In today's episode, Michael is joined by Hikari Senju the Founder and CEO at Omneky. He starts by discussing how he built Omneky, an AI-Driven Marketing Platform. They dive into Hikari's approach to working with customers on brand strategy and content. They also talk about the increasing importance of brands in a digital, AI-driven world. Additionally, they tackle Hikari's perspective on how generative AI will impact the advertising industry. Tune in on how ML is Reshaping The Advertising Industry.

Socials

Welcome back to another episode of Adventures in Machine Learning. I'm one of your hosts, Michael Burke, and I do data engineering and machine learning at Databricks. I'm not joined by my cohost once again. He he just doesn't like this episode, and so he decided to not show up. Of course, I'm joking.

He's on an off-site, for his software team and, unfortunately, can't make it. But today, we're speaking with Hikari. He studied computer science at Harvard and upon graduation founded QuickHelp, an IT services and consulting firm. QuickHelp was acquired by yep.com, and Hikari joined and quickly became the head of growth. But Currently, he works as founder and CEO of OmniKey, an AI driven marketing asset developer.

So, Hikari, starting with the most important question we'll talk about all day, how did being a d one fencer prepare you for the world of tech? Thank you. It's really exciting to be here, Michael. Thank you so much. Just as a quick correction, QuickHub was a education, application.

So it was an ed ed tech company that was acquired by .com. And, yeah, OmniKey generates personalized marketing assets at scale. And so Let me let me redo that. Interesting. Got it.

We must have the the name must have shifted. Okay. Cool. That's the beauty of having all this, edited. Education ed or edtech firm?

No. It's ed it's an edtech it's an edtech company. It was personalized learning application. Cool. And, yeah, that's actually kind of where I learned the value in personalization and the importance of personalization and Yeah.

Okay. Awesome. That that makes a lot more sense. Cool. So quick note for editor, we'll do that same intro and then start it off with the Hikari introduction.

Cool. Hikari studied computer at Harvard and upon graduation founded QuickHelp, an EdTech company providing personalized learning. QuickHelp was acquired by yep.com, and Hikari joined and quickly became the head of growth. But currently, he works as founder and CEO of OmniKey, an AI driven marketing asset development firm. So, Akari, how did being a d one fencer prepare you for the world of tech?

Gosh. I think there's definitely a a a, you know, a high percentage of competitive athletes being entrepreneurs. And I think, you know, there's that famous quote, which is, you know, starting a company, it's like staring the bus and eating glass. And there's you need to have a high paying threshold, and I think being a, you know, competitive athlete kind of trains you for that. I mean, I distinctly remember, at Harvard, waking up at, you know, first of all, pulling an all nighter to finish a problem set, and then immediately after that, going you know, running across the winter, in in, you know, the the Charles River, in the Boston winter, make it before AM, and then, you know, doing these relay races where we push these, like, £200 slides, across, like, the the Harvard's, stadium chip.

And, and, like, literally puking after practice and then going straight from there to, to my, you know, morning classes, then having practice twice a day and and that continuing for, you know, for for a couple years in college. And so, yeah, like, that ability to push yourself and and be surprised by the like, that you have it in you to push yourself further than you thought imaginable and also the ability to kind of eat pain, I think is a pretty critical and valuable skill also to have as an entrepreneur. And so, yeah, I I that would be my answer is just, you know, the ability the pain being a better athlete makes increases your pain threshold, which then, helps you, be a better entrepreneur. That's so interesting. What pain do you face that is similar to that physical type of pain?

I guess it's mental pain as well, like waking up at 4 AM. But what do you face in your professional life that is equivalent? I I it's the the physical pain and the mental pain and the stress and the high pressure environments, you know, I I think it all it come it's all mental at the end of the day. Right? Like, at the end of the day, the the you know, when you're, completely gassed and you're, you know, pushing this, sled or you're, you know, trying to do these hand cleans, but you're, like, completely disorientated and and sleep deprived, and your body is literally just falling apart, and you're like, no.

Mentally just block my mentally block it and just keep going. In some ways, being an entrepreneur, you know, there's so many things happening and, you know, there's so many there's so much incoming problems, like, you know, coming your way and just mentally blocking some things and, you know, definitely emotionally, you know, definitely, like, emotionally, like, compartmentalizing it or definitely even give me even, you know, emotionally feeling some of those those those concerns and feelings, but also just, you know, pushing through it. And there's just I I think, you know, the emotional pain or the the physical pain, they they're all mental at the end of the day. And so just ability to kind of withstand that, I think, is, you know, I don't know. For me, at least, it seems to come from a similar part of the brain.

Got it. Do you look to hire athletes over nonathletes? No. I don't necessarily look to hire athletes over nonathletes. We're looking to hire the best people.

And I and and I and, yeah. And I wouldn't say that, like, you know, it's true for every single world. But I think just generally being, like, the CEO of the company, and and and and a founder, I think just having high pain threshold and tolerance is pretty important to have. And so and, again, you can also develop that through other ways too. Like, you don't it doesn't have to be necessarily through sports, but, yeah.

Like and I think just also being very goal oriented and and and just being focused on a goal and and and and and over a long time frame, that's the only way you can really achieve excellence. Liking competition, I think, you know, businesses competition, even though that, you know, you try to avoid in competition, you try to find these blue ocean spaces where there's less competition, but there is competition still, and you gotta kind of thrive in competition and almost seek it and and enjoy that process. There's the importance of performing in during high pressure environments, whether it's a, you know, investor pitch meeting or customer sales call or, you know, an important recruiting call or, just an important meet event during the you know, for the team, to rally around. You know, that that's sort of also true in athletics too. You you know, you you got a big, you know, game and you need to perform and the team is counting on you and and you gotta, you you you gotta you gotta perform.

And so, yeah, just thriving in high pressure environments also I think is is something that yeah, that I can explore is gonna help prepare you for. Yeah. No. I completely agree. I played sports my whole life, and I use so many lessons that I learned through sports, like, daily.

Like, even right now, I don't know what I'm doing right now that I learned from sports, but I'm sure there's some component. And, yeah, it's it's a really interesting playground to practice these types of skills, like being tenacious, having a high pain tolerance, being goal oriented. All of that makes a lot of sense. And teamwork too. I think that's the key other key piece, how to build how to build teams is individual.

No? It is it is individual sport, but we there is, you know, we we compete as a team against other, other, other, colleges. And so, you know, and and and actually, like, the your your scores are going to be added up. And so therefore, like, you are only as good as the weakest player in your in in your squad, because, you know, if if somebody else gives up a lot of points, then, you know, you everybody else would be great, but, like, that could be pulled you back. And so this idea that you're only as good as the weakest link in some ways and that everybody has to step up and everybody has to play their part and everyone has to, perform with excellence, I think is also true for startups too.

You know, you're operating within a really small team and everybody does really have to, you know, be, you know, 10 x. And, because there's, yeah, just because there's you're kinda constrained in terms of resources and people. It's everybody really has to be, punching above their weight and, how to craft and motivate high performance teams, I think, is also a big part of sports and and and startups. Yeah. That segues beautifully.

You've been an entrepreneur for the majority of your career. Why entrepreneurship specifically versus joining a large tech company? So I grew up in New York, near the IBM headquarters. My grandfather worked at IBM, and he always, you know he got me to love technology at a very early age, and just get really excited about the exponential rate in which the world changes due to technology. Now he'd always say that the world changes every 6 months.

And so I wanna be part of that, and I just felt like I would have the biggest impact as an entrepreneur than necessarily, you know, being caught in the machine necessarily, you know, and obviously, this is kind of extreme example about, you know, the idea of being some engineer that's optimizing, you know, button color on the 3rd page, something like, you know, that was not as, appealing to me as, like, innovating and inventing new products. And I just felt like I could have more leverage to do that as an entrepreneur, than, and really build really cool technology and and ideally move the the world forward through that as an entrepreneur and the kind of the the freedom that the, the entrepreneur, provides than, that than being at a big company. And so, you know, that that was really more of the desire for impact. You know, I just felt like, also, I'd move a lot faster in my career, if I could just, do these things, you know, in the context of of of of a startup. But, but, yeah, I mean, you know, it it really comes back down to, like, the desire to just build great products and have the the the the freedom to do that and and, and and build them powerful products and just the impatience of doing this.

And so that's that's really kind of where, the motivation being our prayer comes. Got it. So from a motivational perspective, it sounds like it's a great fit. Do you think you're actual I mean, you've seen great success so far, but do you think that your personality and your skill sets lend well to entrepreneurship style of working? For instance, being a generalist and being proactive.

You know, it's really interesting. Did I become these things because I'm an entrepreneur, or am I an entrepreneur because of those things? I mean, I think if I was in a different environment and, you know, I've had you know, I may have been a specialist. And and in some ways, I am a specialist. It's like a t you know, you wanna be a t shape.

So, like, you know, you're, like, an expert in one thing. For for me, it'd probably be, like, AI and AI applications and and and and start ups, and, you know, and then, you know, the the t is more broadly related to management and product and marketing and and, in advert yeah. Definitely, the t is also in advertising. You know? So so the t will be, like, the intersection of, like, advertising and AI.

And then, yeah, more broadly, yeah, CEO. You need to be really an expert in basically every single domain of company. So marketing, sales, product, you know, management, operations, finance, fundraising, etcetera. So, I have a question to that end, something that I've been thinking about a lot in my day to day. I have been becoming more senior at Databricks, and now instead of being sort of on the lowest or not lowest, but, like, on the hands on keyboard delivery level, I'm typically either one level above or sometimes even 2 levels above.

So now there's a chain of command of creating the product vision and the solution vision and then managing people below me who will actually do the implementation. What do you think are the key tenants for ensuring that that chain of command is efficient? Because as you alluded to, as a CEO, you need to be a specialist in everything. Well, you don't have time to be that. So how do you create sort of productive handoffs at each chain to ensure that there's the least amount of knowledge lost and you're actually getting a clear picture of what's going on under the hood.

Yeah. Organization structure and communication. I think it's it's about setting very clear well, first first of all, like, a very clear deliverable and communicating that very clearly. So you can have a vision, but if you cannot communicate that vision effectively, then, then it's not you know, your vision is always effective as your ability to communicate it. And not only to just, you know, print out in in in words, but really to write it down, create a doc, create a structure, so that so that everyone in your team can be aligned around it and that, everyone knows what they're, you know, you know, where they're running towards.

I think, and then and then the second thing is, you know, making sure you have systems accountability and creating the structure for accountability so that, not only do, you know, your reports know that they're, you know, meeting with not meeting expectations, but but then there's there's there's transparency, throughout kind of this chain in terms of what role they fit in and and and why what they're doing is important. And it's not just like, I don't know what I'm doing, and I don't know why I'm doing it. But, if it's this larger strategy, and and this is why you need and it's actually motivating when that clarity is provided. And so, you know, setting up systems accountability, communicating a strategy, clearly creating, you know, I should yeah. Like, those those would be some of the examples I think of being an effective management.

But all also for me as well, it's it's it's a, learning process that, you know, I every stage of the company is distinct stage, and you have different layers, you know, what's needed for from a CEO at a 2 person company to an 8 person company to a 40% company to, to an 80% company is is to, you know, a 100 plus is is different. And so, for me, yeah, it's it's a kind of a constant learning experience as well. Got it. Yeah. Yeah.

That makes a lot of sense. I it's something I have I'm still pretty bad at, so I was wondering if if there are some tips and tricks. But, yeah, I think context aware accountability specifically for the person's role, I I think is sometimes underutilized. Like, I know that I sometimes overstep bounds or am not as productive as I should be because I'm not exactly sure what my role is in a given project or even at an organization. And so having really clear delineations of what you should and should be do shouldn't be doing, you don't have to follow them, but knowing what's expected of you is so, so helpful in determining what you actually do day to day.

I think it's like being a coach, right, of a sports team Yeah. You know, of a basketball team. That's basically what a manager is, is that you're setting the strategy and and and and of, like, this is the game plan, this is what we're gonna do, but you're also coaching on an individual basis for each player and say, hey. This is where you need to step up, or this is where you're doing a great job and and celebrating the wins with the team, as they as they happen and motivating team and, you know, being part cheerleader, being part strategist, being part mentor and therapist. I mean, I think it's, it's like it's it's a multivariate you know, you you need a lot of, varied skills for for being an effective manager.

One example, especially in the startup context, I like like to give of being a manager is it's almost like you're in a burning building because, you know, things are oh, you know, it's or, you know, the the classic quote is you you're, like, jumping off a cliff and building the the the, the plane on on the way down. But so, you know, you lots of things are going wrong, but also some thing core things are going really right. And, and your job is not necessarily to fix all the problems. Your your job is in that environment to coach somebody to empower them to fix that problem. And so you also need to have a certain calmness in mind and a certain, equipoise and, a certain authority and a certain, clarity of communication to motivate people to be like, okay.

So, you know, the roof is falling down, but we're gonna, you know, solve it by doing this or and and and not doing it yourself because then you're not really managing, but actually coaching somebody. Okay. Now this is how you're gonna fix this roof. So and, you know, even in an environment where there's other, you know, maybe there's flooding happening and things like this. But, yeah, it's it's and then nurturing talent too.

I think you've answered the other thing to put in is, you know, being effective coach is is, your your it's it's a role of service. You're in service. You're you're you're the lowest person of the company. In many ways, you're there to support everyone. And so, and and to, you know, help them in their careers and and help them level up in their careers as well.

So, yeah, like, creating an environment where people can be the best selves. And also, you know, to to get work out of people that they didn't think they were capable of doing. So, you know, to say this is good, but I know you can do better. And then even and then having them surprise themselves in terms of what they thought they were capable of doing. And I think this kind almost ties into, you know, back to this idea of athletics and stuff as, you know, I remember, you know, the most you know, the the best coaches I've had in sports were people like this.

They were like, wow. Like like, they believed in me more than I believed in myself, and I got they got me to do things that I didn't think I was capable of doing. And so, yeah. Like, Yeah. I think that's that's another component of of being an effective manager.

Yeah. You've hinted at motivation a few times so far in the episode. How do you motivate people? I think different people have different motivations, and so you kinda need to understand what is it about somebody, about that that gets them going. And and it's not, I Sorry.

Let me refine that. Let's say in the last all hands that you had where you needed to, like, rally the troops, get everybody going, What angles did you take specifically for your organization? Well, I think it's I think you need to paint a vision of why it's worth it. Right? And, like, what did it like like, why do you wanna be part of this journey and why it's an exciting thing.

And you have you you want to paint that vision very clearly so that it so it's something that people can rally around. Like, why this is important work. And it's not, you know, people it's not about the salary or the money necessarily, and people can get paid doing, you know, working, you know, at at various companies. But, like, why are they choosing to spend their valuable time with with you and with and at OmniKey? And why is it OmniKey's vision and mission so critical and important, and why is it worth their time to, yep, work really hard to to achieve this.

And, and so I think it's just really framing, like, the why. And and then after you frame the why, then again, really communicating everyone's role in how we get there as well. So, like, you know, every role is important. The other example I could give that was, you know, there's so I I studied Latin in high school and middle school, and, my favorite quote is from Lenny, and it goes for Samedhay, Olam, and, you know, it's it's it's Aeneas' you know, his crew is, shipwrecked in Carthage, and he's lost he's lost, like, 2 thirds of his crew. And he goes in front of him whatever's remaining and says, yes.

This sucks. But perhaps one day, like, you know, it'll be worth it. We'll look back at this moment and with in a moment of joy. The the exact quote is perhaps we'll look back in this moment in a moment of joy. It's like, maybe in the future, we'll be, you know, in a happier environment.

We'll all be having dinner and having a fun time, and we'll look back and say, wasn't that so crazy? Wasn't that so tough? Could you, you know, could you imagine that we survived that? Would we and we persevere through that moment? Because, serves our heart, and it does require a lot of, you know, a lot of excellent work from from a lot of people, like, people really pushing themselves.

And, it is important to have context to say that this is, you know, part of the journey and also that, you know, things will get better and that, you know, this this this place that we're trying to go to achieve, this, is worth it. And when we get there, we're gonna look back in this moment and say, this was a great yep. It it sucked, but it was also part of the experience, and it was, it was, you know, a great experience to have. And so, just having that, like, also provide that long term context as well, I I you know, that's my it's my favorite quote, just in general life, I think. But, because, you know, it just says, you know, this too shall pass kind of like Yeah.

And it's not even this too shall pass. This it's like this will become a fond memory. And after become a fond memory. Yeah. Exactly.

Like, we will look back at this and be like, wasn't that so, like, crazy, but, like, not you know? I don't know. Because time also, like, makes things feel less you know? I don't know. Exactly.

This this will this will be a little back at us in a moment of joy. This will be a fun memory. Like, that was so crazy when, you know, somebody so and so will come up for you to fix this thing because Exactly. It was falling apart and yeah. It was, like, really shitty and and, and and really stressful during the moment and, you know but, you know, that that's that wasn't that so crazy that that happened.

Yeah. Yeah. I I think about that a lot actually because I I like to journal. And so if I reread a journal, it's funny how my purse current perception of the memory is different from the actual experience a day out of Yeah. That experiences.

And, like, staying up late for studying. Like, I remember the startup that I did right out of school, that was absolute hell. It was so stressful. I was weighing over my head. I was trying to do everything, working crazy hours.

And looking back, I was like, ah, good times. But in the moment, I was like, I wanna shoot myself. So it's a interesting paradigm. Lots of philosophical implications on that. But, yeah, that makes a lot of sense.

Yeah. I mean yeah. I mean, there's you know, was it the, you know, I would nowhere, like, compare this to, like, the expo you know, one of my direct is like man's word for me. Right? And, like, there is a thing of, like, how do you get through hard times?

But you have to be realistic that it is a hard time. And it's not being delusional. It's not being overly depressed, and it's being realistic of the situation. And so, I think being like this does suck. And, yes, we understand this sucks.

But, like, you know, we will get through this or things like this. Also, it's it's important to be realistic, but, like and empathize at a realistic level, with with with your team as well. Yeah. Yes. Yeah.

So doing a one eighty, talking about things that don't suck, Gen AI. It's the hottest thing. It's so fancy and cool and solves everybody's problems. Can you sort of explain at a high level how you guys use Gen AI at OmniKey? Yeah.

So we're the 1st mover in generative AI, advertising. So the company's founded back in 2018. I studied, computer science, the focus on AI at Harvard, and I was taking about AI course work at MIT. And and the and my dad's an artist, and so the initial impetus behind OmniKey was really this passion, or even the initial impetus and passion behind generative AI is this idea of generative art, which I found incredibly compelling and and just really, really exciting. And so the way we utilize AI and and general AI is is is several things, and it's really comes down to 3 things.

We help brands manage their brand. We find you to model the brand. We, help generate skilled marketing assets. So we use that brand model to generate copy, ideas, and imagery, and video that reflect that brand, and and is aligned with that brand. And then we have analytics platform, so customers can keep track of their brand.

And even as they're scaling this content, they can see which piece of content is performing and and not only which piece of content is performed, but why it's performing ideally. That's hard, but, using computer vision and multimodal AI and reasoning AI to just, get trends about, you know, certain pieces of content are doing well with certain audiences, because it contains these creative features, for example. And so, brand management, scale creative generation, and analytics, those are the areas that we utilize, AI. Got it. That that's interesting.

It's it seems like a very hot area, so I think you guys are super well positioned. I literally this week, I scoped 3 separate calls that for customers that look to build those exact same systems with Databricks. Like, for Databricks, we have the analytics pretty solved. You have to write some SQL, but it's good to go. My question to you first is, how do you think about fine tuning against brand assets?

Like, what does that process look like? How do you get high quality assets? Do you do it for different product verticals or over the entire org? Like, what's that process? Yeah.

So there's 2 ways that we fine tune today. One way is fine tuning an NLM based on the brand voice. So, you know, getting corpus of historical content that is in the voice of the brand so that as we're generating copy, that it's, reflective of the personality of that brand, the way they talk, the, universe of kind of, of of of, like, symbols and and and references they have, fine tuning on brand voice. And the second way we fine tune is fine tuning product photos. So, we're generating an image we're generating image of a product photo oftentimes, in in different settings.

And to do this, you need to fine tune the image model, and the mission generation model and the and the diffusion model on, that product so that it is as accurate reflection of the product with as, you know, few, distortions, as possible. And so that's how we fine tune, the diffusion based image generation model and and the large language model. Got it. Any tips? So a lot of the people that listen to this podcast are highly technical.

They're ML practitioners. That's a a big chunk of the the technical. They're ML practitioners. That's a big chunk of the the demographic. Do you have any tips for approaching these fine tuning processes?

Like, obviously, collect good data, run your stuff correctly, write good code, but beyond that, specifically for the brand awareness component. Yeah. I mean, I think it's important to have a process for providing feedback. And so we from from the customer. Because even when we're fine tuning, we have the customer, we may be generating things that is not, actually reflective of of that brand, or we may be mistaken or have not not have complete understanding of what that their brand is.

And so we have an approval dashboard where customers can provide feedback on the generations. And so it's the creative director, the marketing director, that can provide feedback to a general model to say, this is good. This is not on brand. And and we actually have, rules and guidelines regarding in our brand management section regarding the copy you can't use, words you can't use, imagery you can't use, and really, making sure that that the generations in is controlled. And it's controlled in the direction of the marketing director and and and brand director, and the creative director in terms of, what they're seeking.

Because, you know, the ideal magical moment is, you know, the AI generates copy and generates imagery, and the and the brand manager goes, this is exactly what makes my company, you know, my company. They're like, this is completely on point. And, to get to that point, it does require, some some feedback process. And so really making sure that that feedback from whoever's managing that brand, is, having that structure in place, I think, is very important. What percent of the time does the initial guidelines provided by the customer align with the end result?

And the motivation for that is, like, what percent of the time do you have to show them the art of the possible and have them say, oh, I like this. I don't like this versus them knowing what they want from the outset. I I yeah. I I what percentage? I'm not sure if I can give an exact percentage, but I I think it's more very, very often the case that the first time you're not generating something that is on brand.

And it does require some, some back and forth. And, yeah, like, if I say, you know, hypes in it because the brand I mean, I think this is really an interesting thing about LLMs and and AI with brands in general is that, you know, it brings these things to life. Right? So brand used to be a really static thing. It was like a one pager with, like, some rules.

Like, you know, this is, you know right. It's, like, very it's very, like, limited actually. Like, there's not a lot of information here. Brand book, couple pages. But then what it does is it brings that brand to life.

You know? It now has personality. Now it's it's like a person. You can have a conversation. It's, there's more, you know, vectors of of of of, like, freedom that this this this brand now has.

And when when if it's for example, you're building a customer's, you know, chatbot, for example, for the brand. Right? Like, there's many ways that conversation can can go. And so now these brand guidelines have to be a lot more richer and has to be a lot more, a lot more thorough, and they actually need a lot more data. And so, helping brands evolve from this very static world of brands where it was just like, this is couple pages of what makes brand and, you know, use some basic rules to now I'm bringing, you know, my brand to life with this outlet.

It's a it's it's it's it's a real person or or or, you know, as real as it can be, engaging with consumers on a personal level, and this is the richness of the personality that I'm trying to create. And and so, yeah, you do need a lot more information for this vision of a brand than than kind of the way brands historically operated. That's fascinating. I the I've absolutely love the analogy of bringing it to life because you're right. You have this sort of 2 dimensional, like, bullet list almost of, like, this is what we stand for.

But often a lot of the human, like, subjectivity, I feel like, needs to be interpreted and you read between the lines of those bullets. But with an LLM, you can actually see it. You can actually feel it. You can actually act upon those those items. Do you ever go back and try to enrich that brand description with the LLM?

Yeah. So we, so the the the the feedback that the customer provides based on the outputs that we, provide to them that is updated to the brand model. So we get a more more holistic and clear like, you know, every single piece of feedback that the customer provides through our approval dashboard gets up updates the base model. And so then we do get a more thorough understanding of the brand and and the brand information becomes more expensive. And in some ways, that is, you know, you know, a unique form of data for us to build to build a business and a motor around is that, you know, as the customer invest more time into providing and training our model to reflect them, you know, that is a high switching cost for them to then do the same kind of effort of training some other maybe, you know, there's another competitor that doesn't have that ability, but, you know, it just that investment of data and time that we get through that is is is, you know, is then becomes, benefits just the fact we can generate more relevant, content for them.

So, yes, always looking to enrich more data, about what makes that brand unique. Got it. Do you ever look to influence or sorry. Go ahead. I think I think the other component is that, like, it is an existing public that brand managers have today, which is that especially if you have, you know, a brand operating many different regions or languages, like, the brand sometimes takes its unique personality and character depending on that region or or or location based on whatever local marketing team decides is that brand.

But that's not something it's something that's contrary to the actual brand values, and and it results in an inconsistent brand experience depending on, you know, region or even the platform or, or the person who's seeing that content. And so brands want to, you know, are are seeking to create more consistency, in terms of, their, you know, their brand and their their purse you know, what makes their brand, you know, them. And so, this technology is also exciting opportunity for them because it lets them create a way a better way of creating consistency across all different agencies they may be working with, all different regions they may be operating under, or the 2 languages they may be working with, that even maybe as you're personalizing brand to the cultural context of that locality, it still hasn't lost the essence of what makes that brand that brand. Right. I could see how that would be a massive issue because McDonald's in one country is quite different from McDonald's in another country.

Yeah. It it is. Yeah. Exactly. Like, you know, one country, it's a high end thing and a low end in other countries.

Like, it's but, you know, and but, like, that cons you wanna create that consistency. Right? Right. So because, because yeah. Because if people are traveling then it's like, well, it's just Yeah.

Distinct experience. And so, and even if even if it is distinct, at least you wanna have control over it and be aware that that is something that's happening. Yeah. Right. So this is a really great point of there there's sort of a objective definition of what makes a good fine tune model for an entire organization.

So let's say, McDonald's in, Japan might be very different from South Africa, might be very different from Texas. How do you guys steer the customer's opinion of their own brand? Do they just are they always right, or do you try to sort of guide them into a specific type of brand model? How much, like, thumb on the on the scale do you have in terms of influencing what the brand outcome is? Well, the customer is always right when it comes to the brand.

I mean, you know, that that's that's for them to decide. And, yeah, like, we don't we, you know, we're we're yeah. Like, our tool is to help them scale their brand experiences. So help them help their brand connect with more people and and on a personalized level, but, you know, what makes, you know, I'm I'm I'm we're not here to tell you that you should be this person instead of that other person. I mean, that's that's their decision.

That's when we provide a lot of data in terms of what types of content is resonating, what kind of messaging is resonating, what kind of value pops are resonating in different platforms and the campaigns and audiences. And and so, that can provide feedback to the brand manager in terms of making decisions regarding, in which characteristics of their brand they wanna focus on. But, but, yeah, we're we're we're we're only here to provide data and, you know, empower them to scale their, you know, their their their brand. So, yeah, they're always right. The customer's always right.

Got it. Okay. Cool. What are some high level trends that you're seeing in the advertising industry that Gen AI is influencing? I mean, gosh.

I I would say, like, all the the generative content. Generative imagery, generative video to start getting out of a job? Yeah. Copywriting. I mean, it's interesting.

You know, there's that famous you know, there's you know, what we like to say is, like, you're not gonna be replaced by AI. You're gonna be replaced with somebody using AI. And so the copywriters are not out of job. The copywriters who can really harness AI probably have more work if you if you do it well, probably have more work and more pay pay better and doing a better job than they've ever been. And not only that, but also companies that are harnessing AI properly have probably more work and more scale and, better margins than they ever had.

So, it's, so, yeah, I don't so I wouldn't say the copywriters are at work. I would say the best copywriters who can harness this technology have more work than ever. The companies that can in the in the best copywriting companies or the best company agencies, you know, who can utilize this technology have more work than ever. And I think there was actually some research done, or I think I saw some issues to us that, like, there's more people working in marketing agencies today post chat gbt and pre chat gbt because now, they're just way more productive. And if someone's productive, then there's more demand for their work.

And so, yeah, like, I I think the other thing other component being is that I think when something is data driven, then you can measure and we can measure the outcomes. Then it's warrants a higher ROI. It warrants a higher investment. So whereas copywriting used to be highly not quantifiable and, you know, not as potentially as results driven. When you can quantify the element, then now you quickly differentiate who the best copywriters are from who are the best copywriters, then, and the people who are the best can say they're the best.

They can get paid more, and they can get more work. And so, yeah, I think it's actually creating more work as for for the for the for the best the best copywriters. Interesting. Do you think generative AI has a place in creative breakthroughs? And the preface for this is when I was interning at Tubi, I spoke to the like, I just asked around, tried to get on people's calendars, looked to chat with smart people just to learn what tech was all about.

And I chatted with the director of ML, and he said data is really great for making iterative improvements, but you still need humans to make those 0 to 1 leaps. Because by definition, there was no precedent for that 0 to 1 leap. It can't be data driven. So do you think generative AI can have that human creativity component where it could make these wild new pieces of content or, do something that is fundamentally, quote, unquote, human? Yeah.

I I definitely think it can, and I've always felt like it can. Back when, even starting this company and before starting this company, I felt like AI can be creative. And this is really what got me excited about the technology is that seeing AI being creative. And and the one specific example I have to give here is and this is what got me really, you know, almost was when AlphaGo beat, Lisa Dull and, in this in this GO champ championships back in 2017. That's really what with 2017, 2018.

I started companies. I this is like, oh my god. AI is really in the, you know, precipice of doing something crazy, fact, you know, good. What Lisa said when he was playing AlphaGo was that the AlphaGo was creative. That it was making moves that were unimaginable to a human but were brilliant and had, you know, like, almost unexplainable, but, like, something that a human wouldn't do.

It wasn't just mimicking a human. It was innovating and creating new, opportunities and and seeing the game in a unique way, and that's how, ultimately, how, you know, he he lost was or he he wants, like, one game, but, like, lost me to the game because the AI was, you know, what we define as being creative and creative being, you know, thinking of something novel, combining different themes and and inputs and data to to create something new and making moves that had never been done before, but we're working and we're effective. And so, I think AI can be creative. Maybe not creative in the sense of a of a strictly define you know? No.

But I I actually think I I do generally think AI AI can be creative, and and AI, will will continue to be more and more creative. Because, also, like, what is creativity about the ability to create something new based on combining many different influence and data inputs? And that's basically what an AI does is, you know, they go through, they ingest they sponge up as much information as possible, and then they, you know, come up with something in new you know, a novel combination of those inputs. And that's what human creativity is too when you kinda combine 2 different ideas and you create something new. And it's something that AI can do, and it's something humans can do.

And, but I I think the key component to mention here though is, you know, this framing of AI versus human. Because I'm not sure if that's necessarily the right framing for this technology. I think it's actually very limiting when you say, you know, the goal of AI is to replace all human labor or or or, you know, it's gonna replace all human labor because I think AI can be more than that. And if, in fact, if you just put the cap of AI value to, or or definitely more different than that, and you put the value of if you just commoditize human labor, that isn't gonna be necessarily maximizing value creation, right, because there is a certain cap to all human labor value. What I think is that AI is very good at doing things humans can't do, and it will create value infinite value in in novel ways that humans couldn't even do.

Right? Like, so, like, potentially innovating, you know, discovering new drugs or, you know, or coming up with, interesting stories that potentially, you know, would not not be, intuitive to to a person. So, yeah, I think we should be pushing AI to do things that humans can't do and that in areas areas that will increase value kind of in in a certain infinite vector than just being like let's just commoditize, you know, the existing vector of all human labor. Yeah. That that definitely tracks.

And if you just look at history, look at prior innovations, all these things augment. They're not replacements. And, yes, they do replace what people are currently doing or we're doing in history, but they then serve to augment future use cases. So if you think about the car, alright. Now you don't have to run everywhere.

Now you don't have to ride your horse. You can actually get there faster. And now people can do a lot more things, be a lot more productive. And Yeah. I mean, gosh.

Like agree. Like, you know, the idea there'll be a job called digital marketer. Right? I mean, that's like like, you know, 20 years ago, like, you'd be like, you gotta be joking. Right?

Or or what or, you know, or yeah. But there's a performance marketer or, you know, even me, the car creating the job with the car mechanic. So, like, you know, there's there's always gonna be a role for entrepreneurs to come up with new jobs, and come up with new ways of creating value based on, you know, under re underutilized resources. And I'm confident in the creativity of of of humanity and human, you know, entrepreneurs, to, you know, come up with novel ways of creating additional value. And they will, you know, they will create the, you know, they will create businesses based on the digital marketing jobs of the future or, like, they're not, you know, unimaginable things, that, you know they will recognize things in this paradigm shift as we move from into potentially an AI centric AI world.

There'll be new opportunities for humans, and they will put new jobs and new companies based on those opportunities and those, you know, just new kinds of problems that arise. And so, because, yeah, people are always there's always gonna be problems in the world. And there's always gonna be problems in business. There's always gonna be problems in and and not all problems will be solved by whatever the state of AI is today. And so as long as there's problems and as long as people are wanting to seek problems, seek solving problems, there'll be new companies and jobs and opportunities.

And so, yeah, I I think this is just a continuation of kind of the capitalism and innovation, and we're gonna continue, you know, accelerating through this trend, and it's not gonna be the end of yeah. It's it almost is a very static, stasis driven view of the world where you say, oh, it's Yeah. Like, I have more belief in humanity, human creativity than that. Yeah. And just evolution and, like, adaptation.

It would be kinda dumb to think that we would be just like, alright. I don't have a job Yeah. Exactly. File for unemployment. Like Well, also, like, I think it's like, the humans will be like, oh, I have everything I want.

I don't want anything more now. Yeah. Exactly. We're creating it. Yeah.

Exactly. Humans always want more things. Humans always get used to the current thing, and then they want more things. They would want new things. And as long as and I think this is biologically true.

Humans always are you know, they're seeking they always want you know, the way dopamine works, we always seek novelty. We're always seeking new things and new ways of making our lives better, new goals. As long as, you know, humans could be dopamine driven and and want new things, and there's gonna be entrepreneurs who come up with ways of of meeting that new need. And so, and then as long as businesses also you know? And and businesses are there to kinda you know, to to to meet those needs.

And and not all problems in business are gonna be solved. And as long as there are problems in business, there's gonna be, you know, potentially new problems that this technology creates, which we I think we all agree that this technology is gonna create new problems. There'll be new companies formed to solve those problems. And so, the idea that, like, progress ends or, you know, history ends or, you know, like, we just now, you know, all just become completely complacent people with no new desires, and businesses just stop innovating and try new things. I I don't think that's realistic.

Yeah. I completely agree. But people are gonna complain for a little bit. Like, people will just be displaced. Yeah.

I mean, Luddites, you know, I mean, there's this there's a good book. It was called, like, blood in the machine about, you know, the original Luddism and, you know, like, you know, the factory workers going breaking up these textile mills because it's like, oh, the robot's gonna take everyone's job. And so the idea of of of, you know, automatons or or kind of robot human like machines, you know, replacing humans, it's it's actually an ancient concept. I think the atomic the the word first word of automatonic is, like, prehistoric, and the idea of, like, you know, some kind of a a a inanimate object or nonhuman taking over human work. And so, yeah, like, the this fear that humanity would be replaced by, you know, by an automata is is is truly, I think, an ancient fear.

But, yeah, as long as, you know yeah. But as as long as we live in, you know, as long as humans continue being humans and we live in a capitalist system that's always growing, we focus on growth, then, I think that yeah. Like, I think the system is structured to create new opportunities based on, you know, entrepreneurial and human creativity. Yeah. I I completely agree.

And speaking of those opportunities, what are the opportunities you're looking to create with OmniKey in the next 2 to 3 years? What's the tech that you're looking to invest in and build? Yeah. So I think in the next 2 to 3 years, agents AI agents, I think are a big a big thing. What are the just out of curiosity and reveal however much you can, but what are the agent stacks that you guys currently use?

Yeah. I mean, it's it's interesting. Right? Because there's agents in in the context of, like, what agents could be, and then there's agents in the sense of, like, what agents are marketed as today. Agents today are kind of these, you know, it's almost like just just kind of like a like a robotic, you know, try to automate some kind of process, and you and that process is is essentially automated by a quote, unquote agent, and maybe that process is very is a very discreet task.

That's, so potentially it's coding, potentially it's research, potentially it's, some kind of chore that you have to do related to scheduling or emails or, and, you know, that can be described as an agent even though, you know, an agent really the idea of an agent just has some form of agency, and it has some form of, like, goal oriented behavior, and it's creative. That kind of agent, which almost I think is kind of synonymous in some ways with, like, AGI or kind of the idea of, like, a of a of an AI being, you know, reflective of being clear. You're given in a modest goal, they can achieve that goal is kind of what, in some ways, what makes human, you know, makes it more human like. That I think is a bit maybe a year or 2 in the future or 2 or 3 years in the future, but I think we will get there, because just the amount of innovation and work and resources and investment happening going into, you know, the entire stack is is such that I I'm I'd be surprised if, you know, we we don't we don't achieve that.

And so, yeah, in terms of, like, yeah. So I just named couple of use cases of, like, agent like tools that, we're utilizing, but I think it can be a lot more than that, and I think it will be a lot more than that. Got it. Okay. So in the next few years, you said agents.

What else? Yeah. So, you know, all forms of well, okay. Yeah. So, you know, digital content creation.

Right? Digital digital content being including code, including video, movies, games, like, all that's gonna be AI driven. And we're gonna see and and, you're gonna see more funner games and funner apps, and it's probably gonna make that digital world more hits driven in some ways. I think the lifespan of of of a successful application will kind of shrink because, you know Wow. Really?

That's so cool. Yeah. Because it's gonna commoditize in some ways the ability to create those things. You used you used to have to have a lot of resources to say build an application, but now anyone can do it. And so that will increase competition, which will mean, generally, you know, time frame that these hits exist will be should be short just because competition will will reduce that, and that kind of lifespan.

It'll it'll, in some ways, turn it more like music business. Right? The music business today is very brand centric, but the, the the the actual lifespan of, say, you know, a hit song is is fairly competitive and limited and, you know and and so, similarly, I think you'll have big brands that continue to innovate and compel create compelling new digital experiences and and applications and movies and shows and, but and people will be drawn to brands because humans psychologically, you know, there's that thing. Right? There's 160 people, like, the Dunbar's number, like, humans, you know, can only have 160 people in their brain, really in terms of managing relationships.

And a brand is really that. Right? A brand is just being one of those 160 people. So if if if a brand can manage to be one of those 160 people, then, then, you know, they're they're you know, you're more likely to get business because people trust you, and they're more likely to come to you for for for for for services than than somebody that they that literally do not trust and is not part of their mental village. And so, yeah, the rise of brands, the importance of brands, therefore, in that kind of world, it becomes even more paramount because that's really the thing that competing, you know, benefit differentiates you from.

The power of brands is more important when somebody's commoditized. And so as software and these things become more commoditized, the brands become more important. And, yeah, just innovative new drug discovery and and simulations, to to propel that. So today, I think we're still, the the amount of resources, I think, that can really be dedicated to creating simulations to to discover new drugs is still very limited, and it can be done significantly at a at a higher level. Just like there's, you know, it's today very difficult to simulate the entire human body with with a with with in a machine and kind of, you know, simulate all those different permutations where, you know, say, a drug interaction with human.

It requires a lot of computational power. But as compute, you know, becomes, you know, advances and and and and, then, you know, doing those kinds of simulations would become easier, and then you will have, a real, like, I think, you know, blossoming of innovations related to drugs that help humans live longer, live healthier, live happier, live more productively. And so I think you'll you'll see in some ways, like, you know, an evolution of of of human life through through that. Yeah. And then self driving cars, I think, you know, we've I think that will be probably, hopefully, widespread by in in 2 to 3 years.

I think part of that is based on regulation, but I think the technology is getting there. You know, I'm in San Francisco, and I'd probably see more, self driving cars than I see, like ride sharing. And so, yeah, I think and and then things are kind of unimaginable too, but, those are some of the things in the next 2 to 3 years. Got it. Your point about brand importance increasing is really fascinating, and that makes a really nice market niche.

If if you can, like, increase brand value and put that into the 160 people category, that's an insane step for anybody. And, yeah, you guys are it sounds like you're betting a lot on brands becoming more important. Right? Yeah. I I I I think yeah.

I think brand as the com as the competition retention becomes fiercer, the value of brands will become more important because when you have, you know, 10 equivalent pieces of content that are equally, interesting, like, which one are you gonna look at? You're gonna look at the one that, you know, all the brand they recognize or trust. And so, because people do like familiar things, and, yeah. So I think, especially as digital economy really, you know, turns even more competitive due to AI, I I think the at least digital digitally, the the value of brands became more important. Yeah.

Yeah. That makes a ton of sense. Okay. Cool. We're coming up on time, so I will quickly summarize.

We kicked it off with a sports discussion. TLDR, enroll your kids in little league, but, if you're looking to learn some soft skills, sports, provides a great analogy for that in sort of a safe environment to learn how to be tenacious, how to have goals, those types of things. In the start up space, your job isn't to fix everything typically as a leader. Instead, you should delegate fixing things and empower people to do this. But you also must ensure that the people that you're empowering enjoy what they're doing and you're nurturing them.

It's important to retain your talent. And the best leaders make people believe in themselves, even more than I thought was possible. Then with that, on the motivation side, it's important to clearly paint a vision of why the work is important and then also make sure people feel that they have a role and that they are needed within that role. On fine tuning, it seems like a human in the loop specifically for brand content generation is super, super important, and it's typically an iterative process. It's hard to know from the outset exactly what you want, and showing the user what they could potentially have is a great way to sort of take it in a stepwise approach.

And then for the advertising industry trends, Gen AI is definitely gonna be a huge driver, and then it's gonna be increasingly data driven. And finally, brands are gonna become more and more important. So, Akari, if people wanna learn more about you or OmniKey, where should they go? Yeah. Follow me on x athisenjuor@omnicky, o m n e k y.

Follow us on, Facebook and Meta, atomnicky and, and on LinkedIn atomnicky or or on link or follow me personally. I'll link it up at at hecari sanju. If you have if you want to sign up for a demo, and see, in in traffic platform, you can sign up at www.omniki.com. And if you wanna send me a note, my email is hiat obnicki.com. So, feel free to also send me a note as well.

And, yeah, Yeah. Really really, yeah, really excited to, yeah, speak with you all and, yeah, just have you, you know, you know, at least help as part of the journey. Yeah. And I'm sure a lot of our listeners will have some really great use cases, especially in, like, b to b and b to c types of platforms. So Yeah.

Awesome. Also, the fact that your email is high at OmniKey, that's pretty badass. Yeah. Yeah. I mean, yeah, it's And it's your name, but still Initially, it started off as a support email, and then I just made it my own email.

Like Yeah. So yeah. I was like because initially, I was doing all I still do we'll do all the support and stuff. Yeah. And then it was like, okay.

Well, I'm just spending a lot of time on this email, so I'm just gonna be explaining well. But, you know, actually, like, Hikari and I also have to be the rightest way. So Nice. Cool. Well, this was a lot of fun.

Thank you for joining Hikari, and until next time, it's been Michael Burke, and have a good day,