? But I don't think we have enough of those people that have got past that, whether it's fear or just like confusion, um, to, to start applying it.

[00:21:13] Harry Syed: It also might just be a natural transition. I mean, a lot of, a lot of people follow, follow the trend, setters, follow a trend. They, they trust a trend or they trust an influence or they trust something. And I, I read somewhere at the turn of the industrial revolution when people were washing clothes on a flat plate with ridges on it and probably clothes on it.

[00:21:37] And then someone turns around and says, well, I just took some electricity. I wrapped some copper around some magnets and I put a drum in a metal box, and then I put the power in and this drum site spinning. So I threw some water in there and some soap and guys just put your club, it's called a washing machine and be like, no, No, this is [00:22:00] madness and here we are.

[00:22:02] But then what will the jobs that were create? Okay, so all the people sit in their washing clothes might've lost their job in that citizen, but those very same people had to increase or grow themselves that capability, which was good for them anyway, because what do they do? They either became. Washing machine engineers, they built laundromats and they had all these washing machine.

[00:22:21] And so all these things started changing. People just adapted. We will always adapt, especially in, in, in the way we work today. Creatives. Okay. Creators will have their job icon. Imagine a machine anytime soon, using their instinct or a machine using its subconscious to come up with these wonderful ideas, Google dream aside, um, to, to come up with these amazing concepts that we see, uh, on, you know, terrestrial TV, or what we see at the Super Bowl.

[00:22:49] Right. Right. I don't know if the machine is going to potentially whip up the Oreo moment in a split second.

[00:22:55]Paul Roetzer: Intuition, creativity strategy. Like those are the things that I always say are uniquely human, and you can argue, can a machine be creative? You can see what they did with AlphaGo and like the moves the machine made and say.

[00:23:07] Was that creativity to like, then it's like, no, it was just really hardcore math in a way. Humans can't even wrap their head around that it did something you would never think to do as a human and I, so, but yeah, I mean, I agree. There's just these, there are these things, but it's hard to look out ahead, look around the curve and say, well, what are those jobs?

[00:23:26] I, I understand what you're saying. I get it. But like, what are they say? I don't know. But I can tell you there, there a bunch are gonna come up that we can't even imagine right now. Um, I, I get, I did a talk at a big healthcare company and I was like 150 employees explaining AI and intelligent automation of different tasks and communications and PR and people were like, What the hell like, do I have a job?

[00:23:52] And I said, this, listen, here's an example. Someone needs to figure out how to do all this. Like someone needs to understand community occasions and PR and [00:24:00] sit there and look at all the individual activities and say, yeah, there's definitely tools to do that. That, that, then they need to find the tech. They need to like figure out how to.

[00:24:08] Work with those vendors. They need to train the team, how to train the machine, how to like evolve their, see, like there's this whole operations and tech, like to your point, that kind of technical marketer. Those people don't exist. Like if you wanted to go hire one, where are you going to go? They don't come out with a degree in it.

[00:24:24] They're not being trained at corporations with it. So there's this entire landscape ahead for marketers to become like these next gen professionals, like totally different capabilities, but the fundamentals of understanding marketing and business would still be at the core. We'll always be at the core.

[00:24:42] Harry Syed: Yeah. People were still peak consumers and we'll still work the way consumers and worse. Just, you might find a different way to sell to them. You might find them a different way to target them or understand their motivation, but it still is going to be the same thing. Yeah. Um, to understand the very function of a brand, uh, what, what a brand stands for, [00:25:00] um, the competitive landscape of that brand and also the audience that want to buy from that brand or from that organization.

[00:25:08] I think those things would definitely stay the same. Um, But the framework of, of which those brands and how they operate the frameworks that they sit on the frameworks, in which how we understand the psychology of all the motivations of audiences and how quickly that they're moving around and changing their minds, that framework yes, will most likely be machine and be more machine and more automation, more cognification more intelligence that will keep coming up.

[00:25:37] Um, how we interact with it. How we educated, how we educate ourselves. These are going to start to become more compulsory in our roles as we move forward. That's for sure. So when you look at it, I mean, you've been working in the marketing space. It sounds like tennis years may be like applying your knowledge and capabilities into the marketing space.

[00:25:59] Paul Roetzer: You've seen a lot, a lot of the things you've probably seen and done would be mind blowing to people. Like, again, some examples you already bringing up are things I'm guessing a lot of people haven't even considered, but when you start looking out. Three to five years from now, knowing what's possible, knowing there aren't many technical limitations that the, the technology and most part exists now to take leaps, it's just not being applied.

[00:26:20] Like what do you see as the transformational things in marketing? Is there, is there a Holy grail to you? Is there like the one thing that could change everything or do you look at kind of more. Like I break into language, vision, prediction, movement. Like those are my four parent categories. I often look at and think what would be the breakthroughs in those areas.

[00:26:39] Harry Syed: But do you have breakthroughs you think about, I mean, all the time, some of them go way beyond five years? Um, some of them probably sit in the 20, 30 year Mark as well. I mean, I, I, I think that there is a natural progression to all things that can make. [00:27:00] Not just marketing better. Right. But because when I look at marketing, I think of, I think again, I go back to, I think people connecting to people, um, and I think organizations in the next three to five years will start to adapt to a single.

[00:27:22]And I think that all of these marketing tech platforms, so many variations, I mean, you remember the LUMAscape you watched how it grew. Yeah. They just went from this to just this. And how did that happen so quickly? And then how many categories that exist five years? I think that's going to disappear mostly.

[00:27:45]And those disappear. I think that in the next three to five years, what's going to really transform for marketing is a central nervous system will be embedded in each organization. And that nervous system will be. The way in which it will constantly understand the psychology, the motivation of the people that that brand cares about.

[00:28:09] And once people to care about that brand, and I think it will be in the form of some kind of evolutionary edge computing. They will not just be cloud. It will be faster to react. It will be on-premise, it will be extremely, extremely intelligent to, to, to. Ingest data process the data, have the classifications on what it needs to do with the data, understand the use cases of, of what the brand, the brand strategies are, what the marketing, uh, historical campaigns, the whole nine yards, we'll all be sitting in this machine and it will be able to come up with these forecasts and these predictions.

[00:28:48] I also think that there is a huge tie up to that neurological system that we part of an organization to empathy. Now you can't really, you know, code [00:29:00] empathy into a machine. I mean, that's just ridiculous. We as humans understand that, but not the way us humans understand it. Right. Um, and, and so many variations of empathy.

[00:29:09] And, but what if there was a way that a machine could almost identify a way to help teach us what it's like to be. In the shoes of a consumer on a daily basis, on a weekly basis, on a monthly basis. What if a machine was able to help educate us, translate information to us that. Your your particular neurological nervous system that sits in your company.

[00:29:41] Paul knows you very well as you are as user on a daily basis. And thus it collects information and tailors it just for you. So you can understand the day in the life of a particular group, a segment or an audience. And by doing that, your [00:30:00] decisions are now fueled. By some form of empty the machine gave you the tools, but your own experience, your own instinct, your conscious mind, or sentient mind was able to take that information from the machine and then do something wonderful with it, such that your, your audience don't feel like they'd been badgered or hounded with, uh, with, uh, with, with, with, uh, with an idea or an ad or come to me combined for me all the time.

[00:30:25] Well, let me just spam, all these negative emotions come from not understanding your audience. Um, and I think that in the next three to five years, there will be a symbiosis by which machines will talk to their users and the users will be the CMOs and the marketing executives. I mean, they'll find a new way to learn.

[00:30:46] And, and I think that that, that then expands now when you use that machine and you go further, I mean, Take Neuralink. If Neuralink really, really kicked off in the next 10, 15 years, who's to say that you can't download a subconscious mind and transfer it. Who's to say that you can't put yourself your subconscious mind into that other humor and actually live in that shoes for a day and really understand.

[00:31:15]What, how they tick, what their motivations are, what is that just became a reality. It's not impossible.

[00:31:21] Paul Roetzer: No, it's not. And if you're not familiar with Neuralink, it's Elon Musk's company, where they're basically creating lace that lays over the brain to be able to.

[00:31:28] Harry Syed: Right. Because when we, when we think of, when we think of the human mind, what really, if you break it down to its bare bones, what is it?

[00:31:37] It's a lot of storage. It's electricity, right? Two of the most important things. A lot of storage, electricity, and then finding these together. There's a lot of input and we use our six senses for input. We store them, we remember things. And then somewhere in that gray area of our brain, we then create an emotion.

[00:31:59] And that emotion talks to our memory. It talks to our incoming senses. It even our outgoing data. Our emotion talks to all these things, inhibits our decision-making and we then call it a gut feeling. Maybe we then talked to as if it's a soul, but what if a machine had enough memory and enough power who's to stop that gray area evolving by itself?

[00:32:23] Is that not what Rico's world defines as a singularity? What do you say? 2045, 2029. Last I saw. I don't know if he's, I don't know if he's changed that, but he thought, I thought he said by the turn of the next decade, that. He saw the singularity happening, but do you, I mean, I loved your, so when you were explaining like our tagline for the conference in the first year, the marketing I conference was more intelligent, more human.

[00:32:49] Paul Roetzer: And I think you actually just explain the more human part of what in my mind I know to be possible, but because I'm not the technical person, sometimes I have difficulty [00:33:00] explaining how we can actually become more human through the integration of machine, because it can make us smarter because it can inform us of things we may be can interpret at scale, um, because it can free us up to think about the meaning of these things and the actions.

[00:33:15] We should take a re as a result like that's, to me, what the more human thing men is. If we do this right, if we use AI for good, we can create better customer experiences. We can personalize for them. We can free ourselves up. To think more freely to do more uniquely human things like that. To me, is the, the grand vision of what we can create in the industry and at a broader in the society.

[00:33:40] But I also have these great, great fears that these same powers can go very wrong. And I think that's part of the reason I built the Institute. Part of the reason for the podcast event, all of it is to like push these conversations forward because it's going to happen fast. These capabilities are going to the marketers have super powers and they're going to have the choice to do good or bad with them.

[00:34:01] And I just, I think we need more conversation around what's possible stretching people's minds about what's possible. And maybe this is like way too far for some people, but like, People are going to be doing this stuff like this. It may seem weird to you, but like it's possible. There are really smart people that can apply this technology and insane ways.

[00:34:22] And that's the reality. And I think the more, we just live in the reality and try and say, okay, how do we do it? Right. The better off we'll be as an industry and as a society. Yeah, absolutely. The possibilities are endless. It's just how we use them. And that goes back to the ethics of AI as well.

[00:34:37] Harry Syed: Something must have spoken a lot about, um, Yeah, ethics of AI.

[00:34:42] Very, very important as we respect the power that it has. And also try to put some form of control it to some of the amazing ideas some people might have that just might be a little bit. Right to your question, like to your point, or like, where's that line? Where's the ethical [00:35:00] line. Where's the legal line.

[00:35:01] Paul Roetzer: Those lines may blur more in the future as more becomes possible. And again, that's my big argument for the industry as a whole is like just at minimum. Have a base level competency in what AI is and what it's capable of doing. If you're not the one that drives it, that's fine. If you don't ever work with somebody to build a, an algorithm, a machine learning model, you don't ever drive personalization.

[00:35:25] Like that's fine, but you got at least understand that you have to give yourself the knowledge to make the choice, not to go further. Like if you get along and you say, all right, I'm good. This stuff's just nuts. Um, I'm retiring in 10 years. Like let the next generation worry about it. But I, I just, I don't know that we have the time to do that. I think this stuff's going to happen fast.

[00:35:47]Harry Syed: It's going to happen fast. I feel like I'm seeing it happen with identity resolution of IDs multi-device resolution, or, you know, just [00:36:00] resolving an ID across social channels. I think that at the rate and pace. Of which privacy is coming in and, and, and stopping all of that from really happening.

[00:36:11]We went from statistical to deterministic. We're going back to probabilistic statistical again. Um, there will be better ways in which we will resolve a form of ID, right. And even if it's not about resolving to an ID that whole ecosystem will in itself evolve on how we understand, um, Who it is we should be targeting and why we should be talking to them.

[00:36:37] And I think that gives a complete different meaning to right message right place, right time, right person, running that completely evolves as well. And I think when that stays, when the realms of ethical AI and ethical within the realm of legality as well, we will see a huge evolution of the universe of marketing, which I think slightly ties back to [00:37:00] what I've been building as well.

[00:37:01] Yeah. Moving away from just IDs, just completely moving away from IDs and understanding people.

[00:37:07]Paul Roetzer: I think that's a great place to stop. I mean, obviously we could go, we could spin off here talking about that. And I think we'll just have to do this again. I mean, this is awesome. It was everything I was hoping for, like in this conversation.

[00:37:18] Um, so we'll, we'll wrap it up the way we always do it the rapid fire questions for Harry,  all right.  I've got a couple unique ones here. These are, I've never asked these ones before.

[00:37:27] Um, your LinkedIn profile says you've held roles in leading marketing technology efficiencies for Pepsi, McDonald's, Unilever, Mercedes, and Booking.com. The companies you've done some work with, or maybe a company, you know, you haven't done work with yet. Who do you think, which company do you think is the most innovative when it comes specifically to AI, but just most innovative overall, like, which are the ones that you look to as just leading the way or the one company in it?

[00:37:58] And again, maybe you Google it, maybe Facebook, [00:38:00] maybe Microsoft, you've worked at a lot of companies. Who do you see as the most innovative?

[00:38:05] Harry Syed: I think McDonald's system. Yeah. I, I think McDonald's is up there. I think that they. With some of the interesting acquisitions, um, in the last few years, uh, the way in which they understand the power of data and not just its data, but the bigger question, why do I need this data?

[00:38:29] What problem is it going to solve? That is the beginning of anything a machine needs. And that's the same with you. Same with me, or same with anyone else. If you walk into a room, you're going to analyze the room and then you're gonna decide whether you want to stay in this room when you want to leave this room.

[00:38:44] The same could be said about, okay, I'm an advertising company. I need data. I'm not going to just bring in all of this information and put it in boxes and then start analyzing it. You have to ask questions. And I think McDonald's is very good at asking the right questions to bring the right data for the right reason.

[00:38:59] And I [00:39:00] think for that alone, when they start putting a lot more cognitive location and intelligence on top of the right data, they have, they're gonna be able to make some very, very interesting decisions as anyone. Yeah.

[00:39:13] Paul Roetzer: Cause I, uh, I know Dynamic Yield was one of the companies they had acquired and I know there's a couple others, but I, if I'm not mistaken, they were getting into like adapting the menu on the drive-through based on the individual, based on the weather like that they had learning in it so that you would, everyone doesn't see the same drive-thru like little things like that.

[00:39:32] Harry Syed: The thing is, it's an interesting use case. There's, you know, 300 million people. And I think their target audiences, everyone that's 16 years old with a mouth and above. Right. So, so that's pretty much, that's a huge group of people. So how do you, how do you make your service more, your offering more unique.

[00:39:48] And I think that that's where they're innovating a lot.

[00:39:51] Paul Roetzer: All right. Favorite place? You've spent time in again, LinkedIn, you say, I know you're born in London, live in Brooklyn. If I'm not mistaken, spend time in Hong [00:40:00] Kong, Sydney, Dubai, Amsterdam, Madrid, London, New York, where you've worked in different areas, but favorite place.

[00:40:07] Do you have a, do you have a favorite.

[00:40:08] Harry Syed:  I want to say, uh, Hong Kong close Hong Kong with a close match to Indonesia as well, Jakarta, Bali. Um, I would, I would, uh, they're all different in their own little way. I think that the, uh, the most interesting people in a square mile that I've met has been in Hong Kong.

[00:40:32] Paul Roetzer: That's a cool way to put it

[00:40:33]Harry Syed: the most interesting people in a square mile, a square mile I've met has been in Hong Kong and

[00:40:41] Paul Roetzer: Amsterdam, probably depending on which square mile you walk in.

[00:40:45] Harry Syed: I was scattered all over the place I was living in wench. I spent time in TST in downtown Hong Kong, as well as central and I, I, everywhere I went there was.

[00:40:56]There was an innovator. It was an inventor [00:41:00] of it. We don't need more lawyers and accountants. We need more innovative thinkers.

[00:41:05] Paul Roetzer: All right. Voice assistant. You use the most Alexa, Google assistant. Sorry. Don't use them.

[00:41:10]Harry Syed: All right. So, so when my friends come up to my apartment in Brooklyn, um, I tell them my house breeds my apartment three.

[00:41:19] And they're like, what the hell are you talking about Harry? And I said, Alexa jacked into pretty much everything. I have a sensor on the front door. And if I come in after 10:00 PM and I'm just too lazy or tired to find the lights after 10:00 PM, the lights come on at 20, 30% dim, Alexa controls all that.

[00:41:37] And if I don't. She, she already has a geolocation of my phone. She picks it up. My projector screens. My sound, my light switch is, uh, I think about 65, 75% of the house is completely run by Alexa. Um, Siri is only when I'm driving.

[00:41:55] Paul Roetzer: Okay. Interesting. All right. Um, [00:42:00] I think I know, yeah, it's more valuable in 10 years, a liberal arts degree or a computer science degree?

[00:42:05]Do the machines just create themselves in 10 years? And is computer science not needed or is the do the creatives rule the world?

[00:42:13] Harry Syed: Um, computer science or liberal arts, to be honest, I think with take I'll take, uh, my nephew is 13 years old. He lives in London. He knows more about IP addresses and networking than I probably do now. I think that children that are growing in today's society are already almost nearly pseudo qualified computer scientists already. Right. They fix problems that we weren't even thinking about. We were running around, playing with sticks and riding our bikes and falling over and hurting our knees.

[00:42:56] It's very different society. Now, I think. With all of the [00:43:00] technology that they have in their hands and have access to I think the way to balance that out would be liberal arts and inject a bit of creativity in your life. Take what you know about the machine and then find creative ways to evolve and mature the world around you.

[00:43:18] But use that with liberal arts.

[00:43:20] Paul Roetzer: That's awesome. Yeah. And I actually thought about that way, that, that, that, that generation is just going to come up with all of these innate abilities because of the things they experienced growing up too. It's like, you might not need it. And the no-code revolution may get to the point where it's all about the idea generation and telling the machine what to build and like fascinating.

[00:43:40] Harry Syed: I mean, my friends have got children that are five, six, seven, eight years old, and they're teaching them coding as part of a hobby as an afterschool class, it's, they're becoming computer scientist. Right?

[00:43:49]Paul Roetzer: My son, when he was five using the Swift app on the iPad, creating the little video games, the monster going after the gems and teaching them how to code man.

[00:43:58] All right. Last one. Matter of fact, over the next decade, more jobs eliminated by AI, more jobs created by AI or it's a wash?

[00:44:06] Harry Syed: More jobs limited or created by AI. Okay. All right. There's two ways of looking at this first way of looking at this is, um, it depends on how stubborn you are as a human. If you're really stubborn.

[00:44:24] You're going to lose a job. There's I think that's probably, if you're, if you're not willing to change or adapt, you're gonna lose jobs. If you're willing to change, adapt, and learn and develop and see the silver linings in all of the machine that's coming out. And, you know, the supercomputers that are on their way, quantum computing is about to become a utility soon.

[00:44:43] Um, in 10 years time, if your mindset is such that I can find a unique way to be the architect of how this machine operates in today's world. It will create a lot of jobs for people. On the other hand, if you're just like, no, [00:45:00] I enjoy flipping burgers, which is fine, but Misa robotics will come in and take your job.

[00:45:05] Right? I mean, it just depends how you look at it. You can either be the guy that stops looking burgers and now starts to imp help input parameters into the Misa Robotics. So it flips burgers. So you are now a user of the machine, the Arctic to the machine. The machine needs you to give it input. You have a solid job because the machine will always need that input or you just become the person that says, no, I want to keep doing this.

[00:45:27] There is an eventuality. You either behind the ship or you're in front of the ship, it depends where you want to be 10 years. A lot of inevitability is going to happen.

[00:45:35] Paul Roetzer: That's I, I often, sometimes I'll show the slide of like, I graduated from college in 2000 and I just have the slide of all the things that didn't exist yet.

[00:45:43] The iPhone, Netflix, Bitcoin, Tesla,  Spotify, like all of it. Like none of it, Google was a year old, you know, like Amazon sold books. It's the world changes fast and it's going to keep changing fast. All right, man. Any, any [00:46:00] final thought for listeners? Just like again, so many of ours are the marketers who are just trying to figure this stuff out.

[00:46:07] Haven't run a pilot project yet. Just any guidance for them, any, you know, thoughts who might be motivated, be like, I got to do this now.

[00:46:14] Harry Syed: Yeah, absolutely.  I think that there is, um, a difference between the. The technologists that are academics and they it's their full-time job to be academics, but they don't understand your world.

[00:46:32] Because they work in embedded research and then there are the innovators and the, the thing, cause there's not enough of them to come in and help you guys understand and learn and understand your world and teach you the innovations that kind of happen. Basically. There's not enough of them. There's just, there's not enough of them.

[00:46:51] I think. That, what would really be amazing for, for marketers moving forward is [00:47:00] find new ways to  learn the true value of how you can innovate with automation, with intelligence, educate yourself. It's all about education at this point. The more you learn, the more you understand and how it's relevant to your specific use cases.

[00:47:20] Well, one obviously will explode your mind, but also too, it will start to help you understand it. It's really just to get started. It's not that difficult test environments are simple AWS and, and the Google Cloud Suite. Um, That utilities now power is right there at your fingertips, as well as storage. You have an idea.

[00:47:41] Find someone find a Wazniak in this case, find someone to come in and help you test it out and then see what the tangibles are. You can create. Learn from those mistakes are notable. Machines are not perfect. They probably won't be, but you would learn as you go.

[00:47:56] Paul Roetzer: Awesome. Well, thanks again, Harry. We will definitely do this again and I can't wait until we can do it over a drink in person someday. And this has been The Marketing AI Show. I appreciate you joining us today. We'll talk again soon.

[00:48:13] Harry Syed: Thanks again.