Hikari Senju Podcast Transcript

33
Headshot of CEO and Founder Hikari Senju

Hikari Senju Podcast Transcript

Hikari Senju joins host Brian Thomas on The Digital Executive Podcast.

Welcome to Coruzant Technologies, Home of The Digital Executive podcast.

Brian Thomas: Welcome to The Digital Executive. Today’s guest is Hikari Senju. Hikari Senju is a computer science graduate from Harvard University with a passion for technology and design. He founded his first startup, QuickHelp, an on-demand tutoring app, which he later sold to Yup.

com and became their head of growth. During his time at Yup. com, Senju observed inefficiencies in content spending and the potential of AI in advertising, which inspired him to establish Omneky in 2018. Senju’s background in computer science, entrepreneurship, and his experience with AI technologies led him to create Omneky, an AI platform that generates, analyzes, and optimizes personalized ad creatives at scale.

Well, good afternoon, Hakari. Welcome to the show!

Hikari Senju: Thanks for having me, Brian. Super excited to be here!

Brian Thomas: Thanks, Brian. Absolutely, and I’m super excited to have a conversation with you today, traversing the United States, not out of the country this time, but you’re hailing out of the great city of San Francisco, and I’m in Kansas City.

So, the 2-hour time difference is not a huge deal. So again, appreciate you making the time and Hikari, I’m going to jump right into your first question. Your journey started with QuickHelp, a tutoring app you founded and sold to Yup.com. How did that early experience in your entrepreneurship shape your approach to founding and growing Omneky?

Hikari Senju: Absolutely. So, there’s an interesting pattern where a lot of entrepreneurs started by starting tutoring companies or they were tutors as a kid. I think there’s a lot, there’s something to that. I think one, I think leadership is education. You know, when you’re managing something, when you’re managing somebody, when you’re, when you’re, you know, educating a market, it’s When you’re educating investors, when you’re educating customers, it’s, it has to have this mentality that I’m here to teach kind of an insight, a unique insight in the market that, you know, nobody else really knows you have some secret.

And that’s really what an entrepreneur is, is they’re kind of executing against a certain secret. They believe they know and understand about the way the world works. And so that was a really great experience because. I was running a tutoring business, managing, you know, having a, helping students learn, I myself also tutoring on the app.

It was an on-demand tutoring app from based out of initially based in Boston. We’d connect students in the Boston area with PhD students in the Boston area. And it facilitates these on demand in person tutoring sessions. And, you know, so I did a lot of computer science tutoring through the app myself.

And, and so that was a great experience, just getting a lot of teaching experience, tutoring experience through the app. And then I think the other big thing is. Education is a tough space. There’s not a lot of money necessarily in education compared to some of the other areas. You know, I think college students start generally start one of three companies.

They start a dating app, they start a food delivery app, or they start tutoring business. And, you know, there’s big names, you know, you can think of in the dating space, you can think of names like Facebook and Snapchat. Then in the you know, food delivery space, you can think of companies like Door Dash and Go Puff you know, all started by college students, but you can’t really think of any names.

Very many names in the, in the in the education space and the tutoring business, because it’s just, you’re competing as a free service. And so there I kind of cut my teeth learning about margins and being super frugal, the efficiency of, of every dollar and trying to stretch out that dollar as much as you can and fighting for every customer and, and trying to make the margins work in a space where, you know, you really are, it is a very competitive space in terms of, in terms of the resources and the budget that you’re competing against.

And so, I think those are the two things. I think one is. Learning to love to teach, learning to you know, learning how to teach or learning how to communicate and convey ideas and convey lessons very effectively and very efficiently and, and two was you know, learning how to run a business in a very competitive and, and kind of education is actually very cutthroat business.

I mean, you know, the biggest tech players are these for-profit companies and, you know, those for-profit universities are incredibly cutthroat. And so, it’s cutting your teeth in that kind of environment also teaches you a lot about. Running a business and how to be efficient and, and how to be strategic.

And so, I think, I think those two are the key takeaways I had from running quick help.

Brian Thomas: Thank you. I appreciate that. And of course, it’s almost like being in the military. You went to bootcamp and you just had to learn Navy seal buds training. You had to go through some of the toughest things to learn entrepreneurship and how it is to run a business and you were successful.

I think that’s amazing. And I love these sorts of stories. So, thank you. And Hikari, many people think of AI as primarily data driven, but Omnekey integrates AI into the creative process. How do you envision the role of AI in creative industries evolving over the next few years?

Hikari Senju: Yeah, that’s a great question.

It comes back to the fundamental thing that really piqued me about the space to begin with, which is, you know, can AI be creative? And when I saw the early generative AI demos or some of the precursors generative AI back in when I was a student at Harvard and cross registered at MIT, you know, seeing AI generate these images, I got very excited because for a while I thought that AI, you know, was just purely robots and automating, you know, already static processes and just a bunch of rules and just automate a bunch of rules.

What I saw a glimpse of in this lecture was AI being creative and my dad’s an artist. And so, I just found that like, if an AI can generate art, which is what I was seeing an early sign of back in 2014, then an AI, you know, has a potential of being creative. And then the other big point, I think that got me to start Omneky, you know, we started the company in 2018, but in 2017, Alpha go, you know, beats the reigning go champion, which is like a board game.

And, and it goes a very interesting game because there’s more combinations of moves on the board than there are apparently stars and you know, the night sky or something along these lines. Like it is a very intuitive and actually a very creative game. There’s no rule, hard and fast rules because the combinations of moves are almost infinite.

And one of the most interesting things is takeaways that Lisa Dole, who was the reigning champion when he lost to AlphaGo said was that this AI was incredibly creative. They were making these moves that no human had ever imagined, but, but kind of orchestrating these plays in such a unique and interesting way that that had never been seen before but worked.

And those were really the two things that I think got me as an impetus to start Omneky back in 2018 was. AI can generate images, and AI is going to increasingly be able to generate images due to Moore’s law and kind of the advancements of the exponential rate at which technology advances kind of had that intuition starting in 2014 and then 2017 seeing that AI actually can be quite creative.

It can make these really creative things that, you know, kind of outside of the bounds of, Of human imagination even. And, and that if they can do those two things, then, and then probably you can start generating effective advertisements. And so, you know, and it’s better to be too early than too late, I think to a trend.

And, and so definitely with the first mover in general AI very early, but I think it seems like a lot of it has played out and just playing out.

Brian Thomas: Thank you. I appreciate that really do. And I know you. You all were considered kind of some of the pioneers in this space, as you know, the last several years has just been explosive growth in this type of technology.

So, I do appreciate that. And like you, I was enamored by the creativity that’s able to be delivered through this technology. So, thank you. And Hikari, Omneky is known for its ability to personalize ad creatives at scale. Can you share more about the AI algorithms behind this and how you balance creative customization with efficiency?

Hikari Senju: So different people have different preferences in the way they want to be communicated. There’s Different types of images that they like to see. There are certain colors, there’s certain kind of linguistic patterns. You know, for example, if you’re an older person, you might want to see more words. And if you’re a younger person, maybe more brighter colors or potentially something more animated could be more engaging for you.

We also have just very unique trends in terms of visual preferences. You know, millennial versus Gen Z versus a boomer, et cetera. And so, and then, you know, you have all the demographic differences, and you have all the linguistic differences and culture differences, and even between Kansas and San Francisco, the aesthetic preferences are, could be different.

And so. AI can learn those preferences based on real world data. They can see how people are engaging with content, what types of content is driving clicks, what types of content is driving sales. It can combine those insights with the audience formation it has to, you know, AB tasks and generate different imagery and then kind of optimize and against a type of content that’s unique and personalized for that particular audience.

And that’s something that wasn’t possible before, you know, it used to be that, you know, human team of creatives, you know, you have thousands of designers potentially just, Producing pieces of content, not particularly in a data driven way, you know, not being really, you know, optimizing in a very real time way and it just being very expensive and slow process.

And so, with AI, you know, businesses can achieve a certain level of scale and efficiency that they weren’t able to before AI.

Brian Thomas: Thank you so much. Really appreciate that. And I like how you extrapolated that on a little bit. The way that these really these LLMs learn as they go, and they see how people interact and make decisions and some of those emotional things that come out.

It’s just amazing. So, I really appreciate you breaking that down for us. Some of the AI algorithms behind it. Your platform,

Hikari Senju: yeah, they can dive a bit further in terms of the LMS as well as we fine tune a model on the customer’s brand. And so, we fine tune LLM on the customer’s brand voice. So that’s an LLM that’s fine tuned.

We also fine tune a diffusion model, like a, like an image generation model on the customer’s brand aesthetic. And their product photos to ensure that even as we’re generating and optimizing the content against the target audiences, that the content, it keeps the brand sank or sank because brand safety, protecting the brand, elevating the brand, scaling the brand.

That’s, that’s the most important thing for an advertiser. It’s, it’s the job of the advertisers to scale that brand and have that brand connect with more consumers. And so. In a more intimate and authentic way. And so having, you know, fine tuning the model and AI model, the brand is what we specialize in.

We’re kind of like the, you know, mission control center of the brand. And, and I actually wrote this blog post recently called generative experience management, and that we historically lived in a world where it was more about managing assets, you know, the world of digital asset management, where you have these assets and folders, and you kind of share the permissions of the various folders with, with other people.

But we’re soon entering a world where there’s going to be a model. That sits on top of the assets, generating the assets in real time to be personalized scans for different campaigns and audiences’ platforms. And so, it’s going to be less about managing the assets, which will be generated on the fly. And there’ll be infinite assets to actually managing the underlying models, the power of these generative experiences.

And that’s what we focus on is, is building a platform for marketers and advertisers to manage their generative experiences. Is that’s what we call ourselves, you know, gem, the generative experience management platform. Yeah. Fine tuning the models against the customer target, us audiences and the brand to help businesses scale their brand experiences.

Brian Thomas: Thank you for breaking that down a little bit further and Hikari last question of the day looking ahead What are the key innovations or technologies you’re excited to incorporate into Omneky platform? How do you see the advertising landscape changing as AI becomes more integrated?

Hikari Senju: Yeah, so, you know, another I want to point another post as well, which is AI agents and advertising agents being, I think, the next actually probably the biggest opportunity in Gen AI when it comes to impacting businesses.

When you think of an organization and advertising organization, it’s not very efficient, right? There are layers of management communication. People are operating in silos. The creative team has certain insights. Marketing team has certain insights. The creative team screen, the content, the marketing team doesn’t have a lot of control over the content that’s being created, but the marketing team, you know, 55 the performance of the marketing is based on the creative, but they don’t have a lot of control.

And yet also at the same time, more than half of advertising budgets are wasted on an effective creative. So there’s a lot of inefficiencies in this kind of organizational system that is through the complexity of managing, you know, organizations of people. And so an AI system. Is real time it’s, it solves all of these inefficiencies, you know, in AI, I can, in real time, use an insight from marketing to generate a creative and then orchestrate an omnichannel campaign, you know, 24 seven in real time and drive better efficiencies and better results for, for that business, both in terms of cost efficiency basis, but also in terms of a results basis.

And so, I think we’re going to see an acceleration of, of automation. When it comes to certain types of tasks and roles, being more AI driven and just from a, from almost like a copilot world, which is, you know, the world that we are currently in, into, into more of an autopilot world, the full self-driving world, where the AI is, is really kind of orchestrating these, these experiences on its own.

And that’s really where generative experience management, our platform is so important is because we’re helping, you know, manage these generative models. But I think we’re going to see a transition from a kind of copilot to full self-driving world. And ultimately these businesses will therefore see better scale.

Better efficiencies and better results from the marketing. And those are really harnessing AI agents will be able to compete in, you know, kind of the next, or, you know, as this, as this plays out in the next couple of years versus the companies that don’t embrace this technology will be caught a little bit more flat footed.

Brian Thomas: Thank you. And I appreciate you diving into that gem, as you would call it, generative experience management. And I like that. AI can help be right there at the minute, those changes in emotion or other types of conversations are being held where you can actually do some creative on the fly with the help obviously of AI.

So I think that’s amazing what you’re producing out there to make the world a better place. And that’s what we’re after today. So Hikari, it was such a pleasure having you on today and I look forward to speaking with you real soon.

Hikari Senju: Thank you, Brian. It was really a pleasure for me as well.

Brian Thomas: Bye for now.

Hikari Senju Podcast Transcript. Listen to the audio on the guest’s podcast page.

Subscribe

* indicates required