In this episode of The Next Next, host Jason Jacobs interviews Shaun Meredith, CTO and co-founder of Omnic AI. Omnic AI specializes in eSports player performance, leveraging artificial intelligence and machine learning through their platform, Omnic Forge. They discuss the inception of Omnic AI, the technology and tools they employ, and how their approach has evolved over time. Additionally, they touch on the different types of gamers Omnic serves, their business model, market traction, challenges in scaling, and the future of AI in gaming. The episode also covers personal anecdotes and insights from Shaun's entrepreneurial journey, as well as Jason’s own ideas about AI's application in traditional sports training for kids.
Elevating Gaming Performance: An Insight with Shaun Meredith, CTO of Omnic AI
In this episode of The Next Next, host Jason Jacobs talks with Shaun Meredith, CTO and co-founder of Omnic AI, a leading company in eSports player performance, data, and analytics. They explore how Omnic AI uses artificial intelligence and machine learning to help gamers improve their skills through their platform, Omnic Forge. The discussion covers the origins of Omnic AI, its technological infrastructure, the challenges and advantages of starting before the big LLMs, and the different needs of casual vs. professional gamers. Shaun shares insights into their business model, growth strategies, and the future possibilities with AI in gaming. They also discuss the potential for AI applications in other competitive sports and how emerging AI tools can aid entrepreneurs in building ambitious companies while maintaining flexibility and control.
00:00 Introduction to The Next Next
02:26 Meet Shaun Meredith and Omnic AI
03:34 The Journey of Omnic AI
06:27 AI in eSports: Challenges and Innovations
12:07 Building for Gamers: Insights and Strategies
21:01 User Engagement and Feedback
29:55 Technical Aspects and Future Directions
32:01 Mobile Gaming and AI Integration
32:35 Challenges in Computer Vision and Differentiation
34:13 Innovations in AI for Gaming
37:05 Funding and Growth Strategies
39:56 Market Dynamics and Competition
45:25 Future Prospects and Expansion
58:04 User Engagement and Awareness
59:33 Conclusion and Final Thoughts
Jason Jacobs: Today on The Next Next, our guest is Shaun Meredith, CTO, and co-founder of Omnic ai. Omnic AI is a global leader in eSports player performance, data and analytics. They use artificial intelligence and machine learning to help players improve through their flagship platform called Omnic Forge. I was excited for this one because as anyone who's been following my newsletter knows I've been thinking about.
Potentially doing similar, but in new sports, helping kids who are playing competitive sports to train at home using ai. So it was very interesting to talk to someone who is utilizing a lot of the same computer vision, machine learning, et cetera, but doing it to help gamers. Better their gameplay. We cover a lot in this episode, including how Omnic AI came to be, where they got started, what types of tooling and infrastructure they use, which was also interesting because they started [00:01:00] before the big LLMs were around.
So it was fascinating to hear how. How they did it then and how they would've done it now if they were starting from scratch. We talk about what types of gamers they serve and how they distinguish between the needs of the high-end professional players and everybody else. We talk about their go to market.
Their business model, the traction they've been getting so far, some of the challenges with scaling a business like this, and of course we talk about the future of not just of Omnic ai, but of AI and building technology companies with AI in general. It's a great one and I hope you enjoy, but before we get started.
I'm Jason Jacobs, and this is The Next Next. It's not really a show, it's more of a learning journey to explore how founders can build ambitious companies while being present for family and not compromising flexibility and control, [00:02:00] and also how emerging AI tools can assist with that. Each week we bring on guests who are at the tip of the spear on redefining how ambitious companies get built, and selfishly the goal is for this to help me better understand how to do that myself.
While bringing all of you along for the ride, not sure where this is gonna go, but it's gonna be fun.
Okay, Shaun Meredith, welcome to the show.
Shaun Meredith: Glad to be here. This is great. Thanks, Jason.
Jason Jacobs: I am excited to have you. Uh, so we, uh, connected through Seth Sivak, I believe. Uh, and, uh, you know, Seth and I have entrepreneurship in common, but one thing we do not have in common is, I don't know much at all about the gaming world, uh, but I, I do know about. Being a founder and, and about, uh, building consumer companies.
And I'm exploring doing stuff in [00:03:00] eSports around analytics and coaching and using ai. And Seth said, well, hey, uh, my friend Shaun is, is, uh, is doing that but for eSports. Um, and so it feels like I have a lot to learn from you and I'm really grateful that you made the time to come on and, and speak with me today.
Shaun Meredith: Oh, well thank, thanks for having me. Yeah, and and likewise, you know, you've got a deep, rich background in, you know, various entrepreneurial activities and the VC world, so I'm sure there's a lot we can learn from each other. So to do this. You bet.
Jason Jacobs: Great. Well, for starters, Shaun, maybe, uh, just talk a bit about Omnic AI and what it is, what it is, how it came about. Just I guess anything you wanna share about.
Shaun Meredith: You bet. Um, so, you know, basically put Omnic AI is, is, you know, what we're doing is we're analyzing video of your gameplay computer vision and deep learning techniques, and recognizing different elements or different, uh, [00:04:00] parts of your gameplay. Uh, to help you play better. So kind of think of it as like an AI coach, assistant, uh, kind of thing.
And we do that, um, in a simple, easy to read, uh, post match analysis that you get after every match. Uh, we currently cover five titles, uh, Valant, Fortnite, Overwatch two, rocket League, and Madden, NFL. And we plan on expanding to many more titles within the next year. Uh. In terms of kind of how we got started or into this
Jason Jacobs: Yeah. 'cause you, you've been doing it a while, right?
Shaun Meredith: uh, yeah, we have since, uh, 2021. you know, basically, uh, I've been, I've been an entrepreneur, uh, uh, several times. This is my fifth startup. Uh, worked for Apple as well. So always interested in that technology, you know, angle. had been watching the gaming industry for, for a [00:05:00] while, saying, you know, you know, like what is something that's kind of unique that we could do kind of in the game
Jason Jacobs: Are, are you, are you a gamer, Shaun?
Shaun Meredith: Um, you know, that's a double-edged sword, right? People always ask, are you a gamer? I'm not a hardcore gamer. Um, I'm definitely like a casual gamer, would love to game more, but I've, I'm always busy building things, so I wind up building things, uh, you know, software much more than I do actually getting to game myself. However, that being said, you know, I really enjoy puzzles, strategy games. And, you know, uh, uh, basically, you know, hanging out with my friends online just like anybody else.
Jason Jacobs: Great. Well keep going with the story. Yeah, I asked just because it's always interesting to understand, uh, what, what, when founders are building, you know, it's like some people say build for yourself, and other people say, well actually it doesn't matter. And in some ways it can be a hindrance to build for yourself because if you built for yourself, you're not listening to your customers.
So I was just curious.[00:06:00]
Shaun Meredith: No, um, that's a, that's a really good point. And I do think, like, you know, there's part of it that is kind of building for myself, right? In a way, because I'm always interested in using data to improve, right? Uh, you know, something kind of burned into you as a technologist is, is, you know, okay, let's take data now.
Let's see, you know, does, is that a good measure or not? How do we improve? How do we, you know, uh. to the process Um, so what was really interesting is watching, uh, kind of some of the eSports titles on the rise and seeing how, uh, suddenly, you know, uh, some of these tournaments, uh, began getting bigger and bigger, the leagues formed, and then you started having franchises in leagues and you could easily see how just like traditional sports and, and, you know, major League baseball took a hundred years to get to Moneyball. Um, some of the eSports, uh, [00:07:00] industry, they were only gonna have like three or four years to get to Moneyball, and so how could we, you know, help or enable that? And, um, also, uh, I went to school at MIT for both undergraduate and graduate work. and uh, for my undergrad and master's work, I was, uh, doing, stuff with uh, plasma physics fusion. And we used a lot of computer vision techniques to look at the plasma to try to predict what was going to happen with it. And that's when I had my aha moment and called up my co-founder and said, Hey, um. I have this crazy idea, what if we take these same computer vision techniques that I used to use for looking at fusion plasmas. But now we watch video of people's gameplay and try to predict what would happen and how they could get better. And immediately he was like, oh, I got it. Uh, my [00:08:00] son just went to UMass Amherst on a League of Legends scholarship. So
Jason Jacobs: And, and this was, this was Chuck, right?
Shaun Meredith: yes,
Jason Jacobs: Yeah, so I know Chuck from, 'cause I mean, Runkeeper started back in 2008, which was right around the apperian, uh, birthing time. So, so I, Chuck and I were, there was like three or four or five, uh, mobile app entrepreneurs running around Boston that were always on the same panels and stuff like that.
And Chuck was one of 'em. So, yeah.
Shaun Meredith: Anyway, so, so he's like, yeah, I'm in, I, I, I see where this could, you know, be big. um, so that's when we kind of started, we got into Techstars. Uh, built a prototype, worked with a couple of, uh, uh, professional teams and some coaches to try to say, can this even be done? And then once we had that up and running, kind of going, yes, look, it actually does work.
It can, it can work. Then we began the, the, uh, process of moving this into something that, you know, [00:09:00] player can use instead of just. A consulting model where, where, um, you know, professional teams or top tier players could use it.
Jason Jacobs: So what year were you saying can this work?
Shaun Meredith: Um, that's it. So that was in 2021?
Jason Jacobs: Got it. And when you say, can this work, what was the this at that time? Because uh, I mean there's a number of things that I would imagine have to work, and also there's the whole for whom, uh, um, proposition as well. So I guess, what were the most important things for you to prove out in the earliest days, uh, before you even anchored that that gave you the conviction to anchor?
Shaun Meredith: Well, that's it. So if, if you remember back in 2021, um, that was before all these other AI tools kind of existed or, or were even out there. That, or that there was an explosion with LLMs, stuff like that. So, so it was first, you know, can we train a model to. To, you know, watch video of this person's gameplay and then [00:10:00] actually try to try to, you know, infer from that, you know, what are things that the person can improve on. Um, and we figured, you know, at the time there had to be some way to kind of do it because you were seeing, you know, companies like DeepMind and, and Alpha Go and Right being able to play yo or chess or even StarCraft They were training these ais to play the game themselves. So, so obviously there was a way that if a game could have some sort of, or if an AI could have some sort of understanding of the rules of the game, there's gotta be some way that it could also have an understanding of, oh, how's the user playing now?
You know, how do we apply that and, and also keep it fun, right? We didn't want to, we make sure that our AI doesn't. Doesn't steer everyone towards the same exact kind of gameplay. Right. Um, [00:11:00] you saw that with Alpha Go where many of the, uh, humans that Alpha go played against. They said, you know, made moves that a human wouldn't make. just things that we wouldn't think to try and, and we didn't want to get into this. Oh, well, it's going to make everybody play exactly the same.
Jason Jacobs: Huh. For, for whatever reason, the analogy comes to mind of how like ways ruin certain neighborhood streets because all of a sudden it's like there's a cut through that never used to be a cut through, but ways, ways sends everybody there. And so the neighborhood goes from being a quiet street where kids can ride bikes and stuff to like being, being rush hour.
Yeah.
Shaun Meredith: exactly. And, and you know, let's face it, it's, you know, um, it's kind of like everybody playing League of Legends always did the same build order every time. Then well, where's the fun of the game of going? Well wait, it's 10 minutes and 18 seconds. I need to purchase this item. You know, oh, you know, I need to do this. There isn't [00:12:00] any, any kind of skill involved anymore, so.
Jason Jacobs: Okay. And, and so, uh. Uh, how did you go about training the model at that time? And similarly, if you were to start from scratch today, given that the LLMs do now exist, how would you do it different?
Shaun Meredith: Yeah. So, so there's several, um, that, that's a really good question. I think there's several different points to it. First is, uh, the way we went about it then was, you know, simply to say, okay, uh, the first steps are, you know, that object detection and recognition, and let's see if, if we can. Just get the basic objects that are on the screen.
Right? So, so that was kind of step one. And now that we have, you know, step one in place, and we know that we can recognize all the different heroes that are in the arena and where they're positioned and, what they're, you know, uh, capabilities are, [00:13:00] and you know, where they stand in terms of. How much ammo they have, or how charged is their ultimate ability now that we kind of have a state of the game, is there a way that we can start to infer from that state of the game what are better actions to take as, as an individual player or as a, as a team of players? And so with that, we relied on talking to, you know, several coaches and top tier players and saying, Hey, in this situation, you know, what would you do? So we were able to start to train the AI to look for certain key situations and then essentially recommend, here's in that key situation, you as this type of player, this is your best bet to try to, to try to improve. And then we match that with outcomes. So we were able to take video of several different, you know, hundreds [00:14:00] of thousands of matches and see. If they
Jason Jacobs: And when you say take, I, I mean, does that mean film or find
Shaun Meredith: uh, that's it. In working with these teams, we were able to get a lot of their scrimmage video.
Jason Jacobs: Hmm.
Shaun Meredith: Um, in addition to
Jason Jacobs: Were they, and they were, they were recording that already. So they had these data sets that were just sitting there not being utilized.
Shaun Meredith: Um, that's it. They record them and then they do video review afterwards. Right. And, and so same thing. That's a strength that our product wind up having is. You can think of, here's a professional coach. They maybe, uh, you know, do run scrimmages with their team of five players for say, four hours a day. Right now, from those five different players' point of view, the coach has 20 hours of video to go through that day just before the next day starts. And so, uh, rapidly you could see how like our tool became useful because if it could. [00:15:00] data or, or say here's a certain key moment to review. They could, they could automatically reduce that coach's, uh, video review time, kinds of things. Um, we were able to take all that and then say, okay, did that increase their probability of winning or not? and then it became, you know, simply looking at that and going, yes, if you do this action, you are. You know, chances for your team winning will improve. so, you know, mathematically we're able to start, you know, compounding a bunch of those different results and, and coming up with the best practices for that type of player or that type of gameplay.
Jason Jacobs: So what, when you were, uh, initially getting going, is there some type of magic number in terms of when you have enough data? To launch and then [00:16:00] switch to getting the data from real people using your product versus just from YouTube or data sets that exist elsewhere. Uh, like what's a, what's the minimum where it's okay to flip the switch and actually get some sample size?
That is a reasonable starting point if there is such a thing. Yeah.
Shaun Meredith: um, so, so what we found is, is that what we like to do is train, you know, it, it's, this is a rough number because it does vary per game. It does vary, uh, uh, a little bit by the, the quality or the variation that we have in the video, but we have been able to find that that initial seed we can do in about a thousand hours of video, which is much lower than I thought it would be. I thought it would be much higher. Um, but that is enough to kind of get it started
Jason Jacobs: What happens if it's below that, Shaun?
Shaun Meredith: Um, if it's below that, we start getting some really [00:17:00] bizarre, uh, recommendations or, or, um, right. What people in the industry right are calling is, you know, hallucinations. Um, I don't really like calling them hallucinations.
I think they're just flat out errors and um, you know, we want to avoid those. Yeah.
Jason Jacobs: Got it. Um, okay. And so when you were getting this initially going, uh. Who did you think you were building for? Was it the coach? Was it the player? Was it a coach replacement? Was it, was it to assist the coach? And also, how did you know, and how much time did you spend with the potential stakeholders that you thought you were building for while you were building, before you were building?
Like, like what, what order did you, did you kind of tackle this chicken and egg? Yeah.
Shaun Meredith: When we first started, we thought, you know, our, our initial market would probably be, you know, higher end
Jason Jacobs: Mm-hmm.
Shaun Meredith: Right. but you know, as, as you know, in that entrepreneurial journey, you, you quickly [00:18:00] start to figure out some things and, and you know, as as we started kind of working with those, you know, higher end teams, we learned that, you know, essentially that market size is much smaller than we'd like. So even though that became, you know, where I would say we could get a lot of training information and a lot of value for it, our, we wound up actually looking at this and going, no, our market is your everyday gamer. Right? So, um, an everyday gamer right now will. spend seven to $10 a month on trying to get better, either by buying game guides or augments for their game, or, you know, you know, actually they even just buy, you know, cosmetics for the game. And, and that's where we wound up saying, look, this is where it gets really [00:19:00] interesting and we can actually have a bigger effect. Right. So, so may, you know, I may as a player like playing. Um, you know, destiny as, or Destiny Two as an example, but I want to actually play, know, of Duty better, or I want to play Fortnite better, or I wanna learn how to play Valant, and that's where here's this tool that can help me get up to speed in an existing game where other people are far better than I am already. Um, that became very appealing, I think, to, to many of our users. And it was spending time with those types of, of people saying, I, I, I still remember way back in early, uh, 2022, um, where we had just some regular people come in and play a few matches of Overwatch too, and. [00:20:00] Having one of them come up to me and say, you know, I'd never played this game before in my life, and this is probably completely obvious to everybody else, but I didn't even know about it. know, the, the Omnic Forge actually came up. One of the recommendations they got was, look, you're playing a support character that has healing capabilities and your HUD was turned off. that you couldn't see the other players on your team and who needed healing, and he's like, you know, that's sounds completely stupid, but I'm glad it actually told me I need to turn that HUD on. makes, that actually made me a better player. 'cause I could start to see who needed healing and run around and try to provide that healing. And he's like, and I know a top tier player probably doesn't need that kind of recommendation, but I really appreciate that. This already told me something that I could do as a beginner.
Jason Jacobs: So [00:21:00] what is the workflow? Uh, if, if I'm playing the game, uh, h how, how do I engage with Omnic and when do I engage with Omnic relative to when I'm actually playing the game?
Shaun Meredith: Yep. Um, so we have a couple of different ways that you can, you can engage. Um, the first is you simply go to forge.Omnic.ai and you sign up, you can get a free account. Uh, a free account will give you a. certain level of analysis and you'll be able to see your last five matches and a couple of insights per match.
Jason Jacobs: And, and when you say, see, is that, um, how does that, how does my gameplay get populated?
Shaun Meredith: yep. And so, so what happens is when you sign up, you have a couple of different, you know, options of how you want to sign up. You can just give a username and password, or you can connect your Twitch or your YouTube gaming account.
Jason Jacobs: Mm-hmm.
Shaun Meredith: And if you connect your Twitch or YouTube gaming account, what happens is, is you're giving, uh, the Omnic [00:22:00] Forge permission to watch your stream. And that's, that's all you need to do. And literally, the Omnic Forge watches your stream, every time it recognizes that you played a match of Fortnite or a match of Overwatch two or a match of valant, um, after the match, it will send you a post match report.
Jason Jacobs: Mm-hmm. Almost like, almost like how granola does that for a Zoom call.
Shaun Meredith: yep. And that's it. So you'll get a post-match report and you'll see like a little email come in with, with, you know, your player card that has a picture of you, some of your key stats, your hero or agent on it, and what map you played on. And then, you know, uh, maybe you also might get something that's like a, plays like that says. Hey, congratulations. We recognize in your gameplay you were playing like tens or like booga or like faker, that that can become a [00:23:00] really big, you know, driver. Um, uh, of kind of that oh wow moment. Like that's really cool. Um. Then if you, you know, uh, click on the link, it can take you to a webpage where you can see all sorts of stats, your career stats, your, uh, percentage wins by map or by agent. Your, um, what, uh, tool that we really, really, uh, like to tout is our aim analysis tool that no one else has. Um, where it will show you your time to target and whether you targeted, you know, for critical hits or for body shots. So on and so forth, as well as, as those insights, right? So those quick little, uh, tips that say, you know, for your next match, focus on this or, you know, try, try using this weapon instead of this one. Uh, you'll be better, uh, with it, you know, based on your gameplay. And then as well, there's a place where you can get a [00:24:00] summary of your match and actually then engage with the forge by asking it questions saying, okay. Uh. Where should I have placed my smokes more effectively? Uh, for instance, or, you know, um, is there any synergies I can have, you know, as I'm playing Fara, with some of the other, uh, agents or heroes? Yep.
Jason Jacobs: Uh huh And, and this kind of guidance, uh, how much of that was human or human in the loop, uh, and, and as you went and trained the AI to do it? Um. Was that sitting with a human who to instruct or, or was it the AI running a bunch of tests in the background and kind of doing simulated gameplay? Or I guess kind of practically speaking, what is the best way to train the model and how much should the human be involved as you scale, if at all?
Shaun Meredith: Yeah. Um, so, you know, those are actually great questions and things we're still learning, [00:25:00] right?
Jason Jacobs: Hmm.
Shaun Meredith: still figuring out some of those things. Uh, so far what we've been able to figure out is, is, uh, like you said, we, we sit down at the, at the beginning and we use a human coach or, uh, sets of players, and we come up and. We call it seeding and we'll seed the AI with, with, uh, some various things for them to look at. And, and those can be as simple as, you know, uh, don't engage on this point from the southern side. You want to come in from the northern side because of these reasons. Or they could be things like, look, if you have this player and this player and this agent on your team. is the type of coordination you should be using in order to, you know, take that site. with those, we will literally probably have anywhere from a hundred to 250 of those that we seed in. But from [00:26:00] there, the AI does begin as we. As we develop more matches, the AI does begin to pull in more things and, and recognize, oh, here's other things that I think are creating wind conditions. uh, so we will see those numbers of insights grow greatly, um, over time. And that's the best thing that I tell people about with any AI tool, right, is the worst performance you're gonna have is what you see today. And it's only gonna get better as it learns and gets more information from the community. And that's exactly what we're seeing. It gets better and better and better.
Jason Jacobs: Hmm. And when it comes to avoiding the misses or hallucinations or errors or whatever you. Want to call them? Uh, I mean, is there, is there human oversight today to, to catch those before the analysis goes out the door or, um, at, uh, what, how does that process work?
Shaun Meredith: there, there isn't, uh, we do get those that do come, come out. Uh, [00:27:00] however, what we're using is the community can
Jason Jacobs: Mm-hmm.
Shaun Meredith: So for every insight that, that the forge, uh, delivers. Uh, you, as you as a user, you can say, yeah, thumbs up or thumbs down. I want more of these. I want less of this. And, and as you do that, you're actually giving, you're actually giving feedback into our AI engine.
So our AI engine, we can say, yeah, look, this kind of information that came through a hallucination because we just received, you know, I don't know, you know. 50, you know, down votes in the last hour. So, um, automatically the, the AI will start tuning away from that and not providing that information to you.
Jason Jacobs: Hmm, and, and this is less of an eSports question, but it's more, I guess it's just kind of product management 1 0 1, but, but if you have these hardcore gamers that, uh, it's a smaller market that you're explicitly not focused on the pros, um, and I would imagine that they're [00:28:00] far more demanding with needs that aren't necessarily, that don't necessarily align.
With this more casual market that you're focused on, and then you have the more casual market that has a different set of needs. So it's not just about the feedback, it's about the from who. Right. And so how do you think about what to listen to and what to ignore, uh, in the, um, as it relates to who you're building for?
Shaun Meredith: Yeah. Um, actually that's a really good question. So,
Jason Jacobs: Well, I learned with Runkeeper, right? Because the, the needs of the ultra runner are very different than the needs of the couch to 5K person, right? So, yeah.
Shaun Meredith: the way that we're going about that, and we'll see it's an experiment, but it seems to be working is, is that, um, like, uh. The way that we built a, a coaching tool that can scale, right? Uh, using human coaches doesn't scale, but this AI coach can scale, uh, to cover users. Um, we're doing the same thing where, uh, we've got an AI that actually categorizes or [00:29:00] classifies or, or we call it thumb printing. we're thumb printing your play style and your, your, your game style as, as you are using the forge. And then we're able to drop you into buckets and, and that way when you provide the feedback, it's for that bucket
Jason Jacobs: Hmm.
Shaun Meredith: Um, we need to do a lot more improvement on that, but it seems to be, you know, working and will continue to refine and have different buckets and different categories of those users. And, um, with that we hope that, yeah, essentially it will, it will provide this bucket of users more of a certain type of recommendation. And those high end users, they go, yeah, I don't want any of that. You know, I want this kind of stuff. And, uh, the AI will provide that information to those higher ed users. Yep.
Jason Jacobs: Got. Got it. And uh, one of the things you, you mentioned before we started recording that you listened to the episode I [00:30:00] did was swap mill from. Swing vision. He talked about how they started going like you did before these LLMs existed, and therefore they built their own models. And one of the, the, he didn't use this word, but it sounded like accidental advantage of that, uh, was that they could do all the compute on mobile and not in the cloud, which meant it was a lot quicker and a better user experience.
Um, how are you doing things? Where does that compute take place? And in eSports, does it matter?
Shaun Meredith: Um, actually it, it does matter. Um, so as an example, um, we've, when we first started, we, we started thinking about this and, you know, look, do we do the compute at the user's device? Um, because it, it's kind of interesting, right? Most of these gamers, at least on the PC side of things. Have a big GPU sitting in their, you know, device.
But that's it. We rapidly found out that, uh, gamers hated that because, you know, having that compute on their device meant that their game [00:31:00] didn't play, wasn't playing as smooth. we actually do our compute, uh, uh, basically in, in a mixed mode of, of, uh, the cloud and actually a data center that we have.
Jason Jacobs: So you've got the opposite, that actually it's a disadvantage. So what's an advantage for swing Vision is a disadvantage for Omnic.
Shaun Meredith: that, that is absolutely the
Jason Jacobs: Huh. I'm so glad I asked because that, I mean, that's gonna be super telling for anyone listening that's trying to evaluate how they should do it, right? Is that, well, the answer is like so many things.
It depends. Yeah.
Shaun Meredith: Yep. Um,
Jason Jacobs: I,
Shaun Meredith: like that is, you know, one thing that is a strength for us is since we're only analyzing, you know, video of your gameplay. Uh, it turns out we don't really care what what are playing on.
Jason Jacobs: mm-hmm.
Shaun Meredith: suddenly we can analyze games from consoles like, you know, PlayStation or Xbox or Switch. they just have to share video of their gameplay and [00:32:00] We can help them get better. Uh, likewise, uh, you know, mobile, um, that'll become bigger as we cover some other games here. But it doesn't matter that, you know, you can just take your phone and record video and send it to us. And once again, we can analyze that and it doesn't require a lot of retraining on our end. we have to do is say, yeah, we have this video, take it and. our, uh, engine says, okay, sure. Got it.
Jason Jacobs: Uh, so one question I have is that it, it, it seems like AI is on this very rapid trajectory and there's a lot of leverage that's being gained with every passing day and week and month as the innovation is happening under our feet. Uh. What's happening in the computer vision world? The sense I've gotten from other guests [00:33:00] that have come on the show is that the, uh, it's a more linear path where you're not gonna get as much leverage and maybe it's even harder to, like, maybe, I don't wanna use the word commoditize, but it, it's just harder to differentiate from others who are doing similar.
So I guess what are you seeing as it relates to the stack that you're building with and, and how quickly it's evolving and. I guess how do you keep up, but also how do you differentiate at the same time?
Shaun Meredith: Yeah. Um, the differentiation is, is an issue, right? Um, so, so, uh, I wouldn't say that we've gotten that figured out yet. Uh, you know, what we try to do is, you know, market our product and explain to people, here's, here's, you know. You know, kind of the basics, right? Of yeah, this is how it works. This is, you know, what it's doing.
It's, you know, watching video of your gameplay. Um, it does take a little bit of time to process, but if you stream to us directly, [00:34:00] can start processing as soon as you start playing. And that way, you know, right at the end of your match, within 30 seconds or a minute, you can get a report. so, so those kinds of things. Um. In addition to that, I think just like any other technology, it's just important to kind of know the lay of the land and, and keep flexible, right? So, so that's what we've done where, um, uh, been around with some different tests for our insights engine. I hope that we can release some, uh, soon that, that make modifications and start to build on some of these other workflows or these other tools, um, so that we can essentially take the data that we get from the computer vision side and move that over into somewhat of an LLM side, right?
So that now you're, you're able to, uh, [00:35:00] you know, better interact or, or see that. Or ask questions about, you know, my progression in something maybe that we didn't even think of. Right? So, so, you know, you might be able to say, oh, well look, right now I play a lot of, uh, I play a lot of dualist in, in Valant, but I'd like to become more of an initiator. How do I do that? And, and have it be able to provide you that guidance and say, I. this is how you know this. These are some techniques to adjust your skills in this area and you know, basically port them over into that different type of play style.
Jason Jacobs: Ha have you, or has anyone experimented with analyzing how the pros play and finding a way to either package and contrast their style with the learners, or even put them in touch as a way for them to make extra income cameo style? [00:36:00] Mm-hmm.
Shaun Meredith: some elements of a play style that a particular, a high-end player. Did. And it will, it will say that, it will let you know that, Hey, we notice these elements of, of your play in this game. If you want to learn more, you know, go here, you can, you know, view their Twitch stream or, um, if they're a, as an example, we just signed a partnership with M 80 and, uh, we will be working on like some rollouts of M 80 with their, uh, academy. So it might make recommendations to you of, you know, oh well look, you can go join, you know, the MA Academy MAD Academy [00:37:00] course, uh, with Floyd or something like
Jason Jacobs: Mm-hmm.
Shaun Meredith: Yep.
Jason Jacobs: Got it. Um, how, how is Omnic AI funded? Uh, I, I guess how much have you raised and from what sources? And don't share anything you don't wanna share. I, I have a follow up question that's relevant, which is why I wanted to know that for, yeah.
Shaun Meredith: very fair. Um, yeah, so we came out of Techstars, uh, uh, coming outta Techstars very quickly. Um, we were able to raise about 730
Jason Jacobs: Mm-hmm.
Shaun Meredith: uh, from, uh, various angels, um, including, uh, uh, uh, rich Minor, uh, who people would recognize as, um, the founder and CTO of Android. And so that was a big endorsement for us that okay, we have something here. Right. Um, from there, uh, we have, uh, done essentially, a pre-seed and we've raised sedate [00:38:00] about 2.2 million. And, um, uh, right now we are
Jason Jacobs: and, and was that from individuals as well or were there, were there funds involved?
Shaun Meredith: and some smaller VCs.
Jason Jacobs: huh
Shaun Meredith: Yep.
Jason Jacobs: Uhhuh and, and, and, and how are you thinking about capitalizing the company directionally?
Shaun Meredith: Well, that's it. So then, um, directionally what we're doing right now is, is we're really focused on growing our user base, uh, right now. Uh, uh. Essentially, um, looking at, you know, ways that we can, you know, generate, uh, additional revenue and, and increase our conversion rate to our paid plans, so on and so forth. And we're doing that a lot through some different partnerships that we've been announcing, such as, uh, the M 80 partnership that we just did, uh, play versus, uh, who works with, uh, something like 5,000, uh, uh, high schools across North America. [00:39:00] Uh, we just did a partnership with Levels Gaming, uh, in Saudi Arabia and Great Britain eSports in the uk. um, we've got a few more coming down the line. In fact, I think this week we start with, uh, combat tested gaming and, uh, the VFW, for veterans the Game. And they're running a big Rocket League tournament. And, um, we're providing our software to, uh, for use during that tournament. Um, and then with that, uh, we plan on basically at the end of the year, uh, kind of raising a, an a round to really scale and, um, you know, essentially up our sales and marketing activities as well as, uh, expand the tech team so that we can hit all those top, uh, game titles that I mentioned before.
Jason Jacobs: Uh, you hear these terms like AI native and seed [00:40:00] strapping is another like buzzword of the day where like you raise a seed round and because these AI tools are enabling teams to build so much more efficiently, they can get to profitability and never look back. Um, uh, I guess what does the term AI native mean to you?
And, and. How do you think about yourselves in that regard? And also, um, it, I mean, it sounds like you'll need a bunch more capital to scale. How do you think about that seed strapping and what types of companies are a good fit versus not? And, and, and why This might be more capital intensive than, than some of those companies that at least, uh, the perception, um, is that they're becoming more capital efficient.
Whether that ends up being true or not, we'll see.
Shaun Meredith: Yeah. Um, honestly, uh, this is an interesting one to me because, um, I think, I think, um, I. You know, right in that journey. Right. So when I, when we, we first started the company, my thought was, okay, this will be, uh, you know, more capital intensive, [00:41:00] just because the size of the market is so large with, with, you know, 3.3 billion gamers out there in the world. 1.8 billion of which put themselves in a competitive or highly competitive category. That just seems like, okay, that market is large enough that that venture, the venture route is, you know, what I would say is kind of the traditional, you know, route to go right to, uh, to be able to meet those, the uh, needs. Um, however, then we kind of went through a period where you, you know, like you said, that seed strapping. You know, seemed like, wait, this might just be fine. And, and, you know, we might be able to, you know, essentially, you know, grow the company with the revenue that we have coming in and, and not hit that. now the way I see it is this is, this is where you get into that interesting. kind of inflection point of, uh, we're getting a [00:42:00] lot of traction. however, we're also starting to see some other people that are starting to add AI tools. We know they can do that quickly. we know
Jason Jacobs: of the larger existing players you mean? Mm-hmm.
Shaun Meredith: Yep. And, and we know that they can do it quickly, however, they'll have a lot of learning that they need to still go through, right?
Like there are just things that. Inherently as we developed our, our software that we were like, oh, uh, that's kind of interesting. It works in this case. Suddenly it doesn't work over here at all, and why didn't it? Right? Um, so we know they'll hit those. However, um, this is where if you can stay ahead of the game or, or be, you know, in that kind of first mover advantage or that bigger player advantage. That might be, uh, know, of, of, of significant importance to, [00:43:00] you know, the company staying, at that forefront or ahead in all those gamers' minds.
Jason Jacobs: Uh huh. So how, how big is big as it relates to these potential competitors and also where do they sit in the market? What do they do?
Shaun Meredith: well, that's it. So, um, you know, uh. You know, here, here's, here's a great example. Um, I think, uh, during, um, uh, game Developers conference, uh, Microsoft kind of pre announced that they would have like an AI gaming coach.
Jason Jacobs: Hmm.
Shaun Meredith: Right? Um, they, you know, it turns out what, you know, what they announced
Jason Jacobs: So that, so when you say how big is big like that, that answers my question. That's pretty big. Yeah.
Shaun Meredith: and, and it turns
Jason Jacobs: It's not like an a round company that might do an acquihire or something. It's like, no, this is Microsoft. Which, you know, yeah. Good news and bad news. It's good news that they care about the category and if they come in and try and fail, then they're gonna be a lot more, you know, desperate and price insensitive in terms of fixing their problem.
But, [00:44:00] uh, but they may kill you first. Right. So that's the, uh, that's the dice roll. Yeah.
Shaun Meredith: does show market validation, right. That okay, wait, there's a market there, so, so there's that. And then also there's just, um. There's also, um, the, the standpoint of which this sounds small and petty or, you know, small and, you know, strange. But the fact is AI is bigger now, or, or on, you know, is around. So access to just things like graphics cards or the cloud compute. right? Like when we need compute power, it's expensive. And, and I didn't see that becoming more expensive over time. I thought it would fall in price, but at least now it, it's, it's rising
Jason Jacobs: Be because of demand.
Shaun Meredith: term it will still [00:45:00] fall,
Jason Jacobs: Yeah.
Shaun Meredith: right now it's
Jason Jacobs: Yeah. Be, be because of the demand that ai, the influx of AI is, is causing. Mm-hmm.
Shaun Meredith: that's it. You know, like I mentioned, you know, we have several servers kind of on-prem and stuff like that, getting GPUs for those servers because the big, big people out there are basically buying all the inventory.
Jason Jacobs: Hmm.
Shaun Meredith: a really interesting problem.
Jason Jacobs: Now a company at your stage, how do you guys think about engaging with these big potential competitors slash potential investors or partners or acquirers? Is it, is it arm's length? Do they know who you are? Do you recommend approaching and spending time with him? Do you, do you kind of keep 'em warm? I know what I would do if I know what my answer would be, but I'm curious what your answer is.
Yeah, yeah.
Shaun Meredith: yeah, I think it's a mix of those. It's, it's, it's kind of interesting, right? There's, there's some that you go, okay, let's keep them at arm's length. There's some, you say, Hey, let's, let's start talking and engaging, you know, um, maybe, you know, [00:46:00] there's, you know, uh, I don't know. There's always that typical business strategy of, well, let's approach the number two player. See, you know how that
Jason Jacobs: Mm-hmm.
Shaun Meredith: Um, because, you know, if there's people out there that have announced something that isn't really there yet, that, you know, same thing. If you walk in and say, Hey, we have something that, that you can start with today, then suddenly they get the first mover. So.
Jason Jacobs: Mm-hmm. Now, we've talked about competition and differentiation being a challenge, but set that aside. If you just look at the overall market. As you think about market size, as you think about propensity to pay, as you think about churn as you, as you think about, uh, um, price sensitivity, uh, I'm, I'm probably missing one or two things to, to think about, but just what's the hard part about, about it?
If or, or what are the hard parts?
Shaun Meredith: I think what you mentioned, you know, is exactly the hard part, [00:47:00] especially for me coming from
Jason Jacobs: Which, which part, which part of that.
Shaun Meredith: I. Mindset is, is kind of all of that, like
Jason Jacobs: Huh.
Shaun Meredith: all together and
Jason Jacobs: Hmm.
Shaun Meredith: oh, like, you know, for instance, um, price sensitivity. We did a few tests with that and we didn't see a lot of price sensitivity. However, when, you know, we go out there and talk to our end users, they all mentioned price being an issue. So, so Right. It's trying to figure out, okay, uh, maybe then. Right now, we kind of have, uh, we have a free model or a freemium model with a paid for, you know, 9 99 a month. You can analyze any of your games, many users are only playing a game at a time. maybe there's a way to say, look, if you pay, you know, 4 99, you get one game. Know what I mean? And then each game you add is another two or $3, but you can still, you know. [00:48:00] go ahead and just pay the 10 99 or, you know, 9 99 a month to get, uh, access to all the games that we cover.
Jason Jacobs: Hmm. Yeah, I mean, I've been thinking about some similar stuff in other categories. So think like, um, I mean, uh. My, I mean, my son's a hockey player, so one of the things I've been thinking about is, uh, is that if there's a live barn has these video cameras in, uh, a bunch of rinks, you can get the footage.
There are human coaches that'll go and review the shifts and charge you per game to analyze how your kid did and sit down with your kid and go through like, you know, I like, oh, I like that decision there. Like, one thing you also could have done is this thing. And like, here's some things you did well here, here's some things you could do better next time.
Stuff like that. Right? And so it's like, well, is there a way to train a model to, um, to, to do that? It, it seems like. You know, VC's mindset, not that I would go raise venture anyways. I don't, I, I hope not to with whatever I build, but, but their mindset is that like, oh, for the reasons you mentioned, like, oh, like small market, hard, low [00:49:00] propensity to pay churn, uh, pr, price sensitive.
It's like, uh, I don't know if that's venture right. Um, like. I mean, I, I, I get that you're already on the train, uh, to, to, to some degree. But, um, what would you say to, to a VC that had that feedback for you, um, sitting where you are today with the last several years under your belt that you have.
Shaun Meredith: Um, I, I
Jason Jacobs: Yeah.
Shaun Meredith: you know, kind of the best approach or, or, or the best thing that we can kind of say is look, um, those are all problems that have been solved with other businesses. Um, they're all tweaks to, you know, your product offering. Um, those are all optimizations that need to come into play.
Jason Jacobs: Mm-hmm.
Shaun Meredith: and there's ways to, you know, basically. You know, measure them, do a test measure, and then make a change. Right. And, um, you know, building a company requires doing that many times, multiple times over. [00:50:00] And, you know, I think I, I share your, you know, thoughts of, you know, if we can do, if, if we can build the company. Without having to take that additional capital, I'd love to do that.
Right? Like I, you know, revenue is, is the best way to, to, uh, you know, kind of feed the, the company. Um, and it may be possible still, um, but. If we do need to go for that capital, I think, you know, all those arguments are essentially things that, that everybody encounters and that's why, you know, there's entrepreneurs to try to figure that stuff out and, and guide the company and make it successful
Jason Jacobs: And if, if, if, like, I mean, you happen to be passionate about gaming. I happen to be passionate about [00:51:00] hockey and, and other things too. But, but I was talking about hockey specifically here. If you take out passion and you just look as a pure capitalist, I. At these markets and you think, okay, computer vision, AI for analysis and coaching, uh, you know, different sports or activities have different value props in terms of market size, in terms of propensity to pay, in terms of whether the cameras are on-prem or whether you have to get the footage another way, et cetera, et cetera, et cetera.
Is there a sweet spot in your mind in terms of like. Like a, like a criteria for a category versus a specific category where the fruit's, the ripest, where it's like, oh, if it meet theses criteria, then it's well suited and, and you can still get it done in some of these other ones, but it's just more of an uphill battle.
Shaun Meredith: Um, I do, I think, I think what you have to do is you have to look for something that's evergreen for one, right? So, so, you know, as you kind of mentioned, uh, this is, this is where it's interesting, like, um, uh, you have to look for something that's evergreen. You have to look for something where [00:52:00] there's a significant number of participants, so, right.
Um, at, you know, to use your hockey analogy, right? Um. Obviously there's room for some small amount of companies to do something like that, for the NHL, right? But I think there is a much bigger opportunity if you can help the youth hockey leagues around the world, And so this is where, uh, you know, same thing we've found that, that focusing on the gamers, you know, at every level and bringing them along. Is, is a much more attractive market, um, and one that deserves kind of that venture scale a lot more than, than, you know, those top tier teams. And then the third thing I'd say is looking for, some sort of, uh, community that's already [00:53:00] leaning towards technology. Um, and that's one of the reasons why like. You know, we think we can be more successful in gaming because they're already used to using some sort of technology tools. Um, you know, there, there were gamers out there, uh, using AIM Labs to practice for their, their aiming and stuff or warming up before they go, into a match. And so they're already kind of comfortable with technology tools to help or assist them. Um, you know, if you were going, I think, uh. You know, something like, you know, hockey or basketball or baseball, something where they already measure things a lot and they're already, you know, they've already got some cameras, like you said in the arenas and stuff like that. I think that's a better bet than trying to go after, you know, maybe, uh, rodeo or polo or, you something like that where they're just not as far [00:54:00] along.
Jason Jacobs: Mm-hmm. Uh, interesting. And if you were starting today, uh, and pursuing either this path or in one of these adjacent categories, how would you think about, uh. DIY versus, um, utilizing these LLMs, for example, or, or other kind of off the shelf tools, and also how would you think about capital sources?
Shaun Meredith: I, I would actually, I would actually, um, you know, look at the tools that are available out there and try to stand on the shoulders of the giants that have already
Jason Jacobs: Mm-hmm.
Shaun Meredith: Um, the reason I say that is it also will prevent some technical debt in the
Jason Jacobs: Mm-hmm.
Shaun Meredith: um, as an example, when we first did our prototype, you know, the. we did was, was, you know, homegrown, uh, essentially code, right? So, uh, a lot of, [00:55:00] uh, uh, c um, and, and, um, what happened was is, we started to notice that many of the chips that were being developed were coming out optimized for tensor calculations. So it made more sense. start to use tensor libraries and build upon that, because that's where we would start to see, you know, costs start to, you know, in the future as, as, uh, these chips were more and more optimized towards tensor operations.
Jason Jacobs: Hmm. It seems like it'd be really fascinating to lay out the different potential categories where this type, this type, these, this type of kind of cv, you know, computer vision, ai, machine learning, like where that could be applied for coaching and analysis, and then. Like, what are the criteria that would be more conducive to cloud?
What would be the criteria that would be more con conducive [00:56:00] to on device? Um, what are the criteria that would be more conducive to a hundred percent ai? What are the categories that would be more conducive to human in the loop? Almost just like kind of laying it all out, because one of the, another thing I've been thinking about is just.
How much economies of scale is there, if you kind of stood up this infrastructure and expertise under the hood to go apply and to build a fit, you know, across markets, across categories, but, uh, but leveraging a lot of the same underneath, is, is that an angle that you guys have thought about directionally as a, as a way to maybe, um, increase the market size, but also, uh, leverage a bunch of the infrastructure and expertise that you already have?
Shaun Meredith: We have, um, and honestly, uh, you know, it's just something where, you know, we've, we've put that down as. know, that might be future growth and let's just focus on a niche that we can really serve right now,
Jason Jacobs: Mm-hmm.
Shaun Meredith: uh, to, you know, essentially, you know, establish a foothold and then be able to say, okay, now we have this layered AI approach. [00:57:00] You know, let's see if we can, you know, apply it towards some other, you know, uh, industries or other things. 'cause as you mentioned, right, so. If you step way back at a hundred thousand foot level, you know, or you know, out in the ionosphere, you're essentially taking and, and trying to make software that can coach somebody doing some sort of activity. And, you know, who knows? Maybe, maybe that's, you know, helping to train doctors, perform surgery better,
Jason Jacobs: Hmm. Yeah, I mean, I almost, I almost wonder is there like a, like what Squarespace is for websites making it so easy for people to build? Like for someone like me that wanted to go, let's say I wanted to go pursue this hockey thing, like it seems like you have a lot of infrastructure that I could already leverage.
Like, you know, could you package that and somehow, you know, could I pay for it on a subscription basis or something and be able to build on top of it? Yeah.
Shaun Meredith: yep. And there's [00:58:00] that, you know, that could be an alternative, you know, business model. Yep.
Jason Jacobs: Great. Well, what does the next 12 or 18 months look like for you from a priority standpoint? And where do you need help if, uh, if, if anything.
Shaun Meredith: You bet. Um, you know, honestly, uh, like I said, our, our true north metric we're using right now is, is users and, and we just need to get word out that, you know, here's Omnic, it's a tool you can use. You know, you should be using it. Like, you know, nothing else, sign up for a free account because. You know, if you want to get better at gaming, which we know every gamer out there does, nobody plays a game so that they can be worse. And, um, so we just need to, we just need to build awareness so that people go, wow, oh, this tool exists. Uh, because that's what we've been finding, talking to our current users is, is, you know, they're like, wow. Um. I didn't know this existed. now that I do, it's [00:59:00] actually great. Like it's actually helping me and I'm getting better. So,
Jason Jacobs: And how do people find you today?
Shaun Meredith: yep. So, um, the, the way you can find us today is same thing. Watch, uh, our TikTok is, is. Uh, a good, uh, avenue. Um, also, uh, as I mentioned, some of our partnerships, right? So we're starting to drive with the Veterans group, uh, through M 80 in their academy through play versus, um, and literally, you know, sign up today@forged.onc.ai.
Jason Jacobs: Great. And, and Shaun, uh, if, um, if Omnic AI is successful beyond your wildest dreams, what does it look like when it's in that form? Like, let your, let your imagination run wild From an ambition standpoint. Yeah.
Shaun Meredith: honestly, uh, like I said, I really want to focus on, you know, like it's that whole thing that I think it's very frustrating. Uh, if you're a gamer and you're playing a game and you get stuck, you know, like you're, you're just like, I [01:00:00] can't get past, you know, I'm, I'm here grinding away and I can't get past gold one, know?
Um. Just anything to, to be able to, you know, help any gamer or have them go. Yeah. You know, like, I'm not the best gamer in the world. I'm not ever gonna be pro, but, you know, this helped me get as good as I can be and I enjoyed it.
Jason Jacobs: Mm-hmm.
Shaun Meredith: I think, I think that's the big thing is, um, you know, life's way too short. Uh, I think everybody should just enjoy, uh, what they have and, and, um. You know, I want to see people, if they want to be better at Elden Ring, they go and they can use the Omnic tool and it, it helps them progress faster or they're really busy at work. They only have, you know, 45 minutes a day to game or a week to game. Well, how can I use that most effectively and, and be able to enjoy that game? Um, because we [01:01:00] have some truly amazing like. Like game designers and artists out there that are making amazing, like worlds and, you know, puzzles and, and you know, um, uh, looter shooters and, uh, open worlds. And it just seems like, you know, we now have an art form that's kind of also in that technology realm I want people just to be able to experience that and, and enjoy it.
Jason Jacobs: Well, I think that's a great point to end on unless there's anything I didn't ask that you wish I did.
Shaun Meredith: Not that I can think of, you know, this has been great. I really enjoyed the, the discussion,
Jason Jacobs: I did too and I learned a ton, which probably means that at least, uh, some segment of our users that, that are trying to figure out this world as well are, are, are gonna say the same thing. So Shaun, thanks. Yeah.
Shaun Meredith: learned a ton too. So, you know, I, I think, I think that's, I dunno, that's what it's all about, right? So let's help each other out.
Jason Jacobs: Well, I appreciate it, [01:02:00] Shaun. Uh, thanks so much. Best of luck to you and Omnic ai and we'll be rooting you on from the sidelines.
Shaun Meredith: Thanks, Jason.
Jason Jacobs: Thank you for tuning into The Next Next. If you enjoyed it, you can subscribe from your favorite podcast player in addition to the podcast. Which typically publishes weekly. There's also a weekly newsletter on Substack at the next next.substack.com. That's essentially for weekly accountability of the ground I'm covering, areas I'm tackling next, and where I could use some help as well.
And it's a great area to foster discussion and dialogue around the topics that we cover on the show. Thanks for tuning in. See you next week.