From 82 Videos a Day to AI Empire with John Gargiulo

Speaker A: On today's podcast, I had the pleasure of interviewing John Gargiulo. John is the former head of global product marketing at Airbnb, the co founder of ReadySet, and now the founder and CEO of AirPost AI. AirPost is a platform that helps marketers to create video advertising using artificial intelligence. John and I got into the depths of AI, how, how marketers are using it to create videos. Now, the importance of prompt engineering and keeping a human in the loop when it comes to creating AI outcomes. The conversation provided a lot of value to me. I certainly learned a lot, and I hope that you will get the same value from it. So, hello, everybody, and welcome back to the Catching Fire podcast. I am your host, Brian Karras, founder and CEO of Playbook Media and growtal, and, uh, very excited to have John Gargiulo on the podcast today. He is the founder of Aeropost and also founded ReadySet, which is an amazing video production company. Does a lot of great shorts. And then AirPost, uh, is a new AI video editing company which we're super excited about and are actively testing within our agency. And so, John, super excited to have you on the show today. Thank you for joining us.

Speaker B: Yeah, happy to be here.

Speaker A: So, you know, the last time we chatted was in 2018, which is kind of crazy to think about. Ah, you were still at Airbnb and getting ready, set off the ground. Um, you know, we like to talk about standard founder talk to founders and hear about their founding stories. So can you give us kind of the background of what it was like launching that and running that company and now?

Speaker B: Wow. Yeah, it was really amazing. So I was at Airbnb leading parts of marketing there, and I was in a unique position where I was reporting to our cmo, but I worked with our head of performance marketing who had actually spun out of the cmos Org. And so, you know, working with both of them, you know, the latter said to me once about the former, you know, I love Jonathan, but we come from different planets. It's like brand marketing world. And Jonathan's a brand genius and incredible human and leader and the art of it and the, uh, sort of overcoming obstacles not with, like, pinpoint features, but like, with belong anywhere. And the feeling of Airbnb so important. I think it's important for clicks and conversions as well. We're both more likely to tap a Nike ad than an Acme shoes ad for brand reasons. And performance was really coming up fast with Facebook just crushing it for people. If I'm not making my money back in 30 days. Why would I keep doing it? Uh, and so I felt like I had a, ah, interesting field of vision of these two worlds. And I come from the brand world, but I was exposed to the performance world as well. And so with Ready Set, it was like, you know what, there's no agency really that understands both sides, so let's try this. Oh, and by the way, vertical video at scale was a big need that we lucked into that clients like DoorDash hims and hers, you know, still clients today had they wanted to go from the Mad Men era of shooting, you know, three TV commercials in six months with 60 meetings and 600 grand to, like, I need like 20 ads this week. How do I do that? So it took me a long time to find a producer in LA who take me seriously about shooting like 50 ads in a day. We called them shared shoots. We'd shoot for multiple clients in the same day. And we were off and running from, from there. We just had our best quarter yet, uh, Q4 of last year.

Speaker A: Wow, Amazing. What kind of the, the, the, the director or the folks in LA who are actually doing that, what was it like actually finding that person? I mean, what was the pushback that you're getting from others?

Speaker B: I think it was just like the, well, but this isn't how it's done thing. You know, more recently and more dramatically, someone from our world of, like UGC Vertical Video Creation said, and we've done this now already said, like, even the word production, it connotes certain things. For example, if you say, well, this production has just gotten bigger, the client wants 10 more ads. The production mindset is, well, we'll need a PA on set to help with this and that. Well, who's going to feed the people? It's just very much in the paradigm of, like Hollywood production, when in reality it's content creation. It's like, how about you just grab a phone and point it at me or yourself or you know what I mean, and like, just get this thing done. Um, and so that was the gulf that we had to bridge. And at first, actually it was like what I call mini TV commercials. They were more conceptual, they were less native y feeling. And I'm like, all right, well, we're shooting 30 of them, you know, in a day instead of three in six months. Great. But they weren't performing in 20, 18, 19. And then we got on the tip of like, oh, actually, they don't need to look pretty, they don't need to be shot on a red. And we started Shooting with phones. And we started shooting at more, uh, scale and not worrying about the lighting being perfect.

Speaker A: The logistics of doing 30 videos in a day with three different clients. Like, what does that look like?

Speaker B: We shot 82 in a day once. And I mean, like multi shot conceptual commercials. Music's playing. And you see, I remember what this was for. Like, someone taps a guy on the shoulder at an office, which we had to rent. Like, tell them to come upstairs. There's the donut thing, the three people in the background, the guy who tapped and the other two. And he opens the donut box and it's all Brock Raw broccoli inside. And he's like, come on. And they like. It's like multiple shots. We shot 82 of those in a day. Three different crews, like 12 actors. I mean, it was. I would. I would have a speech to the crew before every day. It was more like an athletic event, I would explain, than a production. It was, we've got four minutes to shoot this thing, you know? All right, there's like, this person does a thing, and then we need a reverse, then we need a wide shot and then a close above the broccoli. All right, well, we got a minute each. Like, what do we need to do? Point the camera at the broccoli. We gotta go. Like, go. It was. It was fun.

Speaker A: Oh, uh, that's wild. That is a lot. 82 in a day. One of the things you mentioned earlier on was the difference between performance and brand. Right? And I talk about this a lot because I feel like everyone went through that transition. They went through the transition of we're brand marketing first, and then, shit, we need to do ugc. We need to feel more authentic. We need to feel like what consumers are used to seeing from their friends. Right? And I feel like it's almost coming back the other direction now. You know, we talk a lot about performance branding, which is these performance marketers, a lot of them don't even build a brand. I mean, hims and hers and those, you know, uh, Airbnb obviously have done a great job in building a brand, but there's so many that just lean into that UGC in the bottom of the funnel that don't build the brand. Do you feel like it's coming back full circle? And do you think that UGC has the same impact today as it did five, six years ago?

Speaker B: No, I think it's definitely been diluted five, six years ago. There were so few people doing it that you could clearly have more of an impact. I do think That B word was, uh, an anathema. Uh, in 2021, as the Facebook, like wave was surging through everyone, propelling everyone, oh my gosh, I'm getting three to one roas. Like, this is amazing. Why would I spend time like those old school mad Men people spending six months in a commercial. Now you go to the bar and like every commercial is for like a DTC company that's doing UGC ads. Like, they're all doing shiny commercials now for a lot of money. And it's in part because of what you're getting at. I think, though you said it. In general, things that look like ads are not going to convert so well. So I do think it's important that it does look like it's something from your friend that said hims and hers does a beautiful job with what we call a high low strategy. You go to their website, you see their commercials. They are more glossy and branded. That's important, uh, for trust, for any brand, but particularly for medication. But then like, all right, I've heard of them. I saw them at the bar. And now you're on your phone and it's like, now I want to see a real dude that looks like me talking about hair loss or ed or something. Oh, I see it comes in the box, will look like that. Oh, I see he takes the pill. Oh, he does it on his nightstand. Like, you're learning. You're, you're, you're being able to picture yourself getting this product. It's not about the glossy Rob Gronkowski joking around, you know?

Speaker A: Yeah, that makes sense. I love the thought process of hi low. Um, I feel like that hits, uh, it right on the head. You have to have both, right? You have to be authentic with the person, how they live their life. But then you also need to be aspirational, beautiful and trustworthy.

Speaker B: I think for most brands, you kind of do, particularly if it's in fintech or supplements or something. It's just like, do you want to trust a super trashy supplement brand where all you ever see is like, offers and things or one that you see like a million dollar ad on tv. Okay. These guys are legit, so I think that's important. That said, uh, there's a lot of discussion about collapsing the funnel into one ad. Right? It's Facebook ads, not some niche thing anymore. I mean, digital video ads on our phones, which we watch a lot more than the rectangles on our living room wall, are kind of just advertising now. And I think it is possible, you know, longboard people will watch long ad to collapse the funnel, um, to start sort of high level, social proof, trusty stuff, maybe some glossy stuff, maybe some celebrities. But then get down to the point, you know, call to action the offer. So I don't know if in 10 years we're going to be talking about the funnel as this like high funnel thing with brand tv and then you've got your Facebook ads that close the deal. I feel like it's all smushing together as video just takes over all these screens at once.

Speaker A: Yeah, well, uh, it's like we're going to go back to infomercials, right? You watch it for 20 minutes and they go over the thing and they.

Speaker B: Infomercials got a bad name because they're for cheesy products. But like, if you're struggling with ed, maybe you want to watch seven Minutes to like, understand it. I was just looking through one client of ours that does like a, uh, Ozempic and Trisyprotide. There's these new things that are like Ozempic and appeal. And some of their top performing ads are just educational. Literally. It's just a guy going, like, I'm saying it wrong, but it's like, this is Tirzepatite and he's holding up like, oh, I see. Tirzepatite's actually a needle thing. Okay, I want to learn more. I don't know if I'm going to sit there for seven minutes. But like, my point is a lot of what people want is new information. It's actually something we're trying to do with airpost video ads is get away from genericisms. Like, this is a game changer. My health is so much better. I've lost weight and feel great. Like I'm not learning anything. But saying, you know, 84% of people lose 10.3 pounds their first month. Or like, this is terzepatide. Let me explain what it is now. I'm leaning in and I'm going to learn something.

Speaker A: Yeah, I actually, I think it's a really interesting point around the idea of collapsing the funnel like that. And we've seen the same thing as the education ads because it used to be that if you could get a UGC video and someone would say this is the best thing ever. Right? But now consumers know that everyone is being paid to say this is the best thing ever. And so they're asking themselves, why? What, what about this product makes it the best thing ever? And why should I trust what you say versus what someone else is Saying about something el else. Right. And so you kind of, if, if you can get that, you know, user or character or however you want to say and bring forward a trustworthy Persona from them and then educate the consumer about the product, then they're actually going to trust what they have to say. Because now they've learned a lot about the product and it comes from someone who seems trustworthy. Even if they know they're paid, it feels right. Right. You're going to be more likely to test it versus, you know, a fake. This is the best thing ever. You know, have you tried this?

Speaker B: Those you can just smell a mile away. And that's why the other word that is important to these ads is authenticity, which is ironic of course, because, right, they are paid or maybe they're AI or this word is getting coming up a lot for us. But it's true. Somebody saying these genericisms, which again, some of our ads still have loaded into them versus where airpost quality is not yet is like a minute and 12 second video ad. A woman comes on, she looks real, she's got the product there or not. Maybe she doesn't mention the product for 30 seconds and she's like, oh my gosh, my sister is impossible to buy gifts for. I got her that, you know, air fear fire thing, I forget what it's called last year and she got something specific about that. And you're just watching this woman thinking this is a real story, like, right. And ideally it is a real story from that creator. And then they're actually talking about how this thing with their sister's rosacea and they can't talk, blah, blah, blah, blah, blah. Somehow it ends up on curology a minute later. But you're just watching the story like an organic piece of content and you trust that person because it seems real. Compare that to like, you need to try this. It's a game changer for your skin. It's like, don't tell me what to do, like flip.

Speaker A: Yeah, yeah, exactly, exactly. Yeah. There's gotta be that, that level of authenticity to it. I mean, going to airpost, right? So we've tested a bunch of different AI products. They keep getting better, no doubt. I'm curious what differentiates what you guys are bringing to the table now. And then I want to kind of dig more into the AI conversation in general, but. But what, what differentiates your product?

Speaker B: We're going higher end. We're uh, about to blow up our website and do this because it's really working for us. We had some big Companies come to us that want to cynically check the big AI box. Right. The board, the CEOs. Like, we, we know that AI is taking over marketing and advertising. Are we using a growth team? No, not really. Because there's these toys out there, but they kind of suck. Okay, well, we still, you know, let's keep pushing. And then they come to us and we have some credibility on this because we baked a lot of AI into what we do at Ready Set as well. And we've been building this thing for a long time. So I think the takeaway from Airpost and where we're headed is high quality. Yeah, it's really just high quality video ads because honestly, I don't think that exists out there today. You and I could spend two hours on the Internet trying every piece of software to make a video ad and give it to your clients at Playbook. And my guess is 19 out of 20 they would vomit on and say, brian, like, what are you doing? We can't run this. And the 1 out of 20 that would run would probably not work. So our goal is to get them to be indistinguishable from manually produced ads. And once we can do that, I think it's a really big company. We've got a notion doc called 99 Problems. That's really thoughtful. List of big and small things that make up the delta between where our ads are now, you know, and if I show the ad, our ads to you or some people, especially less sophisticated folks, like, oh, yeah, those look like Instagram ads. But an actual growth leader who's got to put their reputation on the line and run that thing, it ain't there yet.

Speaker A: Yeah.

Speaker B: So we've got some humans in the loop right now for these big customers that are really closing that last five miles of AI that we can talk about. Yeah. So it's all about quality for us, not about doing 16 things at once. Analytics, photos. Like, we're really focused on video ads at scale.

Speaker A: Yeah, yeah, no, I think it's smart to keep your, keep your focus tight. And this is what we're building, obviously. It seems to me that you guys are in a position to take, take a leadership position in this just because of your experience with Ready Set. And what I've seen from the product so far looks, you know, fascinating from a underlying, uh, technology piece. Do you feel like the gap that needs to be covered is going to get filled with new models, or do you think it's something that you guys are doing internally to train the data

Speaker B: or adjust the process, it's definitely both, you know, as the models get better and enables new things. So you know, it's like um, OpenAI had a speaker on saying recently, if you feel like man, our AI startup is like so close to being able to do this thing, but it's just not good enough yet. Just stay the course, wait six months because the models will help push you over the edge. So that is true. Those tailwinds have felt really good, um, and feel like they're getting stronger. I feel like they announced two things a week, uh, and we use a lot of models but OpenAI pick one. Uh, and the other thing you said is very true. I actually feel like in this crazy fast changing world of AI startups right now, the ones that are best positioned in my mind self servingly maybe are like niche app layer startups that have a lot of expertise in that area. Like we're the only ones really looking around that actually not just like made a toy that makes us video ads. Like, oh, we think you'll like this. We sort of come from an adjacent space, but this is what we've been doing for years. This is what I've been doing my whole life is making video ads. And so, and we know what we need to do to get there. Some of it is model driven, like can the model write a good script? Some of it is more fine tuning or hard coding, certain things, colloquialisms and that kind of thing to get it to, to that level. And it's super fun Brian, because there's so many creative ways to like cross these streams and pull up the bridge behind us.

Speaker A: Yeah, no, I bet it is. I mean I just like digging into the stuff that's going on with like vibe marketing and prompt engineering and all. Like it almost feels like there's infinite ways that uh, you could adjust things, right?

Speaker B: And so I feel like it's gotta be fun. It's like I've fallen victim to it. Like I like pounded the table a couple months ago was like lovable is amazing. You know, like it's a cursor, you know, like make your own website, whatever. I spent all morning and I made a whole website and it's better than our current website. Let's just make it live and go. And that way if I want to try a new header, I can just literally say on the left side of the pane, hey, try this, put this picture here and it'll update to our site. And the team came back like we love it too. And we all made sites and at that time you can't even have more than one user. And it doesn't really hook into a framer or webflow or something very well. And this, this, this, this, this. Even people that are proponents of these cursor lovable tools that I've talked to, even a CTO of Stripe who's an investor in Airpost, uh, like they're like, yeah, I love these tools too, but they're not there. So you got an infinite amount of demoware, I call the demo apocalypse on Twitter. And you're like, ooh, that one, that one. Did you see this startup? And if you look at the cherry picked video they make, often it's one they make with like uh, a designer with a bunch of Figma files. It's like, holy crap, that thing's gotta be awesome. But you've probably had the same experience I have, which is when you scratch a service, you use the product. All right, let's use it for making our website. Let's use it for making this or that. After like an hour or a week, you're like, yeah, this is not better than what we do now. So that's the problem. Deep instead of wide.

Speaker A: Yeah. There's a part of me that you know, wants to not work and just tinker with AI all day long. But it's like, but I actually only have 30 minutes of window where I start messing with something like that. I'm like, oh, this is interesting. But I have no time. Moving on, you know, but sometimes take

Speaker B: longer to figure out than, than they are worth. Yeah. Or like in our world, the video ads, back to that world. It's like I was talking to my friend at a big company, he's like, yeah, we're using this other tool, it sort of works. But our editor can do this much faster in Premiere. Like they have editors, you know what I mean? Often they have editors in Colombia or Argentina or the Philippines. Like they're great deal, they're happy, they're cranking out ads every day. And like, you got to do better than that if you're going to be in this game.

Speaker A: Yeah, well, and that's the thing is that I look at AI as a cost reduction. Right. I mean that's essentially what it is

Speaker B: because now that's mainly what it is.

Speaker A: Almost anything out there, people can do. I mean we're just, people can do this work. Right. And so the thought is, is either you can reduce the cost of what you're selling now or you can do more with less. So you reduce the Variable cost scale. Right. Those are your sort of two reasons why you want to use AI for, for anything. And so in that, right, to your point, there are great video editors that you can get overseas for, you know, 5, 10, 15 an hour. Right. And so, you know, your AI has to be better than them. And then the cost of production, though, you know, it's fairly high. But to your point, I mean, it's dropped a lot with iPhones and all of those different things, you know.

Speaker B: Yeah, we're trying to, it's exact. Right. We're trying to replace or force multiply the. All three parts of the process. Ideation, you know, because even with that editor, you. Someone's got to be looking at a blank screen. You know, the white bull, like Hemingway called it, like, okay, what should we do? That's a big stressful question for a lot of people that need 30 ads a week. Ideation production can be expensive and time consuming. And post production, um, and also we're not doing one by one. So AirPost is like a wall of an instant gallery, we call it, of these tiles that you can scrub through of finished ads that you can control, tweak, shape, the brief shape, the ones that are coming in every day, rather than like, oh, forget Premiere and your footage library that your editor can just do as well and sit here and click through a bunch of questions and chat with an AI for three hours to sort of get a draft of an ad and wait for it to render. I think that model's more for the lower end of the market. And we're trying to like, really solve the problem for big folks at scale and then to take your binary like cost cutting and. Or force multiplying what you can do for the same cost. I get more excited about the second one. I think you're right. I think the first one's probably the most applicable now. But like, how many big companies do you meet that they're like, we would love to run ads at every different Persona consistently and test every value prop for every Persona and do every carousel and image and video, but they just can't. Even big companies, if you could think of press the magic button and be able to enable that, that's another thing we're looking at.

Speaker A: Yeah, I mean, a hundred percent. I mean, the ability to create so much at scale at once. Ideate, get it all out there. I mean, it's, it's, it's really difficult now. So, I mean, I think it's going to be a huge, huge game changer and I think you're thinking about it the right way. What do you think about. So I'm sure, I'm sure you saw Google's new Release. Was it VO3 3. Right. Uh, I've been, I went down the rabbit hole last night on Twitter just looking at these videos and saying holy shit, right? What? This seems like it's kind of the chat GPT moment for video in some ways. How do you see that fitting into your roadmap and do you see it as a risk?

Speaker B: I. Not really because I hold the again, maybe self serving view that there will always be companies tools that big brands will need to rely on to solve certain problems. I think if you view it as an existential risk, what you're really saying is, and it's very easy to fall into this conversation that I have many times with people, well, but aren't you going to just be able to say make me a perfect ad for hims with uh, this type of person and it's just going to work. Maybe, but there's so many things around that first of all, like I guess what I'm saying is if we get to a world, Brian, where you can say make me the Matrix whatever four starring puppies and it's actually funny and uses John Belushi kind of snle humor and it works and you're like, wow, that was a good movie and you could make any video then I'll die in peace. It's okay. It's like then, then great. Like the world will change it. We'll all be doing something else by then and we'll start different. It's fine. I kind of don't see that happening. I kind of see Nike still being like, we really need distribution and we need to advertise this new shoe and our director of marketing and wherever doesn't really want to sit there and type make me a this and that over and over again. We need something to depend on or at least help us organize that and make it at scale.

Speaker A: The Catching Fire podcast is brought to you by growtal. Growtal is a network of, uh, freelance marketing professionals for all of your marketing needs. If you're looking to build your own marketing team, you can do it through grow tell, starting with a fractional CMO and then building out all of the team members, including your paid social, paid search experts, your SEO experts and everyone in between. Go to growtel.com www.gr o w t a l.com and get started today. I tend to agree. And you know there's always going to be A human in the loop. Some way, shape or form, right? You have to. Someone's got to come in, actually take the action to get it started. And then there's a QA and there's iteration and there's a bunch of pieces in there, right? And in my experience so far, using any of the, uh, you know, LLMs out of the gates and just typing it in, it feels laborious. It feels like, oh, I have to do this over and over again. I can never quite get what I want. And so it seems to me that, uh, a company like airpost, like, the biggest value that you can bring to the table is an analysis and understanding of their business and their goals. And then the prompt engineering in the background to bring out the quality and the consistency that's needed for these companies to have success in creating the assets from an LLM. Right? You get someone who just doesn't really know how to do it and they're going to kind of be. They're going to get crap out all the time. But if they have someone like you to hold their hand, that's much better than probably finding a prompt engineer who's going to figure out how to do it themselves. Right?

Speaker B: I'm sure it's called their brand and how this new shoe inventory just ran out in Indonesia, so they need to pause down this one, but they want to try that while they wait to go up for the big season. Like, there's so many variables and things. I think that's a very smart analysis. That's exactly how I see it.

Speaker A: Yeah. Yeah, it's, uh. I was thinking about this last night and just conceptually, right, uh, at the start of the Internet, right, everybody was putting in HTML and then HTML turns into Java and turns into Python and P and, you know, whatever, whatever, whatever. I almost feel like prompt engineering could follow a similar format, right? Where right now we're writing basic prompts into different types of LLMs. The complexity of those prompts and the accuracy of those prompts and the amount of data that could go in and training and all these other pieces, it almost opens an entirely new set of coding in some ways because there's an infinite amount of ways to prompt these things and to get what you want on the other end. And so that's going to get more and more defined over time. Feels like prompt engineering is going to be the new, you know, computer science or, you know.

Speaker B: Yeah, the wind surf that just got bought on the Cursor competitor for 3 billion by OpenAI was saying, you know, he's a hardcore software engineer. He said, I don't know if in 10 years we'll have software engineers, we'll have this group of people called builders. And if you're sitting there and you're like, you know, I keep getting this a little bit wrong and it's giving me that, and you're like, I wish there were sort of a QC agent that could look at all my outputs. And before it shows it to me, it runs it by these four things that I keep snagging on and redoes it for me that way. And you could just go to a Windserf or OpenAI whatever and say, build that agent for me. Talk to this guy. Talk to my regular ChatGPT instance. And it does all the hookups. And this is probably two years away, not ten. So, yeah, it's. We just had a call with our management team this morning at AirPost about. I saw Sam Altman say something a bit contrarian, but it actually makes a lot of sense to me. We were discussing it. He said, you know, a lot of companies, uh, that are like, this is what we're going to be in five years or 10 years and work backwards from there and then boom, boom, boom, which I read the book, Working backwards from the Amazon guys, and I get it. He said, I don't really see that working historically. And then I'm running with his comment, like we were talking about in the AI age. Everything's going a million miles an hour. Jason Lemkin, the keynote at Saster the other day, was saying his message to the crowd was tough love. He's like, you gotta move so much faster. He said, 2009 and 2023 software didn't need to really go very fast. Like, oh, we're adding SoC2. We're doing some permissioning thing. SaaS software was relatively staid, and now it's like, there's two new models, 3s a week, and you gotta go so, so quickly. So Sam Allman's M thing is like, we just kind of look forward and we're like, all right, what do we want to do this next six months? Well, this new model's out. We gotta do this. That, uh. And that's kind of how we're looking at building quality in video.

Speaker A: Yeah, Yeah. I think I saw, uh, another quote from him that talked about businesses that are trying to build A.I. uh, now to solve incredibly complex use cases, are gonna spend a bunch of time and money doing that, and they're likely to be unsuccessful because it doesn't support them. Now and what you want to do is you want to build for now, what can it actually support so you can get something live and with your customers, something that actually works in the market. But then you need to ride the wave of every model. So as you're, as the models get better, you need to improve your product or extend it into new domains that are, that are parallel and continue building upon it with the models and not try to force the model to do something it's incapable of doing.

Speaker B: This is exactly how we think of it. It's catching those waves. Like there was this one really high value thing that we were trying to figure out recently and I found this startup in China that could do it and uh, had to do with like stripping away overlays and things like that automatically and painting them in. And I was like, whoa, this, they're doing it. Where are the other APIs? Like, let's all research this. The engineers looked at it and there's some stuff out there, but even the good stuff doesn't get to like the 95% quality you need to. It's not there. The models aren't there, the APIs aren't there. And so I'm like, you know what, the waves were sitting in the water on our surfboards. Like we just got to pass on this until someone makes an advance and the waves coming and then we'll ride it. Otherwise you just struggle in the water and waste time when other people are off building other things that, that are ripe and then they're going to catch the wave when it comes.

Speaker A: Yeah, yeah, 100%. You got to, you have to choose your battles, um, and determine what's going to actually work. Now if you, when you're looking at the marketing space now, right, you guys, I mean, obviously do video makes sense that that's going to be where you're focused. What other areas are you seeing? AI touch and marketing. And uh, is there anything that you see that's really exciting?

Speaker B: Video is interesting. It's funny to me when people are like, oh well, video, look at, look at this clip here. Like it's over. I think about text, Brian. Like, text has been out since 2022. I was playing with GPT3, you know, long before ChatGPT, and I've watched this whole movie. It's been years now. Not that many years, but a few. And the fact that most people today are not like writing hundreds of blog posts that are SEO perfect for their website with AI is telling to me that even with text, which has gotten really good, the Quality is still not quite there. My co founder and ready set Sam made an accidental, really astute observation, which is if you're a novice, this stuff seems amazing. If you and I look up, like, tell us more about angles of entry into the unit, you know, galaxy or whatever, the atmosphere with a rocket and how does it work and what does the person have to do, we would be fascinated and we'd assume it's all right. And it's like, wow, AI is amazing. Meanwhile, if you went on chat GPT right now and we're like, what's the best way to structure a Facebook account? You know, for this client of mine, probably you're not going to be that impressed. Like it might repeat what you know and then, oh, that's interesting. It knows a lot of what I know, but it's not going to blow you, uh, away. So it's like, wow, it's still not there with SEO. So what's happening now? And there are companies like GrowthX, AI and others that are like, look, even with text, even though it's years ahead of video, it's still not good enough for experts that really know your space. And you know, we're only going to read good content that actually is additive. So here's what we're doing. We're hiring last week Darrell Etherington, who's been a tech crunch forever. Like, we're hiring a less reporters and then really know how to punch up stuff. And we're using AI for 80% of it and we're using people and experts to punch it up because AI is not good enough. They say right on their website. People are too slow. So we're putting them together. And there's a lot of controversy around that right now. A lot of VC funds are general catalysts and others are spending billions putting money into like literal accounting firms and law firms going right to the outcomes, ripping out their manual backends and putting in AI, uh, and just upping the gross margins massively. And so are those startups? No. How much do we care about monikers? Right? Uh, like Sam's always saying, like, look, as long as your gross margins are 80, 90% software level margins, who cares? Someone calls you SaaS or agency or hybrid or whatever, build a great business. But that's kind of what we're experiencing now. And it's a little bit of a reaction to what you alluded to, which is like, oh, we're doing stuff with AI, but it's mostly unripe bananas. And then people are starting to realize, I'm um, not actually using any of this stuff yet. So people are realizing I can solve that last five miles of AI by carrying this product on my back with good people to the finish line.

Speaker A: Human in the loop, man, it's hard to get rid of it. They've been saying they're going to remove people for a long time and it turns out we do add value, uh, over technology. So I hope that continues.

Speaker B: And the vibes have shifted where it's like, oh, maybe it's not a bad thing anymore. If I, if I yelled the word service revenue at a Saster conference 10 years ago, people would be like, oh, don't be near that guy now. It's like cool to have humans in the loop and, you know, have people and charge for service revenue and things like that.

Speaker A: Yeah, yeah. I, uh, I think pure services is tough. I've done a lot of that in my life. I do respect having a lot of product revenue and better margin than pure services, so. But I think, yeah, great technology with a level of service to, you know, just put that extra shine on it is absolutely a great model.

Speaker B: I mean, it's amazing. Like HubSpot, you know, Salesforce, these pure SaaS companies. You'd think. How many times have you gotten an email from some not even junior level HubSpot person like, hey, do you want to talk? You know, I've got a bunch of time to look at your account for an hour. Like that's called service, you know. So it's, it's interesting. I think this hybrid model is going to settle in for a little while until pure AI gets good enough.

Speaker A: So obviously, you know, with Ready set, you guys came up in the world of direct to consumer E commerce, right? And there was so much money flowing in from VCs at that time. And over the last few years we've really seen that space slow down and there hasn't been as much investment. I think, you know, the ones that won are still out there really pressing and my guess is they're a good customer base for you. But I'm curious, what do you see within the world of direct to consumer and what do you think the next big growth levers are going to be from a customer base for a product like yours?

Speaker B: Yeah, Direct to consumer went from like fake easy in 2021, where, you know, tons of funding flowing in, lots of growth in revenue, not a lot of profits, and that was just being sort of shoved aside to really hard with ATT hitting, which is, interestingly might be getting rolled back in Europe now. And, and the business model sort of, uh, um, you know, the emperor has no clothes kind of thing on a lot of business models. I had somebody at a big PE firm tell me who left that firm that invested in mostly famous TTC companies. Tell me one and a half names out of the like 30 plus that they invested in mostly. Again, well known ones that we both know that were profitable like, and this was like 20, 23, 24 out of everybody. It's really hard. I, I saw a really good um, analysis this guy did. I don't know if it's the answer or not, but he was saying a lot of it is just the shipping. It's the last mile of getting to the door. You know, it's why a lot of D2C brands are like, well, maybe shipping a giant pallet to the Costco regional center of the Northwest is like actually a lot less expensive for us to do than getting each individual package to each individual house. And this guy was saying that, look, you know, drones, if they can make them quieter or other robo, you know, waymo delivery cars, like if the cost of getting a package to your house can go down by 3/4 over the next 20 years, that might unlock D2C like crazy right now. It's a tough slog and you said it. Well, there have been some winners that have consolidated everyone else. Those are our strongest clients. But yeah, ready sets, mostly. Talk about ready set, ready sets, mostly consumer direct to consumer businesses and we crush that.

Speaker A: Yeah, no, and I, it's, I think it was a great wave to ride. We sort of rode the same wave coming out of Meta when we launched our, our business, but now we're starting to see growth from other angles. SaaS is interesting, fintech and professional services and it just feels like there are people looking to innovate in, in different areas. I'm curious, where do you think the next area growth is, is going to come from and what interests you now? Coming out of kind of the VC world.

Speaker B: I mean, speaking specifically to ReadySet, we've got so well known for creative and luckily, you know, we've been doing that since 2018, 19 now it's almost conventional wisdom, which it wasn't then, that creative is the biggest lever. So there's not that many agencies that are known as a creative specialist. So now what we've been doing, where our growth's been coming from is adding media, building out a platform for people to just grow generally. And then I kind of want to turn the question back on you. Like I, we Never did much with B2B. My friend Daniel runs an agency called Bamboo or he left recently, but it's his agency and they're mostly B2B. Uh, I know you guys do some B2B. Like what do you see as the distinctions in serving a client that's B2B or SaaS versus like a D2C, you know, supplement company?

Speaker A: Yeah, Well, I think what's really interesting is you're seeing a bit of a blurred lines between what was a traditional B2B funnel versus say direct to consumer. Right. Used to be that, uh, you had a very strong content forward approach, which I think you still kind of need. But now the formats that people are learning in is starting to become more video centric than it was. Hey, I'm going to go to a webinar and I'm going to go read about it and I'm going to go talk to the salespeople and put it through. Procurement is that they are starting to look to say who has, who's an authority in the space and what are they saying? What do I want to learn from them and then what is that applicable to my business? And so, you know, one of the things you're seeing a lot on LinkedIn is just a ton more thought leadership from CEOs, but also people who are in more of a, uh, kind of, I call them sort of they're pseudo influencers because they're, they're operators and they, they do things in companies. But then at the same time they're always talking about their trade. And I think now companies are turning and they're looking to who has authority, who's talking about this, what can I learn and what can I put into my business? And so you're seeing a little bit more of that consumer type creative work in the B2B world. And LinkedIn is one of the biggest places that's leading that because there's so much content that's being, you know, produced and consumed on that channel and a feed format. And so that feels a little bit more like Facebook or Instagram, just more depth to it than the standard kind of video you might see on general.

Speaker B: Is that an area agencies can play or do you run into a lot of people saying like, no, no, no, we're, we're good, we're gonna just have our CEO post a lot and try to boost some posts.

Speaker A: Uh, you know, we have some clients where we're doing a lot of content for them, putting out a lot of content on LinkedIn and helping with their ad Content and producing more UGC style things and educational pieces. And here I'm, here's how I'm using this software in my startup and pieces like that. Um, and I think you're seeing a lot more kind of community, uh, led thinking versus individuals within companies with highly specialized knowledge, which is pretty interesting to see.

Speaker B: That's really cool. That makes sense.

Speaker A: Yeah. So I think that, you know, I guess for you guys, right, my suggestion would be get some of those UGCs. You know, it's the same thing as someone's in there talking about a product, right. And they're kind of, instead, they're demoing the product and here's what, how I used it in my startup, et cetera, et cetera. Right.

Speaker B: You see a lot of those like, fast motion, like going around software products. You know, comment this to get, you know, access to my.

Speaker A: Exactly. Comment agent. Uh, and I'll show you how I built this custom agent for your company work.

Speaker B: I see CMOs of big companies saying agent and I'm like, wow, okay.

Speaker A: Yeah, exactly. Um, everyone's curious to learn the next thing. It's moving so fast. It's so hard to keep up. Right. And so I feel like that's, that's one of the greatest promises that a B2B company can give you is, hey, we have the next thing. We're, we're there and we're going to help you evolve so you don't fall behind.

Speaker B: No, there's so much curiosity. No one wants to miss out on what's next. Everyone feels like they're missing out on the next big AI thing they're supposed to be doing. And I tell people, don't worry, you're not missing out on much. Everyone's using ChatGPT. Get good at that, keep up with what's going on. But there's no secret DTC brand that's like pressing a magic button and letting their whole growth team go and having everything automated. It's, it's still a couple years away, if that.

Speaker A: Uh, yeah, yeah, I would agree with that. Well, I really appreciate you coming on the show today, John. Uh, before you go, any last words of advice to founders looking to unlock growth?

Speaker B: I would just say off of that last point, you might feel like you're behind. You might feel like you're not up to date on the latest AI stuff or you're doing something wrong or you should be doing this or that or why did that pilot work for me that the Twitter demo looks so cool that all my friends were sending me don't feel alone. It's what Brian said earlier. It's like, all these things look awesome, but you play with them. And they mostly aren't used a great term like ready for market yet. They're not in market tools you can actually rely on. That said, it is true what they say, marketers who actually understand and are up to date on AI, when that big wave comes and the stuff actually does start working, they're going to be the most successful people in this business. So I just tell people, keep reading as much as you can. Try to find areas of AI you're passionate about, but don't feel like you're the only one that's, you know, scratching every surface and having things fall apart. Most things are falling apart right now. And, um, yeah, that's why I think we're entering this era of like, a hybrid humans in the loop with software. And that's what we're trying to build at AirPost.

Speaker A: I think that's. That's totally right. And I would say just keep scratching the surface, though, right? Don't put your head in the stand and say, oh, I don't know, I think, uh, it's good. It is moving fast and so it's important to keep up.

Speaker B: But don't get yourself saying, oh, but look at this and look at that problem. It's like, yeah, we all know and it's the worst it will ever be. And you should have seen it a year ago. And ChatGPT came out literally just two and a half years ago and the whole world has already changed. So, yeah, don't catch yourself being a laggard either.

Speaker A: Yeah, yeah. I can't believe it's only been two and a half years. It's crazy how much it's. Anyhow, well, thanks so much for being on the show and for our listeners, if they're looking to find you, where should they look?

Speaker B: Yeah, I'm on LinkedIn. John Gargiulo or Airpost AI, if you've got, uh, uh, a brand with some scale and you're looking to get video AI and people made ads that actually you're going to, like, come check us out, give us feedback.

Speaker A: Sounds good. All right, thanks so much, John.

Speaker B: Thank you, Brian.