Ryan Steelberg – CEO of Veritone on Harnessing AI to Transform Media and Business

All right, welcome back to the Casey Adams Show. Today we are joined by Ryan Steelberg, the co founder and CEO of Veritone. Ryan, thank you so much for coming the show today.

Kasey, great to be here. Thank you.

Absolutely. So, you know, your, your journey as an entrepreneur has been quite the journey from, from building and scaling and selling multiple companies and now building Veritone, I believe, since 2014. And, you know, you all are at the precipice of this AI movement. And I'm just so blown away by the. Some of the technologies that you guys have launch, you know, to, to just jump right into it. I would love to start by saying, where do you spend your time today? 2025. You know, AI is this not only fundamental technology that has transformed the landscape, but you guys have launched so many unique products. And I'm curious to know what's catching your time and attention today given the advancements you and the team have made over the years in AI.

I think primarily it fits into two categories. Public sector, how we're now applying a lot of our technologies that we really cut our teeth on. Leveraging AI in commercial, really media and entertainment, including podcast companies like iHeartMedia, and how we've sort of been able to migrate those technologies to assist and empower state and local law enforcement, the Department of Defense and the Department of Justice. I'd say that's one and then two is really helping since we've been now ingesting, indexing and archiving tens of millions of hours of premium audio and video. Right. And indexing, leveraging our AI technologies. We're now working with some of the largest hyperscalers and I'll say foundational model developers, the big names that you know, and helping them actually train their next generation models. So when we go through this discussion, some of, you know, as an entrepreneur, sometimes you have a great clarity on what your roadmap's going to be and where market demand is coming from. But there's a weird, there's a weird dynamic that I've experienced several times in my career that when you achieve a level of scale in one area of your business, it could create a generic derivative line of business that at times could be equal or even greater than what you've done in the past. So I would say those two areas, the data, the data work we're doing for model training and our advancement into the public sector.

Very interesting. I think the public sector questions. I'll have so many that we get back to. I wanted to take a step back because I heard you on a podcast. Back in 2022, when I was prepping for this show, you sat down with a mentor of mine, David Meltzer, on his show. And I remember you talked about how at that time, your models and what you were doing were ingesting over 100,000 hours of content per day. And you brought up an example of if these specific types of companies were to hire staffers, they need to have over a million staffers to help with this type of data infrastructure. I'm curious to hear how that has changed over the last three years and why organizing and optimizing data is important and is becoming increasingly more important over time.

So you've heard this statement that data is the lifeline or gasoline fuel for running AI models. And it is. You know, let me kind of explain this in two ways. We started in media and entertainment. For example, Casey, this show we're talking about, this is your primary product offering, right? When you're producing the show, you're distributing this audio and video and you're hopefully monetizing it, right, and turning it into a career. But your main data, your main data for you is the show itself. So it's a little esoteric, but what I mean by that is when we came into this business and we started to apply AI to media and entertainment clients such as, again, iHeartMedia, Amblin Studios, NBCUniversal, those companies main data, both their primary and tertiary data, what is their primary product offering? The audio and video programming they produce. So the point I'm making is even for those type of businesses that are experts in audio and video for a completely different purpose, forget AI, at least they've had experience going back decades of how to store it, manage it. They went from on premise, legacy storage. Now they're migrating those data payloads to the cloud. Even those entities have an incredible need, an insatiable need for AI to help advance their model. But now, if you shift to an insurance company, right, or you know, a financial service insurance or financial services business, their primary product is not their data. Does that make sense? Yeah, but, and then, but, but even there, AI is incredibly important. How can we sift through claims faster? Right? How can we make the better experience? So the point here is, number one, you need to understand and get a good handle on your data first. Absolutely. Before you can start applying and get great yield out of AI, people are jumping at times and it's catching up fast. But your question, specifically, what's changed specifically from 2022 to now, is that people now are finally understanding that before you can get all the yield and benefit from applying AI workloads and payloads, you better have gotten, you better have transitioned and prepared your data sets so you can act upon those efficiently. And it's gotten easier, it's gotten cheaper, right? I mean, as we're, you know, now as cloud and compute continues to get cheaper and cheaper, it's easier for companies to move their data or prepare their data for AI to impact upon it. So I think that's the biggest change I've seen, frankly. We've been around for a long time. You touched on 2014. It was a little bit of the horse before the, I mean the cart before the horse where we were trying to sell AI solutions going back to 2014.

Wow.

And we were, in effect, we were spending the majority of our time helping companies just prepare their data before we could even apply the AI to it. So I think that's probably. And we're still there. We're the very early innings. I still think that the limitation and preparation of data is still a major limiting factor on fully realizing the AI potential.

I love that answer. And just to even double click on the 2014 timeframe, you talked about in a previous interview how past companies sort of have. Obviously there's an adjacent level of thinking towards last companies to current companies and how that overlaps to learnings and lessons back in 2014, 2013 era. When it comes to AI and the opportunity that you saw with Veritone as someone that has a multi town founder, has sold companies to Google, what did you see and why Veritone as an opportunity when it comes to this type of business, business, what was that sort of, I like to call it a white space in the founder journey where it's like you've built companies or maybe you're changing jobs, changing roles, where you have many directions you could go and then you end up focusing on the next journey itself, whether that's a new company, a new state you want to live in, whatever it may be. So for you, what led to wanting to pursue Veritone given the past entrepreneurial successes and why this opportunity back then?

I'm an ad tech guy, right? And so I'll go, you know, the quick history is, you know, I'm at school, I went to ucla, I was maniacally focused on becoming a doctor in kinesiology. I love sports, you know, I wanted to work with athletes, et cetera. Thank the Lord that the Internet opportunity kind of came upon us. You know, we grew up with computers in our homes and let's just say we're not going to go too long into the, you know, the impetus of, you know, why we went into the Internet space in 1994. But ad tech was our spot and we started to build some of the earlier foundational companies in advertising technologies back then, AKA serving ads really fast and well to everybody right online, good or bad, including the infamous banner ads of the old days. But through that journey, the whole thing about ad tech is it's data, right? When you are on a web browser or on your phone and you go to visit a certain URL, for example, data is exchanged, right? We have to serve an ad to you, Casey, when you're browsing somewhere on the web, right on the Internet and what we get is a limited amount of information and we try to infer what I can about Casey and what you're looking at. So if it's whether it's you're doing a search on Google or you're going to espn.com and you see an advertisement, we have to very quickly, brilliantly fast, in milliseconds, understand the ad, request and deliver hopefully a targeted ad that's relevant. Back to you. What I just described there is data, but it's been mostly structured data, HTML like IP addresses, inferences. So through our journey, you know, I've been basically building and thankfully successfully selling or monetizing different ad tech businesses because the world and technology keeps changing. My expertise is ad tech. So you have to keep applying, right, your expertise to the changing landscape and technologies. Case in point, we used to show off the size of our data centers to serve ads. Back in the day we actually had the previous company called Adforce, our very first big company. We actually acquired isp, Internet service providers, data centers because that's how much horsepower it took to serve ads today. I could have run the entire company of Adforce on a laptop. That's how fast technology has changed. So the point is for entrepreneurs is you don't always, you can continue to apply your expertise to a changing technology landscape, right? And that's what we were doing. That was our arc, the big, the big bomb drop. For us that worley was the early kernel to thinking about the veritone opportunity was the iPhone and the explosion of user generated content. Video. Video online was a game changer because video and audio is a form of unstructured data. We machines, we legacy machines couldn't understand what was inside the audio and video. So if people are spending more and more time online not watching structured HTML static websites, but they're now Consuming and listening to tonnage of video and audio. As ad tech people, we better understand what you're looking at. Right. So we can potentially contextually target ads. So this, so the, so the journey that ARC that ultimately led to the adoption and utilization of AI was really an ad tech problem. I. We needed to understand, Kasey, what you're actually looking at, not just the metadata that you're watching the Office, right back in the day, but we need to know what's inside the programming. So like many things, it was a problem that we needed to solve. And legacy algorithmic indexing wasn't working. And that's where you started to see the creation applications of neural networks and the nascent forms of AI models. Trying to look at those audio and video data sets and say, what are they saying? Right. As easy as transcription. We take it for granted today, but back then, right in late in 2007 through 2012, it was, I mean it, it was very expensive and took a lot of compute and horsepower just to. To create a clean transcript. Casey, you and I take it for granted you to transcribe this podcast now you can get a transcript that's 98% accurate. Cost you a few pennies back then.

Absolutely.

Back then it was 40% accurate. Right. And it would cost you $10aminute to transcribe something.

So yeah.

So Veritone was really the idea that we need to create. And to be clear, we created it with the mindset. Eventually this is going to be for advertising technology optimization. But the concept of could we ingest, index and understand audio video at scale could be applied to so many different use cases above and beyond ad tech. So that's kind of a quick journey of the arc. The key there for entrepreneurs is you don't have to abandon. I mean you can have as long as you're disciplined on educating and continue to educate yourself as technology changes. Going from legacy data centers to cloud AI, you know, there's countless entrepreneurs have done this multiple times where they're applying their expertise, they're continuing their improved expertise against a new technology arc that continues to advance, et cetera.

Yeah, no, I mean I appreciate the answer and the backstory there as well. And I think there's so much personal relatability just to give you some quick context. So even on the sports angle. So why I started this podcast back in 2017. I was in high school, I was 17. Nearly got paralyzed playing football. Ended up in a neck brace for six months. And that kind of led to having a lot of time on my Hand ended up wanting to learn about business entrepreneurship and I'm like, hopefully I can meet great people, learn from them and create some type of business opportunity through the show. And fast forward a couple of years. I was sitting down with the great Larry king back in 2019 as a 19 year old podcaster and sort of the podcast really laid the groundwork for now. I've had two different technology companies. One that I'm currently running is listener.com and then my last company was called mediakits.com, which was the easiest way for social media creators and Twitch streamers to aggregate all their cross platform social data when pitching and you know, pitching their data to brands. And when you talked about that concept of like taking past learnings to and sort of like it's a new iteration as a founder and there are different companies, I related that, related to that a lot because my current company, listener.com is very much so all about like AI powered listener analytics for modern podcasts and podcast networks, hyper focus in that niche and vertical. And I always say it's very adjacent to my last business. And when you started to kind of dive down that rabbit hole, it was not only like a confirmation bias of like focusing on skill sets and you know, the, the industry that I see here, but I think it's very parallel to a lot of other founders I've had on the show in terms of how, you know, their first company or first, you know, job leads them to pursuing their next larger, bigger opportunity that's very adjacent to their past learnings. And I'm curious to know if you've seen that with other successful founders and you know, leaders that you've had the privilege of working with or learning from in your journey.

I have, I think, you know, we, all of us, since, you know, news is everywhere and we have some very big personalities of entrepreneurs, the Elon Musk's of the world and others. I think they fall into a couple categories. One is, you know, the Zuckerbergs who came up with the killer idea and iterated that same concept. Right. For the next 20 years. Right?

Yeah.

Facebook, the Facebook, Facebook now meta. But fundamentally the gasoline and the engine that created all those different derivative businesses was primarily the social network. And so he kept the reins of that business as a single. I'm kind of on the opposite extreme where candidly we sold before we ever got to a certain level, the next level of scale. Veritone's probably the first one that we've ridden this horse longer than any of my previous businesses. For a couple of reasons. Primarily is just the size and scope and the magnitude of the opportunity is much bigger than I've ever realized that we've ever seen. Second is from a financial perspective, I can afford not to grab and cash out immediately. So I think, first of all, think about that is even in those extremes is I have started different companies but still stayed within, I'll call that band of media, entertainment, audio and video migration. The opposite extreme is Mark Zuckerberg who has obviously created this huge business and there's different derivative lines of businesses now, but he stayed with it so intellectually and with his passion, because he had success even within the same company, he was able to do different things. And I think that's kept him motivated and excited going forward. So I. And obviously then there's the slew of the 95% of entrepreneurs, unfortunately, that never get out of the starting gate. Right. Or they get out of the starting gate and the business fails. So I'll give one, just one example of like they were, you know, I would say maybe I was a little bit of a mentor of theirs is the Vanderhoek brothers. The Vanderhoek brothers, which you may know are the founders and kind of the run. And Tim is the CEO. Tim Vanderhook of dsp, right. That's their symbol. Dsp. Vayant Media. Vayant is the name of the company and it's gone through multiple different iterations. They started years ago, they sold it to time, they bought it back from time. They took a public and similar to kind of my background, but even more on a common thread. Right. They've kept more or less the same company. And as ad technologies has gone from legacy web websites to now ott. Right. And video and what you and I call TV today, that's another example of how that entrepreneur team has kept the same company more or less and have been able to adapt and grow that business line against the change in technology stack. So the point here is there's many different ways and flavors by you as an entrepreneur to be able to attack an opportunity. And sometimes it's you gotta stop and do a whole new company. Others is if you have scale and you have either investors or profits, you can iterate and try to solve or go after these new lines of businesses organically with your own business. So again, there's different flavors.

Yeah, that's very well said and I appreciate that perspective to change gears a little bit. Something you talked about at the beginning here was selling to the private sector, excuse me, public Sector And I'm curious to hear your perspective when it comes to, you know, selling to CEOs and into the C suite. When it comes to, you know, B2B sales or large scale enterprise deals versus public sector, what have you seen as the biggest difference there and how has that perspective changed over time?

I'm a moron when it comes to selling into the public sector.

No.

So again, so my entire career has been to selling into the commercial sector, right?

Yeah.

And you know, obviously you. Most of my products and services have been sold direct. I have not gone through reseller channels or the like. When we migrated and started attempting to sell, you know, several years ago into the public sector, it's a radical different ecosystem. You know, there are, you know, different procurement methodologies and processes. There's companies called like, you know, there's GSA schedules, there's companies called carahsoft. So that in effect think of them as marketplaces or different procurement methodologies and workflows that you have to go through so your software can be purchased by different law enforcement agencies, the Air Force or the Department of Homeland Security. So I think education on how the procurement process works, how forget how good my product is, even if I have somebody who wants my product, it perfectly satisfies the mission, whether it's lieutenant General or it's a back office. Right. Procurement officer. I have the solution. Getting educated on how you can prepare yourselves to sell that product is a huge learning curve. Also. It is not. You need to have the discipline and depending on how your financial model works, you have to have a lot more time to break into it. Now there are benefits if you get in. Right. It's, you know, you start to see some of the half life of some of these companies. Typically, if you have a good product, they have, I'd say more staying power, longevity once you get in. But getting in can be very challenging. So I think, I think it. What I've learned is sadly, there's as much demand in the public sector that wants our state of the art technologies. But wow, it's been very hard for them to get these technologies seamlessly. And by the time when they get them, ironically, it's months or years. If I take the same version of a technology, I can sell it to you immediately to empower your podcast network. We could do a deal tomorrow and you can start using it next week. And that's when the sad thing is that regardless of your political philosophies, the procurement process and the operations of our government to get them the greatest, it's a disaster. Now it's trying again. We're trying both sides of the aisle trying to streamline that. But it is challenging. That I would say is you take a snapshot in time. Our government is using old or second class of everything for the most part. Right? Yeah, Think of it that way. Which is again, we have to solve that. It's going to be hard for us as a nation to stay competitive if we're using, if we're using Windows 2000 for example.

Totally. I mean that whole conversation around public sector is so fascinating to me. And over the last two years, one of my good friends, former investors, his name is Brandon Max and he started a company called Pursuit Us and it's pretty much helping companies sell more effectively into the public sector. And I'm just a big fan of it and like learning about that space and hearing you talk about it on other shows and how that's a market, it makes me just think about like the company he's building and understanding that like large problem that it really is when it comes to like the discrepancy of, you know, selling to large scale enterprise. So very cool perspective here and appreciate you sharing that. I want to take a step back when it comes to your philosophy on team leadership and empowering your employees. Curious to know how you view when it comes to all things remote in office, how that's changed over the years and how early stage founders should maybe think about things in this new AI environment. You know, you, you, you hear that phrase of there's going to be this one person, billion dollar enterprise in the future with way AI and teams are more efficient. Curious how you view like team size, head scaling, headcount for founders and how it's changed in this new era of AI.

Yeah, I'm learning myself on all these too. Right. So that, that's the other thing. These, this, this world is changing so fast. We went through Covid, right. And obviously the world changed and now we're emerging from that. And then as we're emerging still from that, we have now the onset of AI and agents, et cetera, et cetera. So I think, you know, again, there's not one right way to operate. I think for me I'm, you know, I'm very personable, I get empowered, I get energized when I interact with people. And there is a difference of having Zoom calls versus getting face to face. Yeah, Veritone is a remote first company. I think we have employees in like 45 states. Right. So we're very diversified. So we've I've say we've taken advantage of being able to hire and manage what we deem as the best of brightest that we can afford geographically, not constrained to geography. However, I think it takes that much more discipline for not just me, but having really dynamic and engaged managers. I think the management level, although some organizations are flattening a lot of them, it just shows it is that important that you can continue to advance the engagement of your employee stack. So my style of management is, I'm extremely frank, you know, even at the risk of too much disclosure. Right. Because everything that you know like this is being recorded. You know, if we're doing a zoom call and I'm speaking to my hundreds of employees, you know, again, it just, I've always been, listen, if I can't sit next to you and help you sell and solve your problem, I'm not sitting here in ivory tower, right, speaking down to you. But again, so therefore I miss, I miss at times the human interaction, physical in person, human interaction. That's just me personally. So again, I don't think we're putting the genie back in the bottle, particularly for companies like us where once you've kind of purged ourselves of these expensive office locations in New York and stuff, trying to think about how I would reestablish that, you know, frankly, I'd rather spend more time trying to optimize the remote experience, right. And put those investment dollars in other areas than having these skeleton physical structures in, you know, all the top major cities. I think the AI thing, let me just answer that. I think this is a real big one. I think from an HR work workday, if we want to pick one of the biggest companies in, in the HCM space, you know, if you're an HR director, you're going to be looking at your physical people and your digital people. Right. I think it's real. I think, you know, we are all trying to optimize different areas of our business using AI. It could be again, people being experts, using the tools, but in other areas, truly. And this is being, as you can imagine, this is consultants dream where they're trying to now consult companies like Accenture and others where 30% of my workforce are going to be digital agents. Right. There's some truth to that. Right. We're seeing it right now. There are areas of the business that I think are going to be impacted faster. And I'm not going to go into details which ones I think, because again, I don't want to scare some of my own employees, but let's just say that there are, you know, you're going to see significant reductions in size of departments because they are optimizing or leveraging AI tools more effective, more effectively or aggressively. I will say one thing about the coding side, which is fascinating.

Yeah.

You know, coding, the irony is that some of the areas that are being the most impacted are ironically are the people who created the models in the first place. Right?

Yeah.

And think about it makes sense because coders can't. Their output, no matter by default is structured data. Does that make sense?

Yeah.

So a software engineer, their output is structured data. It has to be code can't execute, it can't run effectively. Right. If it's, I don't care what language. Right. Where it's Java, C Sharp or whatever, it has to have structure. When you have that much structured data and, and that much that's been stored and documented when you're training a large language model, a model, what do you think you can train and sort of emulate the quickest Structured data. Structured data. So that's why you're seeing now, frankly for us, when we've been fighting against the hyperscalers for hiring great talent. There's a lot of coding talent out there on the market right now, more than I've ever seen. And yes, you and I, we see the news like, oh my gosh, this data scientist has got a $100 million contract from Meta.

From Meta, yep.

Great. For every 20,000 of those, I mean it's every one of those. You are seeing an impact on entry level coders who are really having to compete, frankly with decisions from CEOs saying, Am I going to invest more in copilot towards writing code, cursor, whatever your flavor is, or are you going to invest more in people? I think you're going to see a balance there. So that is real. And I think in every department there's going to be a decisioning on how do I structure my organization with the onset of these new tools.

So great answer. I think that's so real. Especially it's funny you say the people who have built the models have kind of created that outsized.

You get what you asked, it's going to feast. And they have structured data to feast on.

Totally. So to go more inward into Veritone, I would love to talk about some of the core products you and the team have launched. I know a couple that stick out. Whether it's AIware, IDEM, IDMS or the even what you've done with CAA and the vault aspect of things. Curious to know, outside of those products that I've mentioned, which ones are. I want to say the Are you the most excited about which one has had the most reception? And for those that are listening, would love to have you give some context on these different types of products you and the team have launched internally.

I think it goes back to the original platform. Aiware was the original concept and thesis for the business. Having a single AI stack, platform agnostic, where I can deploy it on Azure, on aws, on Prem, which would in effect be able to do, I'd say four or five core pillars or service offerings, very highly scalable and dynamic data ingestion seems trivial. You're like, what do you mean gestion? And if I want to ingest live audio streams, I want to ingest harmonics off an electrical grid, I want to ingest live drone footage. Having a dynamic platform agnostic ingestion layer is still very critical and highly reliable. Number two is that data lake. But obviously now it's so germane to us that applying AI upon ingestion, right. Based upon jobs. So obviously some of our output of our needs is near real time. Some of our clients, they need the AI processing in near real time. I need that metadata, I need to know what's happening now versus others. It's kind of post analytics, right? So meaning the orchestration. So ingestion the data lake, the AI orchestration. Right. We've disclosed that, you know, we've used, you know, hundreds of models. Some are general models, some are hyper trained models, but it's the orchestration of those models cost effectively. And let's not forget, I mean we can do amazing things if we're. If I have only my entire job is one hour of audio, you could throw everything in the kitchen sink at that. But when you're processing millions of hours of audio and video, cost is real, right? Yeah, compute and GPU utilization is real. Ingestion, data lake, AI orchestration and then the workflows, how do I turn that into value? If it, is it a end application? Right. Is it a JSON data payload is an API, right? How do people consume the output of this? What I just described is AIware and what, what I'm most excited about is it's taken a long time, but we've proven that we could take the same stack and now apply it cost effectively to not just different applications or workflows in a single market opportunity. Like how do I help cnbc? But also now I've taken that same stack and I'm now servicing hundreds of state and local law enforcement agencies, the DOD and others. So what makes me most excited is our thesis. And I always remind people, like, let's go read our. Go read our S1. We took this company public very early, maybe too early, but if you kind of read our thesis on our S1 going to 2017, it's like spot on. Right. But it's taken us a while to get that yield. So, yes, I'm excited about the passion of empowering our law enforcement agency. Right, with our software. Yes, I'm excited about working with the Masters golf Tournament, ncaa, all these big brands. But what people need to understand is that what makes that all work effectively and cost effectively and at scale is AI work. So it's going back to the roots, which has been reinvigorating, very, very clear.

I love when you can take complex ideas and make them very understandable and digestible, which I think, Ryan, you've done very well throughout this conversation. So thank you for that and I'm sure the listeners appreciate it as well. One thing I would love to touch on in this conversation, something that interests me and I think a lot of listeners is just all things the podcast industry. I know that. I remember last year, I believe it was in October, Veritone had a divestiture of Veritone 1 and you all have been just in the podcasting landscape for a while. I'm curious to hear your general thesis on podcasting as an industry, as a vertical, the content associated and, you know, kind of the context of when and how you and the team were involved in this industry overall, given, you know, as a founder in this space, but as someone that's also very interested about your perspective here.

Absolutely. Well, I'll start with, you know, divesting. Selling off Veritone one was bittersweet. Right? Veritone one, I mean, I'm sorry, yeah. Veritone One, our podcast agency, which Veritone sold to your point, in October, was the. Was the substrate. It was the basis of what we applied AI were to first Audio. Right. We also did ingest, you know, tons of AM and FM streams from iHeart and others. But you know, I've been a big fan of podcasting. To me, it was a little bit of a glimpse of what media was going to be like, truly packetized, decentralized, on demand audio. And thankfully it still had tons of a lot of unstructured data. But it was like a bite sized problem where I could take one show, apply our AI to it and understand obviously, again, back Then it was, you know, it was challenging to make sure I could transcribe your voice, Casey, accurately now, right. I want to analyze everything your sentiment. Can I, you know, frankly, can I tell if you have a cold? I mean, there's so many amazing things we can do with AI now, right. Of analyzing today. But, but, but, but the podcasting agency that we created, we were, you know, we were on the agency side, but obviously we were able to interface with so many different upstarts in the podcasting space and some luminaries in the space, Norm Pattis, who obviously passed away and others. It's something very personal to me. We've been able to create a business about it and I think it still has great legs. One thing obviously, which you've embraced is it's gotta be multimodal, in my opinion, right. With the cost structures today so low, you need to be creating quote, unquote, podcast or media casts supporting every modality. Right. And be prepared to distribute it in every language. Because this technology is there, why not, right? If you can manage it, why not? And case in point, you and I are capturing this video. You're going to be repurposing this video. So podcast, you know, again, thankfully, they didn't call it purely audiocast, so I think podcast still hunts. But again, I think it's a phenomenal platform. It's cheap, but again, I think it's just the tip of the iceberg. And I do think that it works well with AI because as long as you are having the right structure, metadata around it, I think it's going to help maximize the new world of SEO, whatever that's going to look like for indexing of this with LLMs and whatnot. But again, I think there's just so much dynamic, exciting content that again, it's like many things, when you go from 20 channels or networks to 20 million, the economics is challenging, right. Of how you funnel those dollars in. But again, I think from a platform that we, as an audience, we love podcasting, I just think that like anything, you need to think about how the business model changes and how the platform, your end product, changes as the technology changes. Kind of what we've been talking about, that's a big theme here, is you have expertise in this space, you're entrepreneurial by default, your product is going to have to change with the technology advancement.

Yeah, very well said. And I think too, just even for the context for you, I mean, one of our core use cases is just the simple ability for as a podcaster or as a Podcast team that's managing dozens of shows to automatically associate all of the multichannel clips and pieces of content with the audio specific version for better measurement purposes. Because I've always been, since 2017, very video first with my show, very, you know, they call it omnichannel or vertically integrated for any kind of campaign. I do. And I think as someone that for context for you, ryan, like I'm 25 coming into this social media podcast world back in 2017, it was very odd to me of like, wait, people are just doing audio? Why wouldn't they do like embrace this social media era clip discoverability. And I think the use case that we've found a lot of early success with so far since we launched listener.com earlier this year is just the that ability for automatic association with all the cross platform content in this new, you know, 360 advertising deal era. So it's been very interesting to say the least. And I know, you know, we were talking about Zeus earlier. That's when I initially came across Veritone years ago. Was sort of like through that podcast lens and then understanding the larger, much larger umbrella that you and the team have, have created, it's been really cool.

To dig into one thing to that, you know, I think the podcast thing was, you know, we kind of saw early on this transition from, you know, a show to an influencer too. Right. And I think podcasting, you know, was, you know, definitively one, you know, in my mind, even before the explosion, you know, of Instagram and stuff like that, where most podcasts, I mean, it's dominated. You are the show primarily right from your side with guests. And so I think it's one of the best representations of taking an old model traditional radio and really bringing front and center. The influencer. Right. And not just the network, I like to say, is the network guy, Dan Patrick.

Right.

Who's sitting from one to many. You can speak one to many or you can speak one to one. And so I think the influencer ecosystem, podcasting is one of the best examples of influence of that as well.

Yeah, I love it. While we're on the topic something and as we kind of wrap up this conversation here shortly, what are some core podcasts or shows that you listen to that have been influential? I'm curious if there's any that stick out to you as these core reference points.

Oh, gosh, man, I'm going to get back to you on that one for sure.

Of course, no worries.

Yeah, you know, I mean, gosh, you stumbled. You completely blocked me like the ones like, sadly, I know the shows and I don't know all the names.

You're good.

You're embarrassing.

No worries at all.

Like, I can't pull them out right now.

You're good.

The one thing I can say is I'll tie it back to AI, my personal consumption. So we are just from structure, we are internally a Gemini shop. Not that I don't think OpenAI is incredible in ChatGPT 5, but just from a security compliance. The one thing is the amount of utilization I use personally for Gemini in both context. So for example, if I'm going to research your show, the first thing I will do is now go to Gemini to research your show. Right. And that's. And it's just, it's like having, I swear, my entire research and consulting team next to me the whole time. Now, again, do I trust the output and you know, completely. No. But again, obviously, as you know, most of the latest models have attribution. So I just say is, you know, again, I would encourage everybody like, you better be drinking right from the fire hose yourself. You cannot rely on anybody else in your organization to be the AI expert and start with the most obvious ones. Right. The big, large language models that are public. Right. They are amazing tools and I think it just, it really is good or bad. It's the tip of the iceberg.

Yeah, I mean that's. You sort of, you answered my question before I was, before I even asked it. I was going to ask more of just, you know, as someone that's building in AI, like, where have you seen the personal use cases? You know, from obviously the ChatGPT chatbot, but also are there any like, specific tools that maybe like. There's one just to give context. There's one that I love is. It's called Whisper Flow. It's this dictation app where it's, it's like the best dictation where any email that I write, I'll sit here, I'll talk for a minute and it's well structured, it's punctual, it's relatively perfect. And it's something where just a tool that I use personally that it's been game changing for me and I'm curious if there's any AI tools that you've used personally that have come on your radar that you're like, hey, if you're listening to this, I would recommend checking this out.

Well, again, I'm going to go back to the generic one, but in a secure format. So it's amazing, like just you know, two years ago, I'd be going to my CFO and my chief accounting officer in just obsession of like wanting to know subtleties and trends in our business from margin or stuff like that. Finally, once we've sort of secured and have our own very, you know, triple tap on the word secure is now I can load our entire financial model right. In our forecasting into our own tuned LLM. So now I can speak to my financial model. Right. And so I can know what my real time margin is and I can ask it in different ways. And once it's in a secure environment, then like everything I can compare and contrast. How is this comparing against my other public competitor in this space? Right.

Yeah.

So I think for them, my flip side of that is it's so easy for executives who have proprietary information to use. Just chatgpt.com Be careful, right? Because you want to take your entire data set and dump it in there as a macro prompt input. But I'm saying is for entrepreneurs, spend the money, please to spend a little bit more on the enterprise secure account for your business, but then have at it. I just say play grow data sets. Start with a PDF of some marketing material that you just had. Throw it in there again in the enterprise version. I don't care what flavor using Gemini, whatever, Claudia, I don't care, but just play around. I think it's great education for everybody and it's great yield.

Yeah. Very well said. Two more questions before we wrap up here. Ryan, looking back on your entrepreneurial journey, what's a core lesson that has been consistent through your journey that you'd want to tell young founders that are maybe starting businesses today in the midst of scaling their companies that you've learned over the years?

I would say three things. One, for at least me, have a partner. Right. I've had my brother primarily up until just a couple years ago, and he's obviously still a major shareholder and part of baritone. But I can't imagine having to do this all by yourself. Right. And it could be an engineer. It could be, you're the engineer. It's a business guy. It could be a mentor. Finding one person that can be in that foxhole with you, good or bad, and doing it, it's hard for me to imagine not having that. We mentioned Zeus. Zeus is absolutely. My foxhole has grown. But there's just pillars of people that you need. So anyway, I'd say that's something that's been consistent, good or bad. Second is, and it's a hard lesson and with Success, it's so is, you know, time and capital needs for entrepreneur is everything. There's so many, so many great ideas that just, you know, they spend too much or even they see, you know, a false indicator, like, oh, my gosh, we got this, and they start running about it and then they run out of cash, right? So I think just whatever your concept is, you know, try to say, can we do this divided by five in terms of budget, right? Just the best ideas, I like to say. Actually, my grandfather said, like, literally just described this to me, who, by the way, was like, built homes in theory. Almost everything we see and use is second class. And I'm like, what does that mean? He goes, what's like the probability that you have the best idea that was able to be produced? And we like to say, great. They call it the iPhone, right? Every once in a while there's a category. So my point is some of the best ideas have never seen the light of day, primarily because they so capital, capital, capital. And then third is you better be the leader in being, I'll say, just a maniacal optimist. You can call it tenacity, you can call it discipline. Don't give up perpetual optimism. You better believe in yourself that what you're going to do is you don't always have to do something that's gonna change the world, but you better believe that you're gonna realize success, like, right, and go after it. So, you know, again, you know, Colin Powell said, you know, optimism is a force multiplier. It's like one of my favorite quotes, right? And that's why I think that's the big thing for me at least having partners, whatever that partner is, be great understanding of your cash needs and be ultra conservative. Right? And third is you better be your. The biggest optimist in your company and your biggest cheerleader yourself.

Very well said. I appreciate you sharing those three points, Ryan. And last but not least, for everyone that has watched and listened until this point, where is the best place for everyone to learn more about Veritone, to stay in touch with you, Ryan, and just to get informed about everything that you have going on and to stay in your echo chamber of what you're building at Veritone?

For me, it's generic. It's obviously veritone.com, right? It's our website. It's got information. But. But, you know, we do. We're pretty active. By far, my most active channel is LinkedIn, so by all means, you know, I do, you know, try to post as much as I can. So by all means, you know, you can just follow me on LinkedIn as well. So those would be the two best vehicles for Veritone and for me as an individual.

Amazing. Well, thank you so much, Ryan, for taking the time to be on the show today. I know, you know, we could go on for much longer, but I learned so much and I really appreciate, again, like, the punctual taking complex ideas and problems and breaking them down to easy, understandable ways. So thank you so much for taking the time to come on the show today. I really appreciate it.

Thank you, Casey. Have a great day.