Most brands using AI end up sounding the same, but it doesn’t have to be that way. Ross and Drew Brucker discuss how to use deep personal context to turn AI into a true brand asset that amplifies distinctiveness instead of erasing it.
In Episode #91, Ross is joined by Drew Brucker, AI pioneer and marketing leader.
Drew Brucker is an expert in building AI-driven visual and written identities, helping brands turn AI into a true brand asset that amplifies distinctiveness instead of erasing it. With a background in SaaS marketing and a focus on custom “ghost systems,” he helps companies move from manual execution to strategic curation without losing their creative edge.
Ross and Drew discuss why most brands sound the same when using AI, how to avoid the generic trap, and the importance of front-loading deep context to build systems that protect and amplify what makes your brand unique.
Find show notes and episode highlights at [https://nwrk.co/_]
00:15.45 Ross: Okay, Drew, welcome finally to the podcast. It is funny that we are going to be discussing technology, but there have been multiple technological fails in the lead up to this conversation. But here we finally are together.
00:28.91 Drew Brucker: That is the truth. I think we were trying to do this in 2025. Here we are in 2026. So much has changed.
00:34.78 Ross: Yes. And we are doing it in one of the fastest moving technological revolutions that I have ever been part of, which is quite interesting.
00:44.47 Drew Brucker: It is chaotic and a whole lot of fun at the same time.
00:48.02 Ross: And I think when we started, at the beginning of our conversation, we were introduced by a mutual friend, client, and person, Lisa from Buddy. You had done a lot of visual work for them in terms of helping them build tools, systems, and processes to use these things to actually generate information. A branded story, a visual signature that they could do. And when we had our prequel, the thing that I wrote down that I found quite interesting is this idea of how you think about capturing a brand essence in a tool. Because one of the trends I am seeing now is as these tools become better and easier to use, everything is starting to look the same. We have been through these different visual trends with these tools, and everyone is putting out stuff that looks and feels the same. And in our world, different is better than better. So if you are going to look exactly like your competitors, you may as well not do it, do something completely different. So how do you think about using these tools and actually pulling in a bit of personality, a bit of soul, a bit of identifiability into them?
02:15.74 Drew Brucker: This is a really great question. This takes me back originally to getting into this space in the first place. My background for the last 11 or 12 years prior to AI becoming accessible was leading marketing teams, specifically in SaaS startups, doing everything and anything. And when I go back to an inflection point for me, I was already playing with some of the AI tools, specifically Midjourney at that time. Photography is a hobby of mine. I have a really good eye for these things. And obviously I consider myself on the creative side of marketing, more so than the analytical. Obviously both is ideal, but I always leaned a little bit more toward the creative part. What really got my mind going was once I was playing around with tools like Midjourney at that time, I had become such a master at that tool and what was currently available, where I was able to really create high quality imagery that matched what was in my head. That wasn't always an easy thing. It is much, much easier now because the technology and the coherence of these models have gotten so much better. But at that time it was really trying to solve a puzzle. You had to find the right formula to do these things. And sitting on my own marketing team, it was like, okay, we are in this niche space. There is not a lot of stock photography out there that feels like we could use it. And we didn't have this big budget to go and hire photographers to take a bunch of shots. There were all of those logistical elements and finding the right people. So I instantly had this aha moment where it was like, okay, I have become so good at this tool now. I think I could start creating on-brand imagery for us using AI. Now this was back in early 2023. This was a novel concept. No one was doing this yet.
04:16.83 Ross: Mm-hmm.
04:18.36 Drew Brucker: And I just started testing it and trying it. Sure enough, within a month or two, we started incorporating AI imagery into the fold. So thinking about things like icons, illustrations, and even some photorealism elements. I know photorealism is probably the most polarizing part of that, but that was really this moment in time where it was like, oh, if this is possible, I know this is a pain point for other marketers, right? That is a huge issue that we all encounter. And to your point, how do we stand out? How do we differentiate? Visual elements do a large part of that lifting for you. You think about a brand's digital visual footprint.
04:54.77 Ross: Mm-hmm.
04:57.38 Drew Brucker: I mean, it is really massive. There is massive opportunity there. So for me, I stepped into that. And that was a similar problem that I solved for Lisa over at Buddy. She was actually one of my first clients going out on my own. It was just that. They were in a space where hiring a photographer to go on site to these metal recycling plants, having the right people dress the right way, perfect lighting conditions, and taking shots from different locations, just wasn't really feasible from a budget and logistical standpoint where they were at the company at the time. And we recreated that. I recreated that same thing, but in a little bit more of a structural and clear processed way. And even that has evolved since then. So circling back to the beginning part of your question now, which is how do you build out this visual identity? I did something really interesting probably about six or seven months back with a big client of mine. A Fortune 500 company where we built a visual identity toolkit. So essentially it was this massive, comprehensive, holistic document that describes exactly who this brand is, their aesthetic, the things that feel like them, and the things that don't, which is also important with anything that you do with AI, the positive reinforcement, and also the negative reinforcement. By the end of it, it was like 33 pages long. So that is great. But it is also like, no one is going to read 33 pages. So what I would do in that situation was I took that 33 pages, synthesized it with AI, and consolidated it down to a system prompt that then could be used either as a copy and paste to prime an LLM or to put it into an agent or GPT. And that way it is primed already. All I need to do is describe exactly what I am creating for this particular campaign or project. And now it is operating with all of those rules and guidelines that we already set forth for the visual aesthetic and brand identity. And now it just spits out these prompts that are formulated and templated in a structure that I have used in a lot of my work and I know it works. And so that might be a little bit confusing to some people, but essentially what we are doing now is rather than me teaching somebody exactly how to prompt for something, leveling them up on all the vernacular, photography, lighting, camera angles, subject distance, elements that really tell stories, we have already done all that and we don't have to train. Now it is just like, I could have anybody from the team come in, put in a project, and now they can get those exact prompts to just copy and paste in whatever model they are using and get predictable, consistent output. So I think that sort of circle in itself has just been a massive thing. And it has evolved every six months or so because the technology just keeps getting better and better and changes the workflow and the process.
08:11.38 Ross: I think there is a couple of things that are interesting in there. The one is obviously that work on spending the time on things that you wouldn't have necessarily spent the time on before, because people are obviously using this either to replace an agency or because they can't afford an agency. But in theory, a lot of that stuff sits innately in the creative team. So I understand, okay, this is the feeling I want. These are the colors. I have seen your thing. I know you like shiny and blue and big and epic or whatever verbs we are sort of using. You have to spend the time thinking about that for yourself, either on your own or with a human team, making sure that we think the base of this thing is right. So getting that visual identity toolkit right is so critical because I don't think people spend the time doing that.
09:11.52 Drew Brucker: I know, I am a big fan of front-loaded work. I love solving something and spending time upfront, knowing that it is going to compoundingly get easier and easier over time, requiring less and less work.
09:14.42 Ross: Yeah.
09:24.12 Drew Brucker: So that was a perfect example of that, but also to your point. If this is sitting with a creative team and you have brand guidelines, those are really made for people, but at the same time, you are expecting somebody to go through and really internalize all of those rules from a human standpoint, which can fluctuate depending on how much sleep you got or how you are feeling or do you still like your job. That could all depend on how well you interpret those types of documents.
09:52.31 Ross: Yeah. Do you have 24 hours to do this brief?
09:56.86 Drew Brucker: Yeah, right. And that lives with you. But what if you leave the role and then somebody else has to come in and learn those things and then audit those things? So it is just finding the right ways that you want to incorporate AI into your work. AI can do a lot of things. Does that mean that it should be doing everything for you? Probably not. It is really just trying to pick the right areas that you know it can solve for and do it effectively.
10:27.03 Ross: Yeah, I mean, the second piece that I think is interesting is that people have tried to outsource creativity wholesale to these tools. And that is wrong in my opinion, because the tools are just a representation of what has been put in them. If it doesn't, they don't necessarily create anything. They just go and look at everything that they have got access to, and then they make a representation based on that information. And I think people should be spending their time being creative with these tools. They are spending their time saying, this is what I really wanted to feel like. This is the problem I am trying to solve. These are the ideas that I have. This is the weird thought that I had on the treadmill this morning. And then use it to bring that to life in a new way as opposed to going, make me a campaign for my brand to sell more things, which is going to get you something fairly flat and not humanly compelling. It is going to come off as a little bit alien. And all of this stuff is said with an asterisk because I have seen some brands do really crazy AI stuff and that is just part of their idea. They have embraced that and they are going in the other direction as everyone else.
11:44.98 Drew Brucker: Yeah, sure. That reminds me of when AI video was really bad, there was the crowd that was trying to spend as much time as possible to make it look perfect, which was so hard and so tedious to do for a long period of time. Versus somebody whose strategy is embracing the hallucinations or the morphing or the errors that occur in the video as sort of a strategy, you know?
00:00.06 Drew Brucker: Sorry about that.
00:01.30 Ross: That is okay. So you were in the middle of talking about people embracing the craziness and jank as part of an aesthetic piece.
00:13.53 Drew Brucker: Yeah, I think people that were embracing the morphing, the hallucinations, the silliness of it all. And in a lot of ways, those videos performed a lot better. From a creative standpoint, it is probably just more fun and less pressure given where the tech was at. Obviously AI video has come so far and it has gotten so much better. But for a long period of time, there was about the 90% of people that were just trying to push it as far as it could go and make it look as perfect as possible. But it just couldn't get there. And I just thought it was such a cool idea for those that leaned into the opposite side of it.
00:50.87 Ross: I think in all of this is to not forget, there is that builder.ai, that company they raised half a billion dollars.
01:03.03 Drew Brucker: Oh my God. Yeah.
01:04.65 Ross: And it turns out that they were just a bunch of developers in a large warehouse in India somewhere, just programming at rapid speed. That was what was happening there, which is insane. But we have this belief that these technologies will magically solve things. And I think we have an opportunity with where we are at right now to actually embrace the humanity in it. How do we as humans show up more and how do we use these things to represent our own weird ideas? Because the ones that have really worked and really resonated is when someone has created something interesting.
01:39.68 Drew Brucker: That is right.
01:43.02 Ross: Like Buddy, for example, they play with the visual. They have got their visual signature and they have fun with it and it tells you the personality of who they are. I am not looking at that image trying to go, oh this is a 100% real representation of what is happening in the world. They are telling me a story.
02:01.66 Drew Brucker: That is right.
02:04.38 Ross: Using a visual. And I think that is where people attempting to replace a fashion photographer with an AI fashion photographer hasn't landed very well. But people who use these tools to create visuals that tell a very specific narrative that they are trying to communicate, in whichever aesthetic you want to play that in, has been more successful. And I think it is a disconnect of this technology is not going to magically make your bad idea good. It is not going to magically make your boring concept amazing.
02:41.15 Drew Brucker: Totally.
02:42.95 Ross: It is not going to automatically take a poor input and upgrade it to something that is amazing. But it does have the ability to put the creative power into the hands of people who didn't necessarily have the technical ability in the past.
03:00.12 Drew Brucker: That is where we are at. And first of all, what you said was spot on, but I think that is the most exciting part. No one is confined to this formal career path that we were all sort of on. What I mean by that is I wasn't a creative director, an art director, or a graphic designer. Did I like those things? Yes. Did I try to do those things as much as I could when I was a marketer? For sure. But I also wasn't going to be able to just step into a well-known agency and get a creative director job. And I wasn't able to just go out and probably creative well-known clients and do that type of work. This has really just democratized all of that. The barriers to entry, you don't need this formal pathway. And again, with how significantly AI is embedded and will be in the things that we are doing personally and professionally moving forward, this isn't going away. So it is just those people that took advantage of the early window and found opportunities with it have been rewarded and will continue to get rewarded with those things. I just think it is fascinating. The same can be true even with the product side right now. I have got a client, a well-known platform, they need a lot of internal tooling and so forth. And I am not a product builder. I am not an engineer. But I have been playing with Claude Code. I built things. I wouldn't have been able to do those things before. And now they are tapping me and say, hey, can you build this for something that we are going to use for thousands of employees? That would never happen. And so it is just a product of being curious, testing and trying things with this technology, and just finding things that really appeal to you. And it opens up this wide lane for all of us to just say, okay, am I liking where my career is headed? Do I want to pivot? There has never been a better time to find a new lane or explore that new lane than right now.
05:09.72 Ross: I also think for companies, there is a great opportunity. If I think of a company that has embraced technology really well over the years, it is Nike. Remember, they were one of the first people to have that little pod you could put in your shoe. And they designed the shoes with a little hole in the Nike Run app.
05:24.12 Drew Brucker: Oh, man.
05:26.71 Ross: So they were playing with technology, which wasn't their thing at the time, playing it back to their products.
05:31.58 Drew Brucker: Yeah.
05:33.64 Ross: And I think now you don't have to be Nike. You don't need the $25 million that they spent building that experience.
05:40.92 Drew Brucker: Also a great point.
05:42.35 Ross: You can look in your world, around the people that you serve, around the problems that you are solving, and ask, what can I build that solves a very specific problem or does a very specific thing, the kind of things that you wouldn't do under normal scenarios. I mean, a friend of mine is building a chief of staff for mothers.
06:07.32 Drew Brucker: Yes.
06:13.14 Ross: So it is a family running tool. It is going to help the mom do all of the jobs that she needs to do and remember all of this stuff and be reminded that Tina's ballet is at Wednesday at three and it has now changed and don't forget the t-shirt on Monday.
06:25.54 Drew Brucker: Oh, love that idea.
06:31.63 Ross: And then on top of that, the budgeting for the household and all of these things. Nobody would spend the time chasing that particular market, but because he sees that need in his world with his partner, he is like, oh, I can solve this. And he is not a software dev. He doesn't have a team of 50 people, but he can build a product around a thing. Coming back to branding, branding is a promise of solving something for someone. We do this for this group of people. And now there is opportunity to build products, experiences, software, abilities to make things more consistent, more amazing, which I think is super interesting and exciting because it comes back to that humanity again of, I know my world so well. So I know where things come undone. So I can now solve some of those problems. And if they are coming undone for me, just like you said, you built it for yourself as a creative for your marketing team. But now this is relevant for other people because you are not the only one trying to market a product. There are millions of people trying to market their product.
07:42.52 Drew Brucker: Yeah. It is a fascinating time because especially for anybody that would consider themselves curious or creative, there were probably just so many reasons why that person wouldn't have done that before. Whereas now there is maybe one or two obstacles that could be in the way. And it is probably just time and then potential resources.
08:13.81 Ross: Mm-hmm.
08:17.18 Drew Brucker: I mean, the cost is no longer a factor. You could build an MVP of just about anything in hours to a few days. So just the concept of that. Visually, again, the same thing. A lot of people take AI and think that it has got to be this all in or all off kind of thing where we are either using it or not. The truth is somewhere in the middle. You don't have to use it for everything. It could be a small part of this process. It could be a large part of the process. It could be in internal work. It could be in a little bit of external work. It doesn't have to be, hey, we are just using AI for everything, and shouldn't be. But I think that alone opens up so many doors here because I was just thinking about me as a one-person team. I have got clients. My days are really focused on maximizing the hours I have. And I buy a lot of tools. I have to have a lot of tools, but I was just thinking, hey, Calendly, I pay $140 a year for Calendly because I have multiple calendar setups synced with multiple calendars. And I was just like, I wonder if I could build my own built-in-a-day replacement.
09:32.99 Ross: Mm-hmm.
09:36.66 Drew Brucker: I didn't know anything about authentication or APIs for the calendar, none of that stuff. I just wanted to make my own, see if it was possible. It worked. Okay. Cancel Calendly. You know, I just think there is this idea that a lot of these simple SaaS products, if you don't have a comprehensive, complex product at this point, that could be troublesome. Anybody could create these things, start a business, sell things on the side, undercut things at a fraction of the cost. I just think this is the most disruptive time. And I don't even think we are fully there yet. I think probably by the end of this year and into 2027, we are really going to start to see that. We were just talking about before the show. You have a Mac mini, you are exploring the open Claude thing. I am testing and building all these tools either for clients or myself or processes. I mean, the real opportunity here is the personal context that you can provide with AI to break out of the box of the outputs that everybody else is getting. Just as much information as you can share. So for example, Claude for me knows all of my business client contacts, my communications with them, how they like to be communicated to, the projects we have worked on together. I have got a memory layer set up for that. So now I can just say, hey, write an email to Tom about this. It knows exactly the project I am talking about. It knows exactly the past conversations. It is primed with past conversations between me and Tom, all of those things. So I think this just dumping of internal personal context is something I am completely fascinated with at the moment.
11:30.90 Ross: So do you think the companies and brands that are going to stand out are the ones that are going to embrace that? Because I was on a call just before this with someone, and he was talking about how he is so tired of all of this AI stuff that he wants something that is human-made. So I see it has the potential to split into almost two distinct camps. There are people who are deliberately staying away, and then there are people who are embracing it and using it to build these custom things and these tools and these processes and things that are super interesting.
12:05.24 Drew Brucker: Well, I guess the question is, do we have a problem with things that AI makes versus humans and where does the line split? I think when I think about my personal experience, the most polarizing parts of that are things that include photorealism, people, real elements.
12:14.99 Ross: Mm-hmm.
12:23.96 Drew Brucker: So you brought up photo shoots earlier. Could AI replace photo shoots right now? Yeah. Oh, yeah. I have run several experiments on that. I could even show you something right here. It is a whole workflow where I could take any person, any model, as long as I have got an image of them, the items that I want them to wear in the background. I could even create that. Then, using sort of a process of how I want to prompt it, I can build all these different photo shoots. I can change the texture of the shirt. I can change the color of the shirt. I can change the type of shirt it is, the jeans, the pants, the shoes, whatever it is. But I think when it comes to those things, the personal craft of it, people don't like that. That is the polarizing part. Me as a photographer, it would mean more to me if I hired somebody that I knew was a good photographer and I liked their work. That would be exponentially more valuable to me to have them do it versus have AI completely do it. Just because I think it is more personal. And you are helping another person in the process, which I think will also be a bigger issue in the future. But to that point of not having AI do everything, why can't I use both? What if I hire that photographer and we didn't get all the camera angles that I wanted to get? Now I can create those. I have got all the raw materials from the photo shoot that I can build to expand out that world with different camera angles.
13:51.60 Ross: Mm.
13:58.93 Drew Brucker: What if I forgot to shoot one of the shirts, one of the pants? What if I forgot one of the colorways I wanted to show? What if I have got a new product coming out that I want to put on that same model and that model is not available? That is where it is like, okay, those two things together, not one or the other. So I think that is really the opportunity that I look at. But I think all of that art, craft, creative element, personalization, touch, those things I think will be a little bit more polarizing and a harder choice for us. But the things that happen in the background, I don't care whether a human does it or AI does it.
14:38.30 Ross: Well, I think it is also there are times when a human wouldn't do that. Because you wouldn't have hired a chief of staff to look through all of your conversations with someone to make sure they have got it and summarize the thing and give it to you. You just would have done without. You probably would have done it badly. But I love this idea that there is a confluence of things for brands to think about.
15:04.02 Drew Brucker: Yeah.
15:06.50 Ross: What are the processes that you have that are distinctly yours, that make you feel like you? What are things that you do that nobody else does? And then are there ways that you can use these tools to personalize that or make sure that it happens all the time, regardless of whether Drew had a good night's sleep or not? Like it always happens because these tools have got the ability to sort of do that. And I also think I see everything as a bundle of resources, right? So if you are a brand or a company and you have got a budget and you used to have to spend a bunch on tools and analytics and all of these things and now in theory that can go away. I put all of our Google Analytics into Claude and I was like tell me the 10 things I need to do to improve my website.
16:08.98 Drew Brucker: And don't you just remember how painful it was to go into Google Analytics and try to find something that you wanted to find?
16:16.41 Ross: Yeah.
16:16.97 Drew Brucker: I mean, my God.
16:17.21 Ross: And I also don't even have the ability to process that information. So it would be right there in front of me and I wouldn't see it.
16:22.54 Drew Brucker: That is true too, right?
16:25.53 Ross: But now you have freed up all of this budget. And this is, I think, coming back to humanity. Now you can spend it on the other stuff, the stuff that used to be nice to have. Like we wouldn't have hired a photographer or a better photographer. We wouldn't have had a stylist or we wouldn't have had somebody hand illustrate 500 things because we just didn't have the money for it.
16:43.28 Drew Brucker: That is right. That is right.
16:47.86 Ross: But now that if we can pull out all of this other budget, now let's spend it there because I think that is super interesting because now I have got an artist who has done 100 paintings for me, and I can use that as my training. So I have got 1,000 images for the next 18 months to two years, our visual style is completely done. The artist is happy because you bought 100 paintings from him and the rights to do this and you aren't going to look like everyone else because you bought this little human thing that he maybe did something that you might not have done. And I think this is where the wholesale outsourcing of creativity freaks me out a little bit because I will have an idea in my head, right? And I have been doing this 25 years, and I feel like I have got a good idea. But sometimes I will talk to the junior designer on set, and he will be like, oh, what about this? Or she will be like, have you thought about this? Or I don't like that because of the following reasons. Or I was listening to this new song and it made me feel this way, and I feel like we could bring that feeling into this. And suddenly the product, the feeling, the brand, the emotional bit of it is much more deep and complex, which is something that still to this day, I mean, the tools might get there eventually, but only humans can do that stuff. Only humans can capture that indescribable essence. And it is often when more than one human works together on something that it happens.
18:35.64 Drew Brucker: That is a really interesting point that you just made. I agree. I think we are all making that choice and those decisions may evolve, right? What we want AI to do now, we may not want to do later based on experiences, based on the way the technology changes.
18:47.10 Ross: Yes.
18:55.67 Drew Brucker: And the opposite is true as well. Things that we maybe do right now, we decide we don't want to do. I just look at some of the things that are recurring tasks in my own personal work, how much input is required from me. If there is not a lot of input required from me, but I know I still need to do something, I might have AI play a role in that process. Versus something that requires so much context. And that is where I think even using AI just to get you acting, taking you from zero to one also helps because this is a collaborative tool. Just by me providing all of that context, I can start to understand things that maybe I should have asked that I didn't. Or other ideas, or if I don't know how to do something, troubleshooting. And I think AI is so good at troubleshooting things as well. I get stuck all the time. And what did you have to do before? You had to ask somebody you knew, you had to go online and search. You had to maybe find a YouTube video. And even then you had to watch the video to see if they actually answer the question. That is not something that we have to spend our time doing anymore. So I just look at those low stake things, recurring tasks, collaboration on bigger ideas where maybe I don't have somebody that is easily accessible where I can start to take it from zero to one and really feel like I have got a better grasp of clarity in what I am doing and what I am trying to achieve with the outputs.
20:41.26 Ross: I love that. I mean, I guess where this is coming to is a lot of this is around getting to a point of everyone needs to become better at curating.
20:45.44 Drew Brucker: Mm.
20:53.94 Ross: At collecting the things that they like or want represented. And they could be anything. I have got this here on my desk. My son made a tiny little green fried egg.
21:17.56 Drew Brucker: Yeah, that is awesome.
21:17.66 Ross: And I just love it because it represents a thought that he had.
21:19.53 Drew Brucker: Yeah. Right.
21:21.93 Ross: He is seven years old.
21:22.74 Drew Brucker: Yeah.
21:24.01 Ross: He was like, I want to make a green egg. And he wants it because he read a book by Dr. Seuss called Green Eggs and Ham. He was like, why do we need the ham? Let's just make green eggs. And I love that idea. So that is my little curated thing. So sometimes when I am thinking, I am like, okay, what is my green egg here? What can I play with in this thing? So when you are setting these systems up, how do you think about the curation of them so that they can do their job better without falling into that generic feeling that so many people get?
21:59.22 Drew Brucker: Okay. It is a big question. I have a few things I can get into to answer that. I think directly with Claude and really how it has changed the game with these markdown files, these files that have depth within them that are almost like these subfolders. That is where you are really getting into context. So, for example, like Claude, when you start doing stuff with Claude Code, you need a Claude MD file, which is essentially your master document, and it is all about you. It is really just about your preferences. It could be your curation, the things that you like, the things that you don't like. And I think that specifically is one document that I have spent a ridiculous amount of time on. And the reason for that is because Claude takes that into account every time I have a conversation with it. I have already primed it with that information. It is the most important file at this current time that you can have with the system. And so I spend a lot of time with giving it context into me. So I will give you a really good example of this. We all know how AI LLM writing sounds and looks. I have read it so much. I am a former content person. It drives me nuts. I mean, you go on LinkedIn, go on X, whatever platform you want. Even videos now with the scripts. You can just tell an LLM wrote it. There are patterns that emerge. And as a content person, that stuck out to me because I was always like, man, how can I get this to not sound like an LLM wrote it? I want to have it give me the 75% to 80% and me build on it. I am cool with that. But I want to kind of keep making these same edits every time. And I just think it is a perfect way to lose trust with people too. You might be touching on a topic that you care about and you can provide insight into. But when it feeds it to people in the same way as everybody else, it sounds like everybody else and it sounds like nobody at the same time. And so I was on this mission of how can I solve this? So I would say I am about 90, 95% of the way there. I have got something I am building right now called the ghost system. And essentially what it does is the idea is that I can have it write for me at an exceptional level without it sounding like AI at all. And I thought at first this would be a very easy problem to solve for. I provided some writing examples. It knows the way I write. Still doesn't do it. Wants to inject these little pieces where it sounds like an LLM. So I started to identify the patterns of all these things.
25:10.21 Ross: Mm-hmm.
25:13.88 Drew Brucker: So now I have examples, but now I have also got what I call the kill list. All these patterns that an LLM uses, what I wanted to identify and not put in any of my writing in the future. So I started building that out. But then I started building my own voice file. So going into the examples, that would be a portion of the voice file. But to that level of curation that we are talking about, that specificness, I started giving it words that I use, words that I don't use. I don't use "however", I use "but". I use a slang word instead. These are my verbal tics. These are the verbs I use. These are the things that I will never say. These are the hooks that I like to use. These are the architecture types I like to use in my posts for social media. I just went 22 layers deep on this voice profile. And so I have got that, the kill list. I have got all these little components that work together as a chain that happens every time I ask Claude to write for me. And now I have got it to the point where there are no LLM patterns. It sounds exactly like me. I am 95% of the way there with anything that I haven't created. But that is all because I spent so much time on the context, the specific specifications of me. And I think that is the same when it comes to curation and taste of anything. A lot of people say right now, and I was one of the first ones to say this, taste is the moat right now.
26:51.71 Ross: Mm-hmm.
26:53.11 Drew Brucker: I think that is going away too. I think taste is a little bit more of an amorphous thing you need to put in the reps. It is very subjective person to person, what you have been exposed to. But hypothetically right now, I can build infrastructure that helps me with taste through specificity, through context, through examples. It is just spending the time to build those things to then have it operate at that level. It is just where tech is right now, it can take longer than it should. I am seeing the same pattern that I started to see with AI imagery and AI video. It was like, yeah, it can be good, but you got to do all these loop arounds to make it as good as possible. And then the tech got so much better where it is like, oh, it is really easy now. I think the same thing is going to be true here. It is just how much time are you spending to develop that right now? So I think that is a long answer, but hopefully that shines a little bit of light in the way that I think about it. It really just comes down to spending a lot of time providing that deep level of personal context into anything that you are trying to do if you want to get a predictable result that is true to you. And the same could be true for a brand, right?
28:11.42 Ross: Well, I think the thing is people often don't want to spend the time here.
28:15.77 Drew Brucker: And who has the time? Who has the time right now? Things change so wildly. I think that is the other thing.
28:20.25 Ross: Yeah, they do. But I think at the same time, investing in these things, these are assets, right?
28:27.97 Drew Brucker: That is the way that I look at it.
28:27.74 Ross: This is your asset. So you invest time and energy and effort into building this thing because you are using Claude as an example today. If we had this conversation when we started, we would be talking about ChatGPT.
28:39.27 Drew Brucker: True.
28:39.90 Ross: And in six months' time, we would be talking about Perplexity. And six months after that, we will be talking about whatever. Or everyone is building their own custom one. Open Claude wasn't a thing. But I think what brands and people and companies need to think about is what are these assets that we want to build over time? Because the logic will remain forever. Your kill list, your do's and don'ts. But it takes time to really invest in this and to pull it out.
29:12.87 Drew Brucker: A ton of time.
29:14.59 Ross: And I think it gets faster. I have been using these tools as a brain dump. I will start a project, so I know I have got this client coming in, I will just start dumping all my thoughts into this thing. And I say, don't process anything. Just right now, your job is to catch everything that I am doing. And I will tell you at some point in the future to sort of do it. You just dump and dump and dump. And this is the curation piece. And then when it gets time to the ideating and the creating part, now it gets much easier. So I think the shift is everyone always wants to spend all the time on the end, right? Now we are saying we have got to spend a bit more time up front. We need to do this kind of stuff.
29:56.02 Drew Brucker: Yeah, go upstream.
29:57.29 Ross: Because if you do that stuff right, then it can release so much of the painful, time-consuming work and so much of the expensive budget to let you then spend better time at the end on the polish, on the fun stuff.
30:11.80 Drew Brucker: Yeah.
30:14.21 Ross: So now we have invested up front and we can now spend all of our time in the fun part at the end, but we have to do this part in order to make it feel like a brand, in order to make it have personality, in order to make it feel like something somebody would actually care about as opposed to, I think the word you said is this thing can destroy trust and I fully believe that. There is a guy I know who I had a lot of respect for who is commenting on everything on LinkedIn. I am like done man. You can't even be bothered to come talk to me.
30:42.62 Drew Brucker: Oh, no. So that is another thing, though. Hypothetically, if I wanted to do that, if I produced and built a specific thing that was focused on how I leave comments, going back to that ghost system. I started first with how I typically write LinkedIn social media posts, because that is what most of the time I do with the writing stuff.
31:10.42 Ross: Yes.
31:11.77 Drew Brucker: But from a comment side, you could do the same thing hypothetically if you wanted to do it. It just needs to know how you comment. You also would need to scrub all of those AI hints that it leaves. So what remains is true to you. And I agree, man, those AI comments, I mean, they are out of control. They are really bad. But I think, again, if you were to spend the time upfront, just dumping personal context examples, then, in theory, you could solve that.
31:45.43 Ross: So I am aware of time, Drew, so I would like to ask you one more question. We have had the time delay from when we had this conversation to where we are at now. In your opinion, what are you excited about that the next six to eight months are going to do for us? What are the spaces, the opportunities, the exciting things that people can get into now?
32:24.37 Drew Brucker: I mean, it is really exciting right now for me. I don't know what the future holds. I think, not to be too existential about this, but I just hope that this technology at the end of the day ultimately is guided by the right people. It is in the right hands. It is very powerful stuff. And I think in a lot of ways, all of us at a larger macro level are going to be like, jobs are going to be wiped off the board. There will be other new jobs that pop up, but at the same time, there are so many questions in my mind, how this is going to play out over the next year, three, five, 10 years. I mean, I have got young kids, I think about this stuff all the time. I think right now it is really exciting because anybody can do anything. The tools aren't perfect. And kind of what we were talking about earlier, sometimes you have to double check things and maybe it made an error. It notices later, but it didn't notice in the moment. As I am building things that require multi-components, everything looks good. And then the next day, something is broken. What is going on? You got to think, oh, I didn't see this the first time. But I think it is just really, really exciting for those that are playing with it because you are starting to realize that really anything is possible. You can be anybody you want to be in the sense that you could make a career pivot virtually overnight. Essentially this is a competitive advantage if you know these things really well inside a company. Let's be honest, going back to the time thing, there are plenty of situations out there right now where people are so busy in their day-to-day job that they don't have time to play with these things. They maybe have played with them at a very foundational beginner level. And I think that is the reality. I have talked to plenty of clients like that. And even just thinking about my time in-house, you don't have time to just kind of go off on a riff for an hour or two and see what is possible. That is the advantage of being in this work day-to-day is I do get to play around with that stuff. And even then I feel like, oh, I saw this and I want to try this, but I don't have time.
34:48.57 Ross: He is being left behind. Yeah.
34:52.51 Drew Brucker: But that alone, I think, is just so enamoring because it feels like, I am 40 years old, to me it feels like when I was a kid and I got on dial-up internet for the first time and you are just like, what do I do? This is a whole new world. And that is what it is right now. I don't know what this looks like in three to five years, but I think if I am a company, I have got to upskill my employees. There has got to be some sort of way that you are upskilling employees. Yes, you are going to lose people, but you have to be doing that. You can't just hire people that don't have this innateness or curiosity about it. But also they have got to be piloting things. They have got to try micro experiments. They have got to figure out what it means for them. What examples are they going to use AI for? What situations, and scenarios? Where are they not? Because it is just going to be harder and harder to catch up later down the road. You don't have all of that context that existed beforehand. So I think that is part of it. But in terms of something that is very practical, I don't know. I guess what I would just be excited for, maybe this is a little bit boring, but I just like all of those dumb, time-consuming things that we used to do that we don't have to. Like my wife will still call up customer service and hang on the phone for an hour to get something resolved. I stopped doing that like four years ago. It was just so painful to me. But AI is just going to disrupt a lot of these things that we are not going to miss. And we are going to forget about instantly once it solves for us. So I think it is stuff like that. And I don't know, I am just very optimistic for people that are embracing this stuff. It is still early. So if you are listening to this stuff and you haven't done what you wanted to, still early.
36:50.89 Ross: Right. Drew, thank you so much for sharing your thoughts. I have got some stuff to think about in terms of have I gone deep enough in terms of training myself into these things, into building these kind of assets that I can carry from platform to platform. Have we done enough for our clients? Do they have this, have we built this for them so that their stuff can be better? So thank you so much for your time, for your energy. And despite all of the technological failures, we two humans got together in the end and we did it.
37:25.88 Drew Brucker: We got here.
37:27.94 Ross: So thank you for coming on.
37:28.60 Drew Brucker: And now I am just going to be thinking about that green egg. I want one of those green eggs.
37:35.16 Ross: Awesome, man. Thank you very much. And we will catch you in the next one. Bye-bye.
37:39.19 Drew Brucker: Thanks, Ross.