Building AI Ads with Ray Jang (Atria)

It used to be more like, hey, you can kind of fluff. You can kind of do the job and you'll be able to get by because there was no strong alternative. But AI, what it's doing, it's been a great binary filter. Do you truly belong in this craft? Do you treat it as a craft or are you more just there to kind of pay the bills but also kind of fluff, right? And so in that sense, I think it's become this great realizer as well in terms of where you kind of need to go and direct your energy in terms of your profession. Hi, my name is Demetri Bonichi and I'm a content creator, agency owner, and AI enthusiast. You're listening to the AI agents podcast brought to you by Jot Form and featuring our very own CEO and founder Idkin Tank. This is the

show where artificial intelligence meets innovation, productivity, and the tools shaping the future of work. Enjoy the show. [Music] Hello and welcome back to another episode of the AI Agents podcast. I'm here with Ray Jen, the founder and CEO of Atria. How you doing today, Ray? >> I'm I'm doing great. Excited to be here. >> Happy to hear that. Yeah, for sure. Excap happy to have you. Um, just to kind of kick things off, uh, really like checking out what you were doing over there at Atria. Um so you founded it in 2022 is that correct? >> Company itself has been around since 2022. Atria is a pivot that we started um about 22 months ago. So slightly less than two years. Yeah. >> Gotcha. So what was the uh original thing? What was the pivot? And um >> yeah, >> this is a follow-up question

that could be in the question. Um how did you kind of get into AI? What was your path there? >> Yeah, 100%. So I used to work at Tik Tok. I was a senior product manager there. I worked there for close to two years. Um went through the grind. It was a dog eats dog world but in the best way completely um you it prepared me for entrepreneurship. I come from a family of entrepreneurs. I knew I wanted to go start a company. Uh I left Tik Tok and I w I did a 11-day silent retreat be part of meditation and I kind of had this feeling that hey I want to reconnect to my childhood self. >> What was the thing that I gravitated most towards when I was a kid and it was gaming, right? So I love playing games really more for

from a competitive and team bonding and strategy standpoint. So I was aiming to build an alternative to Steam. Um, little did I know it was too uh like it was like um too big of a project and many people have died in the graveyard attempting to do it. But I felt like hey with this conviction I can get there. So I assembled a team um you know slightly a year and a half into that journey. We were flatlining in terms of our growth. We were growing but not like something that would had dramatic scale. Um, and with three months of runway, I decided to call it and say, "Guys, we we got to do a complete 180. Um, we'll have to pivot and find a new direction." And um, while I didn't have a plan B with the um, the time running out the clock,

um, I was summiting like the one of the highest peaks in um, Southeast Asia, Mount Kinbaloo. And I had this realization on the mount uh, the mountain. And I had this epiphany that hey, we're not growing. Why not just focus on building a platform that helps companies grow? Like this has been so painful figuring out how to get distribution and get people to care about us. Why not just solve for our own true pain, right? Um and so that's where we um put the team together. Um had to unfortunately say goodbye to some of the other team members. It was very painful. And then we told ourselves in in 3 months we got to make this work otherwise the company shuts down. And so we built the first version of Atria in January 24. And then within the first week we were already making money

and that's when I realized wow there's something here right um and the way I got into AI is I was always curious and learning about it on the side more a dabler um and I had the stint at a insure tech company um where we were trying to figure out how we would apply machine back then it was called machine learning right uh applying machine learning in terms of claims processing uh policy uh management and that's how I got my initial exposure but uh it was clear to me um with all the um uh uh uh improvements in image and video generation that there was a a a world to be um unlocked through not only automation of um copy which was at that point the main thing right through GPT and using different frontier models but you know going multimodal right? Um to be

able to string together image, videos, the full stack with AI. >> Okay. Um it's interesting to me. So you I mean that that's a pretty bold thing to do, right? Like just completely pivot um be like, "All right, we got to make this work in three months." What was like what was going through your head during those three months? >> Wow. Um I must say that in my adults life it was the hardest period in my life. Um because >> and it was self-imposed. So I think to to that level I was able to go through the pain. But I I must say that in I would say I was and to still some degree am an entrepreneur that believes that if you have conviction, you're able to go much further um in terms of your idea um uh distillation and um distribution. But then

one thing I underestimated in that journey is listening to real world data and signals and knowing how to make that part of the equation in building your ideas. Meaning you should be open and flexible in terms of the arena of like where your idea lives and then also in terms of your process. But I was not flexible. I said it had to be this way. And the the the real pain was letting go. Letting go of that idea that hey, I I need to give up. There's like there's that giving up at this point is actually healthy. It's actually to be able to recognize to take the step back in order to go two steps forward. And so I had to learn to let go not only of the idea, but I put myself um in isolation. I ate one time a like only once

a day. So I let go of a lot of pleasure. Um meaning I I cleaned up my diet. I was eating very like um neutral foods to just calm the mind. I lost access to all sort of external stimulus. So no access to Instagram, no coffee, no caffeine, right? Eating once a day. Um, I lost friends because I was not going to parties. Um, I completely isolated myself and I dedicated myself to the the idea realm to to then go back into understanding, hey, why did the first idea fail? What fundamentals did it lack? >> How do I ensure that if I build a company with the right fundamentals, we will inevitably succeed? I almost made sure I I treated almost the idea like a science if that makes sense like the science of finding product market fit. I was obsessed with it. So you

know obviously it's not only about like going after a big market but going after a customer that you can directly address. I no regulations um higher um ACVs customers pay on day one. there's clear time to value it if people are able to um see effects in terms of workflow with AI and a problem that we could identify it so that we can feel the pain and continue on the journey even though like so that we can truly relate to to the customer that we're building for. So I came up with this framework and I you know through that it was a prism or a filter that I had to apply every new idea consideration. It's not like I went from zero to hey this is going to be the new idea. I had looked through at least a thousand ideas a day on Crunchb

on Y Combinator all these places um to have even more conviction on the next thing and that's how like it happened. So I put myself in a cave um and I didn't come out and until I told myself we had the next thing. Um it was the most painful thing but you know they always say on the other side of pain is what you're looking for and uh that I did find. Yes. >> Yeah. I mean, you feel pain or you feel disappointment, right? That's kind of the it but at some point you feel pain, right? Is the the old trope. So, it's it's a it's like are you doing it to yourself? Um whether it be upfront or on the back end of being like I wish I would have done something. I think that that's a fair kind of assessment on on the

way that it works at least from what I've experienced. But no, that's that's that's really cool. It's um that's incredible. I like that. I like hearing your story about that. So just to kind of dive into more of what you're doing now, um what I would love to to learn a little bit more about is you know as Atria has evolved from like um it's so it's it's an ad intelligence platform. >> So yeah so just to give you a bit of context um we started more as a competitor spying tool and then the key thing that we are today is creative intelligence. >> Yeah we're creative intelligence end to end. So we help you um optimize and create and uh soon we're adding launch as well. So you can directly deploy ads automatically through our platform for Meta and Tik Tok with YouTube coming

in in quarter one. >> Okay, nice. >> Up next year. Yeah, >> that's that's awesome. I know that uh there's a lot of different platforms to choose from. Um so it's got to be uh interesting to kind of know where to go when. Um it's uh something that I'd imagine you kind of how how do you go about picking what platforms to kind of target first? >> Yeah, I think uh you know each platform optimizes for different things, right? So um YouTube is education. People go there to get educated less I mean there's still entertainment but people love to learn. So, if your if your um ads are more tailored to people who've already heard about you but want to learn more, YouTube is a great channel for more longer formats. Um um and then also within shorts, people are there more to uh learn

and and meta is more about top of the funnel, right? It's more about hey building awareness, getting people to know about me, um being able to be retargeted at the point of consideration. So extremely consumer oriented companies, right? Um and direct response companies like e uh e-commerce D2C perform really well there, right? Whereas if you're if I'm a little bit more like a lead generation workflow, there's complexity in the workflow. I'm a B2B SAS app. That's why you get exposed to ads from ClickUp and Monday.com more so on YouTube. >> Yeah. So like different companies um leverage platforms different. >> Yeah. Yeah, there you go. Yeah, they never Yeah. So, I mean that that that seems they they know their persona, right? You're you're a podcaster. You think about AI agents and there you go. >> Well, actually, you know what? I I I will

say that's funny enough. I have had less since I went from more of a feed about like productivity apps to more AI. So, they they got me when I was the AI uh or sorry, the productivity app YouTuber versus the AI YouTuber. So, they they it's incred Yeah. It's incredible how this kind of stuff works. >> Yeah. Makes sense. Makes sense. Yeah. >> So, um just to kind of go a little bit more in depth of what you guys do, um tell me about some examples of different campaigns or like people that you've worked with and um kind of how how you would say uh an example relationship would go with a customer that gets you to help them uh do a better job in the space. I know it's it's like a pay per month thing, but I'm sure you have some like specific

experiences maybe with uh companies that have like gone for the business or enterprise level. Like I just kind of want to hear a cool story. Yeah, >> definitely. So, it's funny. When I first started the company, there was this there's this company called True Classic. True Classic makes amazing um >> Yeah, the t-shirts. Exactly. Like right right for you. Um and their whole thing is like being being like making you like look look make your body look great, right? um with their t-shirts >> and and so when I first started the company, they were a dream customer of mine because uh they just really understood creative experimentation. They were so experimental and they were always at the cutting edge. Um and when I built out uh Atria, they wouldn't even bother to get back to me. I cold emailed the the founders. I got one

of their creative strategists to start using us. They still like it was crickets. um two months two a month and a half ago I was in uh LA actually sitting in their office with both the founder and uh who's now become the chairman and their CEO and we were just locked in talking about everything that's happening in the creative strategy space and and you might ask why and that's because they they realize even at their scale where they're spending more than seven figures on meta on a monthly basis that they need a platform that truly gives them strategic insights on all the different creatives that they're putting out there because if you run thousands of ads, it's impossible to know, hey, which of these are actually worth double down like double downing on on the spot, right? Because you're just not going to be able

to go through thousands of creatives in a day. You're going to have to build an army. And that's not the goal. Everyone's goal right now with AI is to trim the team and have truly the best and the brightest only in that team perform the job of two three people. So that's one of the key thing the the key things that we were helping to solve. The other thing is they love Radar and what Radar does is essentially our in in-house algorithm that scans through all your creatives automatically scores the different building blocks of an ad. So, conversion, retention, hook, click-through rate. It's able to give you a score like a homework ABCD and then gives you a sense of, hey, these are the components that I have to go and improve on the spot. And it gives tailor made recommendations for that specific creative

by scanning through the video creative using computer vision to understand what's what part of the script can be improved, gives you suggestions. So, it's almost like a marketing teammate that's built in for you. The way we're evolving that is we're going to not only um make now the actual creative, but we're also going to couple it with a lot of the benchmarks uh in terms of understanding how other people of similar kind of um uh um like with within the t-shirt or the e-commerce space are performing and giving you better industry understanding. So um that's one u great uh customer case and you know it's it's funny like uh the founders were like actually you know it just continue doing great work out there and people the the the dream customers will inevitably find you. You don't want to go searching for them. So I

think the the reality is probably more the truth is in the middle. But yeah I I found that that was like a great example. And then we also help like growing brands right? So, it's not necessarily the nine fig brands that are only um like uh like uh doing well on Atria, but say you you you're starting like there was a a founder of a protein bar um like that that really got inspired by David the protein bar, right? And he was building his own version and he he he he was in this um kind of spot where he knew slightly enough about creatives, but he wasn't quite ready to hire an agency yet, right? So he started using our platform to not to to go into the ad inspiration library where we pull from meta ad library and then we he started clicking clone.

We have this tool called clone ads where you can actually go in, he can look at, you know, protein bar companies seeing what kind of ads they put out there and then he can click the uh the button clone, put his website, we automatically generate a creative brief and then we start creating creatives in his brand tone so that he has a suite of ideas that he can work with, right? And so our goal is not actually to help put enable people put more AI slop out there because Meta has already count is already counteracting against that. Our goal is to put as more human inspired more hum more in like ingenuity behind the idea more strategic thought so that you can actually do with more with less. Right? So that that's really the the two drivers of our platform today. And of course like

actually the just just so you know and it's really important to highlight we're we're not here to replace agencies. We're here to empower agencies, right? So a lot of our customers are like the Vayner Commerce of the world, right? Gary Ve's um agency uses >> me some Gary Vee. >> Yeah, he's you know he's the OG sort of >> Yeah, there you go. So like uh you know and you know hopefully our us being able to run better ads for his team will will help him um be an owner of the Jets but like you know like uh that's that's what we do is we also en you can imagine for for an agency their metric is insights per minute. So they actually love platforms like Atria because it helps them be more useful, more contextual, more um strategic across their uh their their their

clients. So we we actually see a good split between agencies and brands using Atria. >> Okay. Okay. Very cool. Um I I don't know. Did you get the By the way, did you get the blueberry joke or No, >> no, no. Okay. So that's So G Gary has um there's a there's a YouTuber called Captain Sinbad who did this series of videos called um watches blank once and he did a meme of like he did it's like a meme series where he he mocks or like just in in a playful way like does a impression of the people and in it he talks about blueberries because I don't know if you've actually like and a lot of people do this impression with Gary Vee because for some reason he has an affinity for blueberries but it's hilarious when when he talks about he's like blueberries,

man. They're like a superfood, dude. Like, it's so funny, right? So, I was just checking afterwards. >> Look up like Gary Vee blueberries. The amount of memes people have made out of this is ridiculous. But I really like to hear that you guys are, you know, not trying to replace agencies. I used to work in a marketing agency myself. Prior to starting a uh content marketing agency, I worked in paid ads. So, I was a paid search manager um at a decently large firm in Chicago that >> um managed uh I managed uh like 70 million a year in ad spend on paid search. So, it was a pretty good experience. Have a lot of background in that. Had a short stint at a smaller agency that was doing uh pretty much social search programmatic. I it was like a more boutique agency, so he

had to do everything. Um and yeah, no, I think it's pretty interesting just from a product standpoint. How does it work in the sense that it's able to recreate different ads well without kind of being repetitive and whatnot? cuz you know I can imagine ClickUp has a a style and you know its own thing and they changed from that whole one app to rule them rule them all messaging that kind of burned some bridges because they just they were just like a little too annoying online and they've chilled out for a while and try to be more No, I I had a one-on-one with someone at their company that that was one of the higherups cuz I was I was focused on making productivity app videos at the time and I made a video called Click App Sucks and I had a I had a

meme series I had a meme series going but I made another video that was actually like a not a meme. And I was like, "All right, I'm going to be honest with you guys, ClickUp. It's actually ridiculous. You need to stop doing this thing where you pretend like you can do everything cuz then you release too fast and then you have so many bugs no one wants to use you anymore and then they're mad. So like you got to tone it down." And funny enough, I had a one-on-one with someone who's the head of their marketing. >> Oh, no way. >> You're right. >> And I know exactly. >> Was it Chris? Chris Cunningham. >> Mhm. Yeah. Two years ago, I had a meeting with I had a meeting with Chris and he was like, >> you you you called us out. It's true. I

was like, I'm glad you listen. He's very nice. Yeah. Um, so that was good. And then they from there they they've pivoted the last like two years from that sort of over the top style. And I was like, good, man. Like you're relatable now. Cuz before it was ship, ship ship ship fail fail fail fail. >> Yeah. Yeah. I mean, >> and it was anyways, but to get back to it, like how do you kind of repurpose stuff from other companies without being repetitive from what they're doing and uh without meta kind of freaking out as you just referenced? >> Yeah. So, I think the key to a mutation is to ensure that as part of that evolution like you know, I take a lot of inspiration from um like Darwinian effects and um like evolutionary um science, right? So essentially in in part of

this mutation you have to ensure that as part of the input you're introducing a little bit of randomness but that randomness coming from a source of ad principles that they haven't tried before. Right? So it's it's a key ingredient to ensure that you are able to put as part of the basis uh ad principles but also sprinkle on top elements or ad angles that they haven't tried and because and the reason why we're able to do that is two things right we know their historic data because we're we're able to plug into their ads manager and then we also know what they're currently running. So it's important that we index on a bit of the random randomness, right, to be able to see what actually picks up instead of trying to um repurpose their their thing. And then the second is we're able to also

be more um grounded in terms of the validation of these new ideas because we're able to source them from nonadjacent industries, right? So you'd be able to find that, hey, actually because I'm a productivity app or a productivity platform, it's not necessarily that I should go and copy exactly what monday.com does, right? I should actually look for, for example, what Grammarly does or um maybe like another product, maybe even non-productivity that is a great B2B SAS company, right? So, you know, um, Sprouto does interesting ads out there. Yeah. Right. On on LinkedIn, and that has nothing to really do with, um, project management, even though they do have elements of it. So, it's important to take that as part of the equation so that it's easier for companies to be able to draw inspiration from non-adjacent um, industries as well. >> Yeah. No, that that's

actually a pretty good that's a pretty good thought because I I don't necessarily think that um just because it's in the same industry it needs to >> or sorry just because you have something does it need to get inspiration from the same industry right >> you might even argue that it's worse right because it's it's kind of like it caps out >> yeah exactly you cap out much faster you you become more of same of s like you you perpetuate sataness right um to your audience >> yeah no that's fair I I hadn't considered That's really cool. It's a good thought. Um, nice. I I'm And then you're putting it out there, huh? >> I'm pondering that for a moment. Well, I I you know, hey, uh, it's, uh, it it tells you what the difference is between you guys, which is ultimately what I

try to get to in these convos. Um I I think something key right now that I want to just ask your opinion on is a lot of people are you know as you had mentioned concerned about AI taking over to an extent of uh people's jobs. What is kind of your opinion on the creative side of things here? Right. We have graphic designers and whatnot that exist and they'd be like well I think I should be able to make these images. Do you feel like this is actually giving more people an opportunity to do these types of things that have the skill set but at a larger scale? Do you think it's allowing people who don't have that skill set as well to become a part of the creative side of things? What are your thoughts on kind of how not only what you're doing,

but this AI creativeness is going to impact the jobs that would usually be the only ones to do it. >> Yeah, 100%. So, I think uh one thing that AI is doing is it's making outcomes and jobs and like the the people who are in it more binary. kind of fluff. you can kind of do the this craft? >> Do you treat it as a craft or are you more just there to kind of pay the bills but also kind of fluff, right? And so in that sense I think it's become this great realizer as well in terms of where you kind of need to go and direct your energy in terms of your profession. Meaning if you were a creative and you found that actually being in the creative industry is a lot more data oriented, more scientific, less creative so to speak. then

maybe you find a path in more um in the in a in a in a type of job that really truly more fulfills your creative purpose, right? Um maybe it's in cooking, right? Where where maybe AI is less likely to take the job, right? Um so that that is like the first kind of highle thought. And then the second level thought here is that AI what it does is it reduces the floor. It brings down the floor. It gives more access to people to enter because there's less gatekeeping, right? It it used to be you had to prove your worth by going through like realms of both politics and time in order to reveal your true strength. Now AI has equaled the playing field. So it's reduced the floor but also increased the ceiling. And then so the the middle is kind of getting wallowed

out, right? So there's more people being able to enter which is a great thing. It gives more opportunities, gives more exposure. We we find a lot of affinity towards interns and recent grads who just have natural innate talent because they're able to just prove their worth. We don't have to uh rest and believe in and just only index on a resume. Now I can see that hey is the 19 or the 22year-old as strong of a marketer versus someone who's been in an industry and is 37 right it it's so so it's it's kind of brought down a lot of the preconceived notions right of what makes a great professional and it's also increased the bar and what I mean by increased the bar specifically as it pertains to AI is that AI is very good at pattern breaking meaning it's it's like you you

can actually use a AI and then go and build CGI effects the level of a Hollywood studio if you have the patience and the creat creativity and the technical prowess to actually leverage the nano bananas right of the world. And so effectively what the the the true creatives that are coming in are not trying to come in with some formula. They're saying, "Hey, actually, if I'm a dentist in LA and I can't get clients, maybe I actually dedicate a month to learn about how to create great um videos." And then from there on, I put two gorillas or orangutans with the widest smile, the cleanest teeth, right? um that's picture like like so white that it's glazing on the camera and I have them doing fun things adventuring around life and at the end it's it turns out it's a commercial for a boutique or

a small dentist dentistry in LA right so in that sense I' I love that AI is empowering the the creatives without um by by reducing that gatekeeping right it used to be if if you wanted to go into um uh the creative industry, you had to work at an Ogov or WPP, right? There was all these like credentials um indexing where you you had to do your you had to pay your dues there before you can come out and start your own firm. Nowadays, you have a landing page, you you're you're thoughtful, you put enough consistent posts on LinkedIn, right? And then you're just putting out interesting videos of you leveraging AI, people will get interested in you. you know, there's this guy on LinkedIn. Um, I I I don't quite recall his name, but he's an Italian guy based in Milan, >> and he's

putting out crazy AI crazy AI videos, you know, all the time. And and I'm suddenly super interested in this talent, right? And so, in that sense, it's been a great equalizer. and and and and I think uh it's it's important to also see the positives despite some of the negative um sentiment around like the job replacement um um sort of a um uh story that is out there where where it also has a lot of weight and and truth to it as well. Right. We are going through a transformation definitely as as we understand the nature of a job or a nature of a firm even um uh these days. >> Yeah. No, absolutely. I mean, I run a content agency that um utilizes AI heavily. And I think what the thing that I noticed about just talking about an interesting topic prior to AI

being the thing, I mean SAS and just being more organized and automation, that kind of stuff was was very hot. And um I made a tutorial every day for 300 days and I just was able to quit my job. >> Like it actually was that simple. Like I don't want to be like rude to people who are like, "Oh, it's not that." I'm like, "No, it it is like you you just like talk about a subject and be good at it for and granted, did I spend a couple years getting really good at understanding how this stuff worked before I made content?" Yes. So, but in a couple years if you just continuously talk about a topic and show proficiency on something that other people couldn't, I mean, it's it's crazy nowadays what's kind of available for you. >> Yeah. I mean, what what are

your thoughts after having all these discussions on like AI, you know, like uh >> I mean, so there's kind of three cans. Basically, it's now my fun little thing to talk about when everyone's going and asking me about it at like parties or whatever cuz they feel like, "Oh, you run like a AI podcast. That's weird." And then they look it up. They're like, "Oh, it's decently substantial, so you must have an opinion." I'm like, "Yeah, I mean, I try to have one, but this is where I kind of land on things. Everybody has an opinion on AI, which I think for the most part decently warranted. um whatever. If they're running a company, they usually have some reason for it. But you kind of fall into three camps. AI is going to replace jobs and it's going to be the the nightmare scenario. I

think that's the least popular thing I've heard on here. Um second most popular one would be it is going to enhance people's outputs and we maybe just won't hire as fast as we used to. And then the other one's like, well, you're going to create a lot of people who can do exactly what you just mentioned, which is like hyperfocus on a skill or hyperfocus on something and be able to do a lot of things with AI so they can build a one product or like one niche really good business. And then from there, I kind of fall between number two and number one where I think employees who are going to be able to utilize it are going to be very effective. And I think there's going to be a sub substantial amount of soloreneurs or two three company prneurs that are essentially cranking

out a specifically really cool thing. And um we will not see this the nightmare scenario. I think that's a little bit of I think that's a little nonsensical cuz practically speaking I I just don't see it happening. There'll always be jobs found. Like the best example I could think of is uh they got rid of um people thought there was going to be no more bank clerks because the ATM was invented. Um bank clerk jobs went up like 250% when the ATM was uh invented because um banks were able to be decentralized and they opened up more banks. >> Yeah. Yeah. Exactly. >> So I'm like we don't know but so I'm like practically I don't think it's going to all disappear. I think there's going to be somewhere between those one and two options is where I sit. Yeah. >> Yeah. Yeah. Yeah. It's

definitely power to the individuals. I I'm seeing this left and right like I'm seeing all these solarreneurs uh come out and they're able to get like yeah they AI agentify their sort of like processes right and then they're able to get so much further. Um yeah many many people living that kind of 4hour work week actually this time >> I I'm trying to work figure out because I I do a lot of stuff for the content agency. I'm trying to productize everything like sort of what you're doing. I'd love to take how my mental models work for all these different categories and do what you're doing. And it's it's definitely top of mind for me. I'm like I'm trying to identify I'm trying to implement so many things. It's um it's fun. It's actually it's actually so much fun. It's kind of like a video

game in a sense because I I grew Yeah. It feels it feels like cheating or feels like um I don't know. I played a lot of World of Warcraft. I don't know if you've ever played >> Yeah. Yeah. I played Dota, too. Um, yeah, but not world. Yeah, not Wow. >> it kind of feels like an RPG where I'm just making a bunch of different characters and then I'm getting more and more gear and the and then randomly there'll be an update and I don't I I log in one day and the world just gave me gear like because, you know, cuz the models get better. It's so fun. >> Yeah. It's it's it's it's a the best time to be a builder right now. Yeah. >> Yeah. if you really know how to mentally model these things out, I think you're gonna do incredible.

Um, so kind of talking a little bit more about that. Um, and what is your long-term vision then for what the the company will look like and what the product will look like in 3 to 5 years? >> I I think uh I mean because of the first failure, I try not be too deterministic. Um, but I have different flavors that I think uh are going to be interesting in terms of the future of the company. I think uh one um idea that I have is um a like a product market fit engine. Meaning if we're so good at um creating and identifying um consumer signal and consumer interest in products, you know, as a company, if you want to put out a new skew or if you are a new company that has no product and you want to be able to see if

there's going to be actual buyers in the market, true intents for what you're building, can we actually build a platform that can PMFI your your products as soon and and as quickly as possible and give you the validations without either having to go build it or having to do thousands of um customer calls. Right? So that's one flavor. I think the other is what as well that I think is interesting is to build our own advertising channel, right? Is to build our own advertising network. People are always looking for new pockets of distribution. Um to me, what's interesting with all the AI um chat apps going on, right? or maybe uh in a year's time it's going to be glasses. There's going to be new form factors as it relates to advertising that's going to create new pockets of distribution where it will create opportunities

to create um to enable platforms like us to turn into a channel itself. So you know um so that we can um complement the efforts that are going into Meta, Tik Tok and YouTube because specifically people are increasing spend in these channels but at the same time also diverting spend to looking for new net new um platforms that they can arbitrage right so that's uh another one and then I think lastly I the the the the focus is like how do we become the most insightful platform uh in terms of your customer, right? How do we ensure that that hey, if you are a product and a platform, how do we synthesize all the things that are coming in through like um your if you're a consumer brand through like census data, credit card information, um like uh like like people who churned uh people

who are still retentive. How do we like really give you a sense of who are your personas, who you should be targeting and enabling you to go and get in front of these people, right? Either through organic or paid efforts. Um, and then of course like uh if you're a B2B company, that's like going to be CRM, sales data. So I think I think there are multiple flavors that are out there and I think my key learning is to not be so beholden to one is to be able to actually internalize the trends and data that's coming in and also develop a bottoms up understanding in company building. I think that's that was what was missing in my top like my first two years of building Atria um or the company itself right because you know it started 2022 that I was very top down

driven I had this idea the rest of the world had to follow now it's I have this idea the world also has its own understanding and its reality and where do I find the merge of the two um in terms of like the future development of the company u but who knows maybe all of this kind of goes to to the dumpster and then you know it's all about robotics and you know I'm building a robotics company so we'll see. Stay tuned. Right. I think uh it it'll be interesting to do um a podcast like this with you um this time next year. >> Yeah. Yeah. I I really appreciate that um you and I are kind of keeping it. Uh, I I I think, you know, sometimes people have very pie in the sky ideas um about where the future will go. And I

do think a lot of that is stuck in what they'd like versus like maybe practically how the world's going to go. And um you can never really predict that kind of stuff. So I appreciate that. That's fair. >> Yeah. And I think and just just one more quick thing there is like the become the world has become a lot more nonlinear because of how things are progressing right it's just like the the the your prediction cycles are so it's like it's like of course you can never predict even tomorrow but I mean just in terms of macros the the the level of dynamic trends that are concurrently happening at the same time is never been more like each each of these specific trends are at an accelerating pace, right? Robotics is at an accelerating pace. Computer vision is at an accelerating pace. LLMs are at

accelerating pace, right? So, there's just like this in if you if you consider that as part of the infusion and the equation to trying to see the world, it it's kind of really uh myopic to kind of assume that you can actually even build a a year plan for a company, right? I think my job is like to just take into account everything that's happening in AI and almost be willing to go not 180 as much because doing that twi like every two years I'm going to question my life and you know what it means but like to truly every every sort of like 90 days to be like hey do we have the right set of principles or fundamentals assumptions in terms of where the world is going right um yeah so I think uh I don't have three to five year plans I

I only operate on 90-day plans now. >> H that's fair. I I think it's hard like you said that the time just does not feel the same that it used to. Like I I was talking about it with somebody recently. There's a insane level of uh quick um sort of turnaround on on model improvement at at a minimum, right? Like we're we're talking text inside of AI generated images was impossible until like March, right? And now Nana Banana comes out and you're like you can basically Photoshop anything in a sense like in the the colloquial term of what Photoshop means. >> You can replace anything with anything. You couldn't do >> video at all in a reasonable manner. Now, there's multiple competing models that are killing it, that are making better, funnier meme videos than you could make practically because people can come up with

crazy ideas and it'll bring it into fruition. Um, you have models that at first coding was like maybe helpful. Maybe you can do form like it help you with formulas and stuff in certain products, but you know, like is it really going to do a great job when it comes to coding? No. Now, we got models like Claude 4.5 Opus that send you from like a recurring can't figure it out cycle to insanely good suggestions and insanely good speed of coding. Um, I could go on and on and on, but you get the point. Like I thought text inside of images was for to use the kids lingo cooked um and was going to be for a while but uh now we're in now we're in a position where it's where it's cooking. Yeah, exactly. That was that was a good layup. Yeah, it went

from like that was dude cooking like it's it's insane >> the 180 >> and you don't know when it's going to happen. >> You don't when the models are going to reach certain levels. These companies are just continuing to innovate in ways that I would have never predicted. Um, like for example, Gemini is one that a lot of people aren't considering, >> but their visual uh onscreen context thing. >> Incredible. >> Because I I Yeah, they the I think the score on the benchmark went from 11% to 70%ish in >> Oh, wow. >> from 2.5 to three. Yeah. No, it is. And that that was like so I I run an agency that does like mainly content for SAS and AI companies. So we care about tutorials a lot. So we had a workflow where we were taking a screen recording and turning it into

a tutorial script. So that cuz the problem with like tutorials is you can't have a script that is based you can't have a script prior to recording the tutorial really cuz like you first you have to click here and then here and then here and then here and nobody knows where the clicking is occurring really from a >> right >> from you can't use AI to write that script. So we realized oh well what if we just screen grabbed it first and uploaded it to a model that understood it and in like May it could do basic clicking. I tried to make a video QA agent with it. Wasn't really that good. But then 3.0 comes out and I go 70% visual understanding. Okay. Now it uh skewing our videos. It's skewing our videos like uh in five minutes as well as a manager. It's

crazy. >> No way. So, so what what does it put out that has been kind of like a slight >> So, I mean this is my theory on how you can do anything in regards to agentic workflows now like uh micro processes or uh understandings of what is good or bad especially with content production. It's basically creative if then logic. So, all I did was I have a two or three hour course on how to make a good tutorial so that when we onboard a video editor, they basically become a good tutorial video editor within like a week. It's very simple. But I realize that it's just a set of 20 minute modules, like 15 of them, and I'm like, could I just like take the transcript, figure out the real advice I'm giving, and if the eye can understand things visually, it could run

through a Gemini 3.0 module every single time and QA at each level whether it's good and then at the end have a master double check to be like are any of these pieces of editing improvement advice contradictory or overlapping each other and then it sends a final JSON to my frame.io account and uh comments on any uh different videos that my editors would be editing. >> Wow. Yeah, but that literally was I had the idea in June, but the the model wasn't ready, >> right? >> You know, so that's something I've noticed. >> Yeah, like a lot of thing and that's that's that's just video. You can do it for everything. It's just a matter of the models weren't ready to understand context in large portions and they still can't do it like for all of the things at once. You have to segment it.

But it's like if I put that inside of a an app, right? I could make a bunch of parallel Gemini prompts and it could do it. It could like go out and then in because I was just doing in make.com so it takes like five minutes because it goes in order, but you can do it all at once. It could be done in like 45 seconds if you do it in parallel then bring it back together. So yeah. So >> as long as you know how these mental models work, people, you can do things. That's why I'm That's why I'm bullish on this cuz like Microsoft would have not implemented that for like a year. They're slow, you know? I did that in two hours after a meeting. >> So, it's like if you're a small company, you can move faster than these big guys.

>> Yeah. I mean I mean for the listeners, it's it's like uh I I would just learn systems thinking, right? It's like uh it's just incredible the like incredible things are happening. And if you're a great systems thinker, you can be a autodirect and just like teach yourself like your way into being being a great builder. It's we're all being enabled and empowered. We just have to come in with a positive frame. Absolutely. Systems thinking is a word that was popular a couple years ago and then I feel like it fell off a cliff from like a public conversation standpoint. But if you had had that prior, you jump right on this AI because it's if you don't have that going into AI, I don't really and and automation for me was a great building block because it helped me understand systems thinking and processes.

So now that we have workers, so to speak, that can really like take advantage of it. I think it's u it's an incredible position to be in if you know what the system's thinking. So yeah, sorry that was a fun rant, but uh this just happened to me the other day and I'm very passionate about the fact that like >> you guys need to know that even if quote unquote >> they're going to take our jobs, nah, we got it. We're okay. >> I mean, yeah, they're new jobs, right? So, yeah, >> they're new jobs for us. >> Yeah, >> for everyone. Um All right. Well, I I have a couple more little questions before we kind of close things out. I mean, I just talked about something I personally love. uh doing what is one of your personal favorite u tools right now to

use that's an AI that uh isn't you know your own company and product >> oh interesting um I love using whisper whisper flow wispr flow um so I you know I am able to write I I I used to have the biggest inertia to writing out long emails I'm just like ah I have to sit there type it out. Now I just ramble. I'm just rambling and then I'll dump it into GBT and be like, "Hey, rewrite this into a more professional email and it does all the bullet pointing, the grammat grammar, you know, check up the, you know, the start, the end." So, it's been it's been pretty gamechanging in terms of my productivity. Now, even when I have like product docs or like company strategy ideas, I'd be like, "Ah," like, you know, I'd be sitting there being like the cold starts sweat

being like, "Where do I begin? I have to outline things." Now, I just like subcon just let my subconscious kind of run, you know? I'm just rambling, you know? I'm just like saying things, right? I don't even have to care like which part it picked up, which part it didn't. I'll just dump it and it be and it go into GBT or claude and be like, "Hey, I'm I'm I'm, you know, this and then I'll even on claude I'll be using like, you know, prompting it with my voice, right?" So voice to text has been pretty gamechanging. I'm a I'm a very like verbal um thinker, right? Like I I I am able to learn and realize through conversations. So, and and you know, I think that that's one. And I've been loving um uh you know, during my travels turning on voice mode um

to ask a lot of questions. Like I used to feel lonely sometimes while traveling. Now I have this ever ever curious, ever patient u tour guide who's there for me to pick up on all the kind of questions or questions. You know, now the world is truly my oyster, right? So I think wi voice to text um whisper has been a lot of fun. Um and then you know one one kind of side project uh tool I want to tinker with going into the new year is definitely yeah n8 m.com right just trying out different workflows to see um what kind of like uh yeah automations that I can build out there. >> Yeah they're they're really fun to make. They're incredible. Um it's one of my favorite pastimes just making automations and stuff. It's also for work, but it's a pastime. It's kind of

a double. I love it. So, um, and with the power of AI, it's really like the best two for one you can experience. So, definitely recommend trying that out. My favorite thing right now is, uh, crisp. So, it's funny enough, >> what's crisp? Yeah. >> K R I S P AI. So, you, Funny enough, were trying to figure out whether background noise could get blocked out, um, as we're talking with some behind you. >> Yeah. It is a tool that manages to take the background noise and AI remove it and also is like a meeting notetaker. So, it's a two for one >> and um very cool. >> It's like a 0.5 second delay, but it works great. >> Yeah, very cool. I'm looking at I'm going to Yeah, this is this is definitely something I'm going to implement and and that's a great uh

brand name. Makes makes your you know soundpris sound more crisp. >> Yep. All right. Well, uh, last things, uh, I'm going to ask you right here is where can everyone go to check out the cool stuff you're doing? >> Yeah, 100%. So, I'm I'm, uh, uh, it's tryatria.com. Tri r y a.com. Um, clearly we need a new domain. Uh, we're going to aim to buy atria.ai um, next year. I hope I I wish I knew that, you know, prices were going to inflate like crazy. Um because that was actually the first domain that I was going to buy, but uh stuck to the.com. Um yeah. So, and then and then uh we're we're big I I'm I'm personally big on LinkedIn. Um we put a lot of like marketing trends out there. So, R A Y um space j um and in our YouTube channel

as well, Atria AI. But you know the key thing that you know I I would love people um and your audience to take away especially if they're are um B2B SAS founders or marketers is that like we you know Atria is is building uh AI agents um right now as we speak uh where you you're going to be able to have a AI teammate there with you to be able to autogenerate um creative briefs and then soon um um like readym made um image and video ads that will actually um per outperform most of your um actual uh ads at the moment. So that's in the pipes and so I'm super excited to announce it um in Q1 of uh next year. >> Absolutely. So with that being said, I hope everyone enjoyed this episode. Make sure to go check out Triatria. That's Triatria. T

R Y A.com. And make sure to hit that like button, subscribe, leave a review. Obviously you as well, Ray. You want to get this out to as many people as possible. Leave us a review on Apple Podcast and Spotify and uh check out everything they're doing over there at Triadria. Thank you so much for watching this episode/ listening and we'll see you in the next one. Peace.