How Great Brands Scale Beyond the Operator (#CollabTalk Podcast Ep.192)

[music] Hello and welcome to another episode of the Collab Talk podcast, the show dedicated to exploring how and why we collaborate. My guest today is Edwin Choy, founder and CEO of Jet Fuel Agency, a growth marketing agency that helps e-commerce and omni channel brands accelerate revenue through paid media, creative strategy, conversion optimization, and customer acquisition systems. And I'm excited to get in to talk about how your world is changing. We'll get to that, too. But welcome, Edwin. >> Thank you for having me. I'm super excited for the conversation and appreciative uh to get a chance to talk to you. Yeah, I think there's a lot to jump into around this this topic and and talk about with your company, but today we're exploring a challenge faced by pretty much every successful founder. So, how do you transfer transfer the instincts, your judgment, the decision- making that help build the business into a team that can continue scaling without being involved in every decision? This is that problem that every uh startup owner operator experiences as growth happens. Uh and and especially how do you do these things in this era where AI is transforming? How a lot of this work gets done? What becomes uh even more important to teach to our teams? What do we need to preserve about our culture that to as it grows and scales that organization? So Edwin, I know there's a lot that's a mouthful. Um before we dive into the topic, maybe tell us a little bit more about your journey. What led you from being an e-commerce operator to launching Jet Fuel and what problems are you trying to solve and what are you seeing with the companies that you're working with? >> Yeah, absolutely. So, I'll give like a little bit of background to myself and my journey and then we'll chat about you know some of the problems that we're s we're facing and we're also trying to solve for. So just first of all I've been involved in ecom direct marketing since 2009. So I've been doing this for what feels like forever you know basically 17 years now which is so crazy to say aloud. In 17 years you have seen several cycles go through right like you have seen SEO rise. You have seen people say SEO is dead. You have seen paid media. Now you have seen the proliferation of AI. So I I felt like I lived through three or four of these huge macro seismic shifts. Right? And then when I started, you know, the world was relatively simpler. Let's call it that, right? Like when I started in 2009 to 2011, you know, just fresh out of college, went into an agency, started my SEO and my paid search career, and I was kind of learning the ropes there. Uh transitioned to e-commerce operator, super rapidly uh scaling company. We scaled from five million a year to 30 million a year in four years. Uh, and at that point, you know, I pretty much touched every part of the machine, right? From affiliate marketing, marketplaces, email retention. Uh, we were like one of the first 10 advertisers on Meta Ads when they first were released in 2015. Uh, so a lot of that DNA foundation was built there for myself, right? And then, as you can imagine, if you're growing a company that quickly, every day is a new problem. [laughter] Right? Like every day is a new problem. Sometimes it's problems that you've never seen before. Sometimes it's problems you've seen before. But you essentially build this war chest of problems and solutions and you start to build this mental playbook like oh holy moly uh we lost our billing for the day like what's the playbook to get it reconnected and scale back up to where we were so we lose as little momentum as possible. Right? So I feel like in everyone's brains there's there's a database of these situations that come up and then generally the more years you spend at a company and the more problems you're exposed to the greater your database in your brain right uh so in 2017 you know I I left the company left on good terms and I intended to take a six-month sbatical right u that sbatical lasted about six days unfortunately [laughter] so I had just a lot of companies reach out to me say, "Hey, I heard you left. We really need some help here." And that's where Jetfield Agency was born, right? And Jetfield Agency is was born out of several different challenges I had to myself, right? I think one thing I wanted to ask myself was did I catch lightning in a bottle or is this something that could be replicated against many many companies? Right? That's the first challenge. The second is the agency is just a personification of what I like to do on a day-to-day basis. Right? Like if you were to not pay me at all and I just work for free, this is exactly the type of work that I would still be doing, right? Growth marketing, solving growth issues, um, and really, you know, building the infrastructure and the scale for e-commerce companies to get them from point A to point B, right? I find that just immensely from a creative perspective, uh, from a challenge perspective, I find that immensely rewarding. Right. And last but not least, I think I've I just had so much pleasure and so much satisfaction in coaching and mentoring a group of individuals at my prior position. And I was thinking to myself, well, like, could could I build another strong team like that, right? And that kind of wound up us launching JetFo agency. So 2017 we've been business for nine years. We had the pleasure of working with uh some of the fastest growing brands in terms of direct consumer uh within the last nine years. And that really kind of brings us to this topic, right? Because this topic is a problem that we've been chatting about pretty much for the last nine years. And the question is, you know, if you had someone come on board, let's say they're a brand new client, and obviously their their preference is I want your entire team to be just the senior leaders, >> right? [clears throat] They want all the best senior operators for every single, you know, one of their campaigns. And there's it's a fair ask, right? They ask that because those senior operators, they have the playbook in their heads. They're they're battle tested, right? they just been through so many scenarios and issues and problems that that they're able to have an answer for a lot of these scenarios that come up. They have the internal benchmarks, you know, within their brain. And a challenge for our agency, any service- based business, really any business in general, is how can you scale up that institutional knowledge? So employee A who could be the most tenured employee at your organization to employee Z who could be the newest individual at that organization how do you basically make it so the level of quality of work and the sophistication is nearly exactly the same that's the dream that we all want right so it has always been a challenge right so before AI I think we've done things in a very manual basis and we try to do it the best we could we even done things like worksheets, right? So, we would have common client problems placed into almost like homework sheets or homework assignments where you answer questions and then in doing so the team would really learn like oh if this problem comes across these are the data points I pull these are some common answers to these issues. Um but obviously as you know the a a problem with that in the business world and digital marketing in general that thing changes every three months >> constantly. Well, yeah. I mean, if you go back 25, 30 years, it was, you know, the call tree kind of thing. So, I remember building a a chat bot back in 2005 for a client where that's all it was was it was a a script that was a call tree. It was a a a database of answers that they' go and search and pull up there. And so, making those things more intelligent because you're right, it it changes very quickly. um depends on the industry uh you know of the the client. So you've got to be ready to to go with those. That's where a lot of I think the initial thought with with AI is that we'll just put all this information, all this knowledge into the database and the AI will just regurgitate that. Somebody calls in and we don't even have to answer the phone and the AI will do it and uh which can do a certain amount of that that kind of automation but it's you're right it's incredibly difficult to to keep up with that. >> Absolutely. And we had other manual processes too. Another common one is role playing. Like I would pretend to be a client and essentially come up with problems that the team would have to address on the fly and we would do that for certain individuals. We would do that weekly and we would just battle test them, right? Like I would be interrogating them and I I would preface this before the you know before I talk. Hey, I'm doing this to learn like I'm going to just amp up the pressure. So then when a client comes to you with that same level of pressure, like you're you're already battle hardened, right? Like you've already kind of faced this, right? >> So we had all these techniques that we were doing with varying levels of success, right? Like I wouldn't say we really hit the holy grail. We had a process where every single person became the strategic mastermind individual. But I thought it was fairly successful, right? Because individuals come to us and we'll say hey with this level of coaching and preemptiveness I feel quite equipped for the challenges of business and for you know being a true advisor to our clients right so that let's call that the before times right like we would do things like that now in the before times I think a lot of those things still hold true today we're still doing a lot of that right so it's not like we abandoned it but I think something that's really crazy now that we have never had before is the ability to use AI to really supercharge that. Right? So, this is a huge cultural pillar of the company that we basically repeat adnauseium. Whatever we're building in terms of AI infrastructure is to supercharge your human ability, right? We're not looking to replace individuals. We're not looking to take your work away. We are looking to take certain things away so you can go in and use your unique human abilities and superpowers and spend the majority of your day-to-day on that. Right? That is the intent. And in one way that we're doing this to help transfer some of these instincts and playbooks and the wealth of experience we have is building what we call centralized brand. Right? So uh our agency is called Jetfield Agency. We have built something called Jetfield HQ or Jetfield headquarters. And really the goal and vision of this is that we want it so institutional knowledge across employee A all the way to employee Z is constantly building being stored and compounding upon itself pretty much on a day-to-day real-time basis. Right. Um, before I kind of get going on describing what this looks like and how we're using it, I'm I'll do a quick kind of just ask like are you familiar with things like MCPS and skills and Okay. All right. Very cool. So, if I had to draw it out, you know, we have HQ as as the hub and we have a few categories of data connectors that we have for it. One is called the uh the learnings database. So basically what this does is that it'll go into every meeting transcript historically, every Slack message, every email, every piece of communication that it can harvest and it'll essentially build uh a collection of observations and it'll automatically categorize it. Right? So it'll say something like, hey, this is an observation I made. Uh this is a strategic test that's been running for a few clients. This is our hypothesis and this is our finding, >> right? So, we're constantly building, mining, and cleaning up this database uh of tests and things that we're running across the many years and many hundreds thousands of experiments and campaigns that we've been running across clients, right? So, this level of institutional knowledge used to be so incredibly difficult to put together and build, right? Like, so let's call it the before times. I would have to do postmortems with individual teams and say, "Okay, what what do we test last month? What do we learn? What was the results and why did this happen?" So all that is now being done automatically, right? And the beauty of this is that someone, let's say in theory, I just hired an individual and they've only been here a month and they know what we stand for culturally. >> Um, they know what we stand for from a DNA perspective, right? And I'll talk more about the company DNA document that we're building that's going to help hopefully uh turbocharge this. But now that being brought into this system where they can go in and sign a new client and they can cook up a strategy from this database of prior strategies that's incredibly sophisticated, >> right? They could be subject matter experts, you know, in a in a fraction of the time. So instead of and then I I I say this all the time to a team you know I've been doing this for you know 15 plus years I have faced many different types of you know issues and problems but even if you put me in front of a client I personally would not have the breath of knowledge and institutional uh learnings across the entire team combined like I personally could not even match up to that >> right >> right so even I would and I say this to to my team even I would benefit from such a database where I can cook up something like, oh, I've never really heard of this scenario before, but let's see if I can come up with a good strategy to help grow a company, you know, in maybe in a vertical I've never played with before because of this database, right? Yeah. U so I'll I'll take a quick pause because I know that's that might be a lot to digest, but I'm happy to answer and it's very high level, but I'm happy to answer additional questions. >> Yeah. I I I'm tracking exactly with what you're doing is that I'm working with a number of uh uh with companies that are doing something similar le leveraging the the the latest technology. Part of what just kind of a thought as you were talking came to mind about how uh you know back in business school in the late 90s and you you kind of learned this formulated model for building out um you're going to start an entrepreneurial effort. you're going to start a company and you want to go and build like your vision statement, you know, like you've got your the executive summary of what we're going to go and do and you kind of build the the the early model of what you think the business will be about and build that model. Uh and then over time as things change, as the market changes, as customer demands might be different than what you had envisioned initially and and you kind of shift and change and grow things. Um it it's always healthy to go back on a regular basis and look at what what is our vision? What's our mission statement? What did we think we're doing? What how do we define the various lines of operations? Our business do those things, but it's still a very manual process. It relies on you to remember all of the learnings and input those things. And you might hire highly paid outside consultants to come in to help restructure things. part of the benefit of what we have like what you've just described and this could be like an individual like an owner operator. It could be, you know, somebody who's got 25 people, you know, 500 people under that are part of the company already that can go in and build out the skills, build out that, you know, outline the company and then, as you said, start feeding it the real world data of what's happening. Then you're not relying on your memory, your understanding, your insights to go and do that. you're going to start to get all of this other collective knowledge feeding into that and get insights that you would have never reached on your own from that. So, it's I it's from a again from a consultant, a business consultant looking at this and the way that I work with clients, building out a profile, helping them establish that initial model and those skills so that they can better understand what they thought they were, where they are today, and then based on where they want to go, what's the best path to get there. And we've we've never had this kind of automated tooling that allows us to do this. Some companies have have attempted that, but again, it was still reliant on the humans building the things versus the machine uh reading and analyzing and assessing off of our re own real- time data >> which I absolutely >> it sounds like exactly that's what you're going and doing >> 100%. Because if you kind of think about an organization, right, and what makes a lot of organizations successful is is the connective tissue >> between people. And a lot of this connective tissue, if you kind of break it down, I call it almost like organ organizational lore, right? It's like these stories that happen during the day-to-day. Some of which are spoken, some of which are written down, and some of which are quite hidden to be honest. But these are important experiences that the organization and his members within an organization [snorts] >> uh draw upon right as a collective memory source in order for them to figure out things like hey I've seen this challenge before I know exactly what to do right and the problem before is that let's say let's say we had a very senior member on team and they left for a new opportunity they've been here five years and you know we're so happy for them to get this new opportunity But when they walk out that door, a lot of that lore disappears. >> Yeah. >> Right. A lot of those great habits, the war stories, the the playbooks, you know, try as you might and you we've done the exit interviews, we've done the whole, >> you know, go through this >> document the time that you're >> I took I I always share the fact I took over I worked for Pacific Bell years and years ago uh in uh in in Pleasanton, California, San Ramon in Pleasanton, California. And uh I took over my first week there for a guy that was retiring. I had one week >> where we over overlapped. That's it. And it was three ring binders trying to document his brain, his knowledge of all these different things. And it failed. >> Just failed. >> But yeah, it's it's incredibly hard. Well, Edwin, I would like to ask like so most founders develop an almost intuitive sense for their customers, their products, and their market. And this is kind of like it's like that senior person who retires and trying to pull all of that knowledge out of his brain. Um, but you also have this this CEO there's this point where intuition all of that can become a bottleneck instead of an advantage. Like it's all in that person's head sometimes. And I worked for a company, a startup, a venturebacked startup over on the peninsula. Um, we're headquartered in Redwood City. They've since been acquired out of the they're out of Texas now. But um the the first CEO, the startup CEO recognized that he was not the right CEO to take the company to the next level of growth. That he was plateauing what he could do. He was a startup CEO CEO. So he left CTO became the CEO and then he was the second one. They've had two or three since then that that were outsiders that came in to help it grow to the next level. Contrast that to another 25 plus year software company that had decent success but just plateaued. And it was widely recognized in the industry that the CEO who owned like 75 80% of the company and refused to let it go was the one that was the bottleneck keeping it from growing and doing what it needed to do. when they finally the board finally got him out of that role. >> Uh then the company grew very quickly and then was acquired um you know very quickly after that. It did it did very well once he was gone. So again it comes back to that that that bottleneck. I mean how how do you have that conversation or do do most leaders even understand that they're having that bottleneck? >> A thousand%. Right. And earlier in our conversation I I mentioned something called the jet field DNA right so the jet field DNA is exactly that because it's one thing to have a database of playbooks and situations right but it is another thing entirely to understand what is the lens at which you view those things and how should you make judgment calls based on what you see in the database right so we're actually building that right now because we came to a collective conclusion that if we don't write this down, right, things could quickly get very messy from here because at this point, every single individual on our team, their out their output is on the cusp of going two, three, 5x versus what they were able to produce in the past, which is great. But the challenge is if that is not appropriately sharpened or going through the same judgment flows as everyone else around you, it could quickly get very messy, right? Um, so we're building something called a jet field DNA. And what that is is that we categorize basically my brain and my approach to any sort of growth challenge or growth problem into five to seven different buckets, right? essentially how I think about testing like how how would I test something like if I use the word test what does that really mean right like what's the actual definition of that and how am I cooking up the test and how am I evaluating and then what am I doing after the test has finished right so that's one piece of it the second one could be creative strategy how do I view creative strategy what to me is important for creative strategy and what is not important like what do I spend the most time on on and what do I spend the least time on and so on and so forth. There's there's about seven of these buckets that we've cooked up so far, right? And essentially what we're doing is we're going into Claude and we're having Claude come up with the questions for myself. So we're saying like, "Hey, pretend you are the world's most fanatical podcast host." Maybe pretend you're Christian Buckley, right? You really need to integrate uh interrogate the heck out of Edwin Choy, right? What are the deep hard-hitting questions you would ask them about creative strategy? Give me 60 of them. [laughter] Right? And then I'm taking these questions and whenever I have a free moment, I'm literally dictating answers to these questions into this document. Once I am done, we will have the jet fuel DNA. And this is the methodology of which it again it doesn't matter if you've been here for three days doesn't matter if you've been here for three years you should be doing this type of marketing or creative strategy or advisement for clients through pretty much this methodology it should be universal across the team right and then we'll constantly update it so I think there's a lot of different advantages to this one is that for onboarding it's going to be tremendous like it's going to be a living document. They're going to come in, they're going to look at our methodology, they're going to go, "Oh, holy smokes. I totally understand how Jet Fuel would approach this." And then when I am faced with a similar uh issue or opportunity, I know exactly how to act. Right? That's number one. Number two is that it's going to be tremendous for business development. So, if someone comes to us as a lead and they're saying, "Well, you know, what separates you guys from other agencies?" And then my answer would be like well there's no there's no magic answer to this but we have very disciplined methodology across all these pillars right across you know how we collect and analyze data across how we test and how do we push the envelope and how do we think about displaying your brand creatively right like we have a very very tight answer to that and we can say if you like that methodology we're a good fit if you don't all good you know let's let's part ways Then um and then the third is that our growing collection of AI skills and agents will be referring to that document, right? So then when the team, you know, when uh Joe and Mary are working with their own skills, you know, maybe they remix the skills a bit to suit their own preferences, but it still has this unifying connective tissue across all their skills. So they whatever the output is is going to be shaped and and sharpened by this DNA document. Right? So um and that's our way and I would love to hear better ways if if you've seen anything but that's our way of trying to address this you know how do we get the operator owner founders DNA um and how do we pull like extract it from my brain and how do we make it so all this tribal knowledge is not just floating out there in the ether right it's actually codified and written down >> I I think we're learning so much about this this process we're all experime experimenting together with this. Um, I I've often talk about that I read in I don't remember if it was back in the former like Red Herring MA magazine or Fast Company or something or other, but it was it was in the early 2000s. There was an article about a a an a gentleman who was uh older uh was based in San Francisco working for Microsoft, part of Microsoft R&D where he was I need to go find this article again uh because I t talk about it once in a while where he was recording again 20 plus years ago everything that passed in front of him. Now obviously he was turning it off to use the restroom and to go to sleep, that kind of stuff, whatever. But everything else during the day, every document he read, every book book, every television program, everything was being captured either photographed and or videoed and then stored away. And he was experimenting with things like, you know, how much does the brain retain and how do I get access to this data to this super memory around that? And I was so impressed by that I actually started my blog which I started the blog technically I in 2002 I started blogging but but my current blog started in February of 2004 as an extension of my brain to catalog all of this data and then the question that's persisted for the last 22 23 years has been now how do I get access to that data and do something with it. And I feel like that's kind of where we are. We're starting to see through these tools through AI that we're be we're able to access this data. There's still structure that needs to be done. There's still rules. There's the skills we need to build. We need to be clear on the output. There needs to be we need to be trained on how proper prompting to get things when we want something. What's the right way to go in and access that? But we're getting so close to being able to automate all of that, having access to all of that uh all of that that that memory and having access to it at the right time in the right way. Um and it's just it's it's going to exponentially allow us to grow and do things as individuals accessing all of that data. So I I'm excited about that. I I I think that again this fits right into directionally where things are going with information systems in general. >> Absolutely. And think about the tremendous amount of discipline that individual needed to remember to hit the record button. Right. >> Exactly. Yeah. >> And that's a lot of day-to-day friction. But now the friction is gone. Right. Because every move we make in a campaign, everything we send in a meeting, everything we send in an email, all of that is being recorded, organized, and stored away pretty much on a completely automated basis. Right? So now the question is if we have that data, how do we clean it up and shape it in a way where it's useful and then how do we get it to function as the brain or the hub of all that we do, right? Because my dream vision is that when someone says, "Hey, I went and checked up, you know, we just signed on with you guys. I love you guys. I think you guys are great operators. U but I I checked out the client team you gave me on LinkedIn and they've only been there for nine months. You know, what's what's the deal?" Like my my ideal answer would be well there's there's almost no change in what they're able to produce for you because they have taken you know the 15 16 17 years of every what everyone has been doing here and they have at their fingertips their output shouldn't be different at all right so that's that's the difference that we >> but I think when I think there's a difference between if you think of like the transactional the technical task to go and do like you've document ment a process you want how do I do the process well it's documented right here walk you through step by step what you need to go and do then you have the more intangible aspects of what a leader like the the cultural aspects and so maybe you could talk about that briefly like how and and how do you identify which of those founder instincts which are usually on the cultural side of things the way that they're doing something not just what they do but how they do that thing. Um, how do you turn that into a repeatable framework? >> Yeah, absolutely. So, I think it all stems well, first of all, let's define what culture means, right? I think it's very important to have a clear and universally agreed upon definition of culture. So, culture to me is is I think there's a couple of scenarios where it really kind of brings itself up to the forefront. The first thing for culture is on a day-to-day basis, how do people work and communicate with each other? I think that's number one. On the second kind of pillar of culture for me is when faced with an extremely stressful scenario or when things go wrong, how do we respond? >> Right? Obviously, I think culture is a little easier when everything's going swimmingly across like you're winning all fronts, right? you don't really see the true culture of a company until you are challenged or placed into a stressful state. So, I think those two buckets is kind of where I spend a lot of my time in in building culture. Like what are the day-to-day ways of working with each other and how we speak to clients, prospective clients and and amongst ourselves and then what do we do when when faced with a pressured fil situation, right? That's I think that's where culture truly is. Um I I read one quote which I thought was amazing is that if you have a good culture is like a rope. Like let's pretend that everyone is on a ship and it's going through a hurricane, right? If your culture is strong, right? It's a big strong rope that everyone holds on to and they all stay there. But if your culture is weak, everyone gets blown away because the rope is not strong enough. Um so that's that's that's how we define culture, right? And if that's how we define culture then at that point you know if that's if that's the case then we kind of codified it into you know six pillars right basically uh how do we speak to one another? Well we have uh what we call you know pleasant but very open discussions right almost um sometimes maybe almost offensively open is what we call it. Yeah, I there's a I worked for a company that was very salesdriven and uh yeah, they they they had a phrase for it. It was um uh uh charismatically aggressive. >> Yeah, charismatically aggressive. Ours is uh uh is it's basically like incredibly, you know, candid, right? We want to be we want to be pleasantly but incredibly candid. So, we want it to be very safe for someone to say, "Hey, I took a look at your campaign and honestly, I don't think it's going to work. Like, it's definitely not the best work we've we can put, and here's X, Y, and Z, and I'm going to sit down and help you get to the level that, you know, the bar and standards of what we would like to get to, right? And we make it so it's safe for someone to say that." >> Um, and we make it so it's completely acceptable and people expect that, >> right? Um, so that's like a big piece of of our cultural pillars. And then what we do is that once we have that pillar written down, we write down uh these kind of fake scenarios with fake names. So Mary would would uphold this pillar in this scenario and we come up with a fake scenario and then we we come up with John Smith would not uphold this cultural pillar with this scenario, right? He would respond to that, you know, radical candiveness. He would get emotional about it. He would shut down. Like we kind of list all the things of people who did not successfully exhibit that cultural pillar. And we just we basically have a visual for it. We do that for all of our cultural pillars, right? Who is growth oriented and who is not? And we have examples of good and examples of bad. Um and then one of our our really ones that we really hold near and dear to our heart is are you planting seeds for forest that you may never live to see? Right? Are you planting seeds for a forest that you may never live to see? Meaning that you are always looking to improve processes and improve things, maybe even for a generation of team members who may not even be hired here yet, right? Are you are you looking forward enough to make working at Jet feel better for the the two or three generations that are going to come after you? Right? And we have like examples of that, right? So, uh, hopefully that answers your question. >> Yeah, and I think that's that's good insight. I I do wonder like how you're handling how do you prevent um like a growing team like your own company uh and increasingly the use of AI tools, automation workflows from diluting the the original vision that you had for the company, the personality of your brand. >> Yeah, that's a challenge, right? because I think that's something that we for the last two or three months we spent a lot of time talking about is as part of our culture how do we use and integrate AI what is acceptable and what is unacceptable we have a pretty solid list of what's unacceptable that are even um they're pretty serious right if you get three strikes on this you're out so to speak right so you you you never are supposed to run something through an LLM and just hand it to anyone >> like your team member for a client. Like if if there's no human intervention, there's no human review layer. >> It seems like common sense, but we both know there's a lot of slop out there. Yeah, >> there's a lot of slop out there. So that I that's the first piece is that really we really want to establish that home and we take that extremely seriously, right? To the point where we have very serious discussions like, hey, did you look at this before you sent it to me because I do not think it is at all accurate what I was saying, right? >> [snorts] >> Um B is a ed educational and uh cultural component to it. Right? One is really educating our team. How do hallucinations happen? And why do LLM sometimes make these incredibly charismatic yet like completely wrong statements? Like you're just so overconfident in certain things. >> Confident but incorrect. Yes. >> Confident but incorrect. And then me sitting down with a team and saying very often like, "Hey, if something doesn't pass the smell test, meaning it smells funny, you need your your future career in any company depends on being able to sniff that out and stopping it before it goes any further, right? And then we have, you know, systems and processes for you to, you know, help spot those things, right? And then the third is you know making sure from a cultural perspective you know what is one of my favorite things I say and nauseium is that AI is really good for looking backwards right it has training data that goes backwards but it's it probably will I I don't know but it probably will never be good at taste making looking forward >> yeah I I know I there's been the discussions around that it's like well you know the future is really just a a reframing or learning from and then building upon you know the the past that kind of stuff. It's like but you know it it also if you if you understand how it comes to its responses it's averaging out the answers big across the the the various data sources and so it's not going to be unique original thought ever. It's that's not the way that it was built. That's not what it does. So, we run a lot of meetings and trainings where it says, "Hey, this scenario would really benefit from AI." Like, one of our favorite ways to use AI is that it's really good at taking information that lives in six different systems, tying it all together in one place, and then cleaning it up and making it so it's easier for you to make a true determination, right, or conclusion about what's going on. Very good for that, right? Yep. What it's very poor at is is it able to come up with creative or ad copy that is going to engage a 26 year old on Tik Tok, right? And Tik Tok is flavor of the week >> and and some of the stuff that goes viral on Tik Tok, you have no like AI has no idea what it is, right? Has no idea. Like we don't, you know, I'm not a creative person, so I have no idea. But for our creative team, like they're on the pulse of this individual, right? Or this cohort individual. So they're able to cook up creative ideas that, you know, AI cannot cook up and I cannot cook up, right? Because they they're tastem, >> right? >> Right. They're taste making the hottest new trend that will stop the scroll. Um, so they are not allowed to use AI to do that sort of work. >> Yeah. >> Right. Because it's very poor at that sort of work. So you have to understand where it would exceed humans and where it would be behind humans and really chunking that out and saying making sure that you know when you use these tools and skills and agents and enhancements and LLMs that you're making sure you're in category A and not category B. >> Yeah, that's interesting. I It's good to hear that you've got a a a role where you're like, hey, really no AI for in during this role. This is we we want kind of the natural organic creative. I I there I talked with another company where their first year uh developers uh coders they don't let them use any AI either. They want them to learn how to fail and then to improve upon that. And it's not like they're missing out on the technology. We all know they've got all the AI tools at home. They're doing it on their own. They're keeping up with the technology. but in their jobs they're not allowed to use it. I think that is fantastic having that intentional way because of what it does to the way the brain learns just like on the creative side. You get incredibly lazy and also uncreative. It's a it's a recursive effect on the creation of content to let AI create content for itself like purely AI. >> But absolutely. Well, let me ask a couple couple last questions here. I mean, when you're when you're looking at I want to get back to like the the leaders, the owner operators. When you're looking at some of the most successful leaders that you've worked with, what really separates them from those that successfully scaled themselves and those because I'm sure you've worked with some that really struggled or maybe even were permanently stuck in that operator mode. They're unable to escape that. >> What's the differences? Oh, uh, I have a very long answer to this, so I'm going to try to organize it in my thoughts. Um, one is that, you know, I I think is going back to that imagery of are you planting seeds for a forest you may never see? Because I think something that separates the really successful leaders from the ones that haven't been able to separate themselves from operations is are they sitting down and putting in the time to build something that isn't, you know, output driven for the next day, right? So let's say you you come in and you have a campaign and you have to build it for tomorrow, right? Um, are you sitting down and just simply building it for tomorrow and getting that problem kind of out of your way and out of your day or are you sitting down and saying, "Hey, this problem is going to occur a second time and a third time and a fourth time. And if that's the case, you know, instead of me just doing it and getting it off my plate, right, how could I make it so it's really easy not just for myself but other future team members to address this in half the time, right? So, it was really a mindset shift, right? It's very different from just simply I have a problem. I'm going to take care of it to I have a problem. I believe this problem will occur more than once. Therefore, I'm going to cook up I'm going to sit down. It's going to take more time. Um, but I'm going to sit down and cook up a solution that's scalable and repeatable and durable, right? Um, so I think that's the biggest thing is a mindset shift, right? That's that's really number eight. Number is being operations driven, right? It's and I use this analogy all the time. So, you know, obviously one of the most famous companies in terms of, you know, just making sure they're following the process and the quality there, the quality asurances there is Toyota, >> right? How they became a company that's known for repeatable quality all the way down from the assembly line u and the culture that they have there, right? Do we have something like that in play? Are we doing things just to do them once or are we doing things to build a process to do it a hund times at the same level of quality? Right? Very very different set of skills needed for that. Some people are very brilliant but not very process driven. >> Right? But we would rather take 20% less IQ but 20% more process driven because eventually the process driven individual the collective impact of that output would be higher. >> Right. Well, >> you can't learn if you're not moving forward. You can't you don't get data from that. You don't have experience. So, being more processed, you're going to get a higher volume, a higher output of those things. Therefore, you're going to learn more quickly than somebody who's getting stuck and and slower because of creativity or whatever it is that's slowing you down, but you know, from from delivering. But that is a I mean, it's just a a a truth. It seems almost counterintuitive to some folks, but >> absolutely. Yeah. >> You got to be moving forward. You can't I always say it's a one of my buckleyisms that I repeat all the time. It's like, you know, you cannot measure what's not moving forward. >> Absolutely. So, some people are really good at fighting fires and some people are really good at preventing fires from being lit. >> Yeah. >> Right. So you you you need to make sure that if you have someone who's really good at fighting fires that that individual's also equally becoming better at making sure those fires don't become lit in the first place. >> Yep. Well, Edwin, the last question is more a little more like about like your company and what you're building because I think what you're doing is like at the it's at the cutting edge of where agencies are going with with AI. And so the question becomes like it's three to five years out how do you think the structure of agencies and marketing organizations and how companies you know non-marketing agency companies are are leveraging these things. How does that change with with AI? Is it do you see it becoming is there going to be more consolidation? Do you think it's going to uh you know reduce the number of agencies that are out there or just do you see other fundamental changes like how are you looking at the future of what it is that you do? >> Yeah. So my my crystal ball if you want to call that and I want to say I've been many I've been wrong many times before. Um but what what I want to say is that service businesses are becoming more like software businesses. Software businesses are becoming more like service businesses. So what what this means is that as an agency, you know, we need to rely more on things like clean data, uh making sure we have the right data pipes, making sure that we have the right features put into play, but with an outsized emphasis on our human superpower. So, our human superpower is there's never really going to be, we think, um, in terms of LLMs or AI, the complexity and ability for it to go into a business and become your trusted right-hand person, your trusted advisor, right? You know, in in the Godfather, you have Tom Hagen, right? Like, will it ever become Tom Hagen? Probably not. We want to become Tom Hagen, but Tom Hagen, you know, powered by this incredible suite of, you know, customuilt software and customuilt conclusions from the software because we've been gathering this data for, well, wow, like three years now. Um, we didn't know what we quite wanted to do with it until very recently when, you know, Claude and other LLM got to that level of sophistication. But we have this like plethora of insights and knowledge and institutional knowledge in these databases, right? And then on the other side, a software company's got to be way more service based, right? You can't justify charging $5,000 a month for a software company that never talks to you anymore. They need to go white gloves. they need to add more services to and you've seen you've probably seen this Salesforce all the way down is that they need to provide much stronger service layers on top of the software to add value because your software could be cloned in probably eight days >> or a lot of the core functionality can be cloned >> or or get have incredibly strong and refined partner uh uh strategies. >> Absolutely. >> It's another another way. But uh yeah there's I I agree with that wholeheartedly. >> Yeah. >> So I think the companies or agencies that struggle will be the ones that struggle make that transition and then the software companies that will struggle will also struggle to make that transition. Right. So it's almost like a ven diagram. We're getting to this point where a lot of the rope tasks are being taken away. So like what is the you know what is the outstanding enterprise value that your organization will provide to another is really going to be you know a combination of the software and the service aspect. You can't really just rely on one anymore. >> Right. Yep. Exactly. Well Edwin I really appreciate your time today. Uh again, thank you for being on the show and uh Edwin Choy, so CEO of Jet Fuel Agency and we'll have of course all of your contact info in the blog out on YouTube as well and so really appreciate your time. >> Lovely. Thank you for having me. >> You've been listening to the Collab [music] Talk podcast. New episodes are published weekly and you can find us on Spotify, Apple Podcast, [music] iHeart Radio, and most other podcast platforms. Thanks for listening.