Welcome back everybody to Sacking Gross. This is Cassidy. Hope you're everybody as well. Carl, how you doing?
Doing good. Ready to rock. Yeah, this is exciting. I'm looking forward to this conversation.
I know Carl, you and I have been talking a lot about AI, a lot about generative AI, what it means to marketing and sales teams. But, you know, we haven't actually done anything about it. We just talked about it. And so what we've done today is we brought Grant on an Omar who've actually done something about it.
We want to hear all about it. So Grant, Omar, welcome to the show. Thank you, Cassidy. Thanks guys.
Great to be here. Yeah, it's great to have you. Why don't we just have you guys do a quick introduction. I know Grant, you're kind of a friend of the show.
We've had you on before. Love you here a little bit about what you do and then over to Omar and then we'll jump in and get going. So why don't you take it away. It sounds good.
So I'm the VP of marketing at Xingtree and Xingtree provides Nespex Action Software to about 700 companies. It can be used in many ways, but the most common way is for boosting contact center agent proficiency. Essentially, you can turn every agent into an expert by guiding them in customer conversations through step by step or close. And it's really fun to be leading the marketing team here with Omar and others.
Yeah, Omar, what's your role? So I am a senior brand designer here at Xingtree, but I actually do more than brand design, like, you know, the digital as I saw like AI stuff, one pages, email campaigns, ebooks, whatever, like considered as brand design, digital print, whatever it is, but I also do like UI design, like website. I did an overhaul of our entire website and yeah, I do also protocol, product illustrations, animations and all that good stuff. I want to.
Okay, that's a good context. I think it's important for this conversation, both in terms of the role, but also a little bit about the Xingtree. Obviously, you guys are doing something that's important and great. You mentioned contact centers, et cetera.
But it's not something I would normally think of as kind of the marketing team is trying to do some cutting edge stuff around how to use gender to AI. So we saw this on LinkedIn. We'll obviously share it as part of the show notes. So everybody can see what we're talking about.
But can you guys describe like what did you do? Like how did I want to get into this, but like explain to the audience, like what they're going to look at when they see the ads, et cetera, in terms of the process it went through. Sure. Great.
You want me to go? Yeah, please do. Yeah. So I'm going to talk about the ads that you're going to see.
They're like, it's called like turn your agent into rockstar, which is playoff of our vision of Xingtree. The vision is turn every agent into an expert, but we turn it turn every agent into a rockstar. And basically the idea came, I know Grant was in a meeting with Blake. He is our demand gen freelancer.
And basically they were playing around and Blake did like a rendering, but I'm not sure which one he used, but maybe Dolly. He used and it was of the band kiss as an agent. So that kind of what started this rockstar campaign. And I knew because I've been using mid-journey for a while.
So I took that and I applied it more on my end. We'll get into the process later, but basically our very first ad was just like we did the band kiss as a rockstar agent with the headset on, with the laptop in front of him. And it kind of started from there and eventually the person who actually motivated us a lot to start with AI design is our CDO Brandon. He advocated for AI, not just like we're kind of like having that being integrated with our product, but also he just loves using mid-journey as a creative tool.
And he started showing us some stuff. So you got it excited about some stuff, but they were not really related to our product, which is stuff like he was doing on a personal level and he was just sharing like excitement about like design elements he's working on. And yeah, that intrigued us to get a subscription to mid-journey and Grant basically told me just to have fun with it. There was no like actual intention yet, but the ideas flowed as we started using it more and more.
And now like it's kind of like we can't stop. I want to say it's on my end, especially as a designer. I literally use this like if I'm watching television or something, I'm doing something brain-on, just using my phone instead of social media and doing scrolling. I just use mid-journey and I'm popping an idea.
Like what does AI think about this, right? Or like how do you imagine this view or whatever it is, right? So that's how I've been using personally too. But yeah, we're having fun with it.
Like you said, it's not usual that we use that for, you know, when you think about, you know, an agent scripting company or decision trees or things like that, but we've been finding ways to do it. And it's been very, very fruitful. I think I have a lot of questions. Yeah, I have a lot of questions.
But Carl, why don't you jump in. Well, I love this stuff and I totally resonate with you all more because like I was like running out of credit on I use night cafe and I was generating like a whole bunch of stuff. And I was trying to produce actually your ads because I was trying to figure out like what prompts did this guy use to get these? And that's like a question that I have.
I don't know if you're open to like sharing that or if your prompts are confidential. But I think it's fascinating that it's so addictive. You just can rip through prompts so quickly and iterate on them. And yet these are wild man.
And so anyways, I don't want to get all like excited yet and I'll just talk forever about that. I want to go back to something that you sort of said or implied. One of the things that struck me when you were talking was that like you all have like a culture or something like an air of innovation. Yeah, go play around.
Go do this. Oh, the CPO actually is playing around with mid journey. It seems like everyone over there is sort of like, can we do weird stuff? And not only is that what I think incredibly rare, I'd like to hear if there are other ways that you all have used this kind of culture of innovation to like what other weird stuff have you done outside of AI.
But yeah, so that's just like that's just what other ways you feel like you're going to be using it, you know, internally because I'm like sales decks, talk tracks, website copy or design. I mean, it feels like it's endless. So anyways, a lot of questions there. Can we start with culture?
Like, how does that is this common for you all? And how do you feel like it's like a strategic advantage in your space, which is kind of boring, you know, it's like call center tech. This is a perfect marriage here. Anyways, thoughts?
Yeah, I would, I agree that it does take a level of culture to innovate in this way. And it's something that we try to do as a company. So we have a Slack channel for instance called AI's entry and everyone's just putting in links or new tools that seem interesting and people cross functionally are in there. And so, you know, I will drop in a sales marketing one, someone else will drop in an engineering thing.
So having that just kind of breeds some creative sparkers for people. But, you know, we're really brought into the kind of refine labs frameworks for creating, capturing and naming demand and to do that well and to capture people's attention when they're scrolling through the feed, you have to be creative and innovative if you want to get someone to actually stop and pay attention. So that's really something that we think about a lot from a marketing perspective. And so this naturally fits in as a new way to, to try breaking out of the mold.
And Omar, once he really caught the spark here, he started using it for many other things as well. You know, like zoom backgrounds that were partly curated with, with mid-journey blog post images, website images, et cetera. We've used AI voices for demo videos, not even everyone at Zingtree realizes this, but they sound so realistic that people didn't realize this. So, you know, iterating on copy, these types of things are other ways that we're exploring it.
But making sure to give people the freedom to do this is what's helpful to explore these from a culture perspective. Some objections potentially that it sounds like you all are like Zingtree is like a free for all. Like as long as it's creative and like just reasonably inside like brand guard rails, like just let it rip. But a lot of companies are going to be like a lot more conservative about some of that.
How did you all sort of balance like creative freedom and liberty with like making sure that we're within, you know, guard rails of brand? You know, the nice thing is our CEO, Joanne is a proponent of making things bold and standing out. So that helps that it's from the top. But, you know, Omar created our brand guidelines.
So we kind of got to set the criteria for us to say, Hey, let's, let's be creative in these ways. If that's not necessarily the case for a company that someone's working at, I would suggest trying to do it in small ways and iterate and maybe call it an experiment. And that might help you get closer to it. Yeah, we're big fans of that kind of narrative.
It sounds like you've got a little bit of a framework there, which maybe we can jump into. But yeah, imagine like, let's say Grant or Omar maybe hear from you, like you were dropped into just like a stodger environment, right? Think of maybe like a fintech company that sells to banks. It's like, no, you know, you hear this classic objections, right?
Like, no, they're not online or no, they don't laugh at TikTok videos. Like it's just like this weird impression that either, you know, it's like other humans, but what do you do when you're dropped into a culture like that where it's like, you have to be more careful. No, our brand is conservative. Our voice is conservative.
We got this very cold and kind of professional brand. What would you suggest and maybe our CEO doesn't like he's an old bank executive, right? Or she and so they don't necessarily think like how we pushing the bounds of our creative and our marketing, et cetera. What would be like the first couple of steps or how would you think through tackling that and getting your company to a more innovative place where they can experiment?
That's a great question. And honestly, it's something that I learned recently because my background, like before jumping into tech companies or staff companies in general, I had a five years of working at a real estate company. And when I tell you the word conservative, it's like not enough to explain it. I'm so conservative about things.
Every design, like I wouldn't say boring, but I still push boundaries there. And I want to say like this is something I learned recently is that we're not just, I understand the B2B, but at the end of the day, you're not literally talking to like an object or a business. You're actually talking to a human behind the business. And I would say we treat it still B2C.
Like that's like something that I think I learned and I think it still resonates with people. Whatever. It's like TikTok or whatever. We all kind of laugh at the same things.
We all enjoy the same things. Even if you're in whatever business you are, I think the idea is the same. So I always advocate for it in that way. There's levels I would say maybe of professionalism versus being playful, right?
Like when it comes to design or messaging, you can use certain terminologies when you're talking more to actual consumers. They would enjoy stuff like that or like it's curse words or whatever, right? Like some people just go for things and some others are like more professional, but I still wouldn't say staying conservative or like not trying these new things or that doesn't resonate with people anymore. I feel like definitely you're talking to a consumer at the end of the day.
So sometimes you need to treat it as B2C. Like I said, that's from a design perspective and also content. So for design, don't do like two playful stuff, but at the same time, just professionally at playful. And I feel like if you check out our website, like Zintree.com, it kind of has this good balance.
It's like nice. I don't want to call it playful, but it's like still talks to the consumer a little bit, but it's also professional, like legible, easy to read, all of these things. You're not doing an art project where you can have fun and be playful with it. So yeah, I noticed on your website that there's not actually, at least on the homepage, a lot of like large graphics.
You still have a lot of copy. And so it actually is, it's almost like you went on your advertising a lot more playful and a lot more edgy. And then you sort of dialed it back a little bit on the website. Is that what you would recommend do the weird stuff on like advertising or maybe even blog images, et cetera, or maybe even start on like, where do you start to do weird stuff?
Where's like a first channel? Is it ads or is it like, oh, let's introduce some of these maybe in sales decks? What would that progression look like of introducing some of this artwork and creative if you were going to in a more conservative environment? Yeah.
I would say like you kind of nailed it when you said. So blog, we do that. Like if you go to our blog page, you will notice there's more playful elements that were introduced. They are still part of the brand like guidelines that I did.
Like I designed design elements and they're part of it, but they're more playful that I would not use them except for ads blog. The things that like more like I said, like talking more to the consumer, well, you know, the homepage, I would never recommend like being too playful with it because this is like your window shopping where like someone is walking by. They just want to look at whatever they want to look at. They don't want all these things jumping at them.
But with other pages, you can start to be a little bit playful and like, you know, draw the line where basically like you said, if it's ads, be as I would say, like have fun with it because this is like trial and error. Me and Grant actually have like great like the way of doing it where we're just testing things. Like I think it's also the beauty of like being a startup in general where you can test things. We want to find out like what works, what doesn't.
So we're testing like even with colors, we're testing, you know, is it like a lot of copy that resonates with people or is it like an image and copy? Is the image of people or is the image illustration of, you know, whether it's people or like a drawing or something like that, we test all these things and we see again, there's no size that fits all of these ads. And we see like maybe if in an industry, like the faces of people resonates with whoever is going to click through it and actually lead to a lead generation or something like that. But when it comes to our target market, like we know what we're testing.
So we stop using something. If it's not, you know, there's no like click clicks, click enough clicks to actually convert or anything like that. We stop using that or we lean into maybe testing it with a different color with that work better, different image with that work better. And if not, you know, either stop or increase using it.
And speaking of that, like our AI generative like ads. So obviously we've been doing like ads for like a year now or something. And the click through rates, we like these they jumped the rockstar campaign, we just did them like maybe I think a month or maybe max two, I can't recall my brain is much ever since I had a baby, but a month or so. And these like jumped immediately to the top like CTA, like the CDR we had, like literally the rate is much higher than others.
I think there's only one left from the past, but these are now like the top. So yeah, I don't know if that answers questions. I think I went on. Yeah, you went off.
That was good. I wanted to, unless Cassie had questions, I wanted to go into like how are you, you know, measuring the effectiveness of these? Are they just fun? Or are you seeing anything?
I'd love to know if like people are like even calling them out in your inbound form, like saw your awesome ads. I would love to know what kind of quantitative things you're looking at. What qualitative things you're looking at to measure effectiveness. But before we jump into maybe that, I didn't know if Cassie had any threads you wanted to pull on here.
Yeah, let's go down the thread. But one thing that I would say in listening to this and in seeing the ad kind of without the context is I think you do a really good job. And I want to point this out because it's a marketing fundamental. The campaign strategy was really good.
And then the copy and the design was phenomenal in meeting that strategy, right? So what I want to get across is as you go into Carl's question, Omar is, it's not just go off and do this for the sake of doing it. The way you would set up from what the company's trying to do to what this campaign is trying to do to how I use this technology to enable that and then iterate on it, I guess is what we'll get into to prove out the results. That to me is the phenomenal work.
And I just want to kind of point that out from a marketing in like a branding perspective, really well done. And that's what caught my eye when I saw it, I'm like, this would catch my eye and a feed because it's a right on message. It's unique and it sticks out versus everything else I've seen. So well done.
That's kind of my two cents, but I'll just turn it over and kind of get in the nitty gritty of like, how did this one win Cassidy? Yeah. So like what Cassidy just said, like nailed it, right? You all like your product marketing is on point, like you have a message that's clear that you're aligned to and that message is like turn your call center agents into experts.
So you started with like a foundational premise and then it sort of got playful and it was like, Oh, what about rock stars and you just like kept going. But like that can seem like, Oh, cool. They had a brainstorm session and that was it. And like they just got lucky and somebody Omar thought of a cool idea.
But that's actually not true. What's true is that you started with like this really crystal clear direction and premise. And that is like 90% of the work. Because after that, it kind of made it a little bit more simple to then evolve into like, if you don't have that foundation.
So I guess maybe I don't know if that's something you want to speak to, but like what was the briefly the process of just getting to that great statement that you all have that kind of represents your brand of turn your call center agents into experts. Because that's very strong and clear and concise. And we're all marketers and sellers here. So we all know that that is a lot harder to get to that than it seems.
It's very challenging. It takes quite a lot of skill and time to get to just like, this is the thing. So can you talk a little bit about that journey? Yeah, it's definitely a journey.
And as the product evolves, we have to think about how our messaging and positioning evolves as well. But a key piece for us is having a positioning messaging guide that we're aligned on. And this, I try to keep it simpler. It's not like a 20 page deck.
It's like one to two pages of a Google Doc. To get to that was a lot of brainstorming within the marketing team and within the other cross functional leaders in the company. Getting there helped us then say, okay, how do we play this out across the website, across ads, et cetera. And once we had that, then we can think about these other themes as you were saying, like, how does it play instead of an expert like a rock star?
Or like after the Super Bowl, we did a football player where it was like turn your agents into MVPs. You can see how like you can continue to play on similar themes there. But to really get to that statement, that's kind of like a one liner, it was really like a lot of brainstorming with many people and we were able to align on it. Now if I was doing this today, I would probably also use something like chat GPT to ask for variations of this.
We've done this with, for instance, our homepage above the full text. So it's like, okay, we have the main statement and a few sentences underneath it and asking like, create five ways to rephrase this in a catching manner or in an amusing manner. And just ideating with chat GPT instead of just other people. That's awesome.
I mean, it's such a strategic advantage for your brand and your voice and everything that you all do because again, no other company is really doing that. And as far as your creative goes, what's striking to me is most of the B2B ads that you see is like there's a copy in them as the image. But if you were to take the copy away, the image would communicate nothing, right? It's like, you know, a professional looking gal in a sweater typing on a laptop, you know, it's like the classic B2B ad, right?
And it's like that in and of itself doesn't communicate anything. But if you were to take the copy away from your ads, there's, I still like, I actually my brain fills in the copy for you. It's like, oh, yeah, turning like the collars at a rock stars or something. And that requires, I think Omar would love to get your mindset here of like, how does a marketing team transition?
And like you just, you just care more about creative. You see more value in it than your average kind of B2B design or marketing team. You believe that the creative should tell the story and be, you know, self-sufficient almost in a way. But a lot of teams are just like, yeah, we'll just say something and then we'll just try to find a graphic to kind of plug in with it.
But really, we just want to click through to a piece of content. We don't take this creative here in the feed seriously. Have you always been like that? Or like, how did you make, like, how did you grow into like taking the creative elements individually?
You're going to talk about swapping out colors. Like, why do you care so much about these elements and how did you get there? It's a great question. So it's actually like on my end design, obviously I just want people to notice it, right?
Like, how do you notice this the most out of literally scrolling through? It's a term I actually learned from Grant. Like I always did this, but the terminology I learned from Grant, which is basically scroll stopping. But basically something that you're scrolling through this literally endless, like just everybody trying to advertise their product or whatever it is.
And as a designer, I'll tell you, I don't care about most products that I'm seeing, like on LinkedIn or on Facebook or on Instagram, but I literally stop when I see a good design, even if it's something I don't care about, I just stop and read it. And sometimes, sometimes I actually am interested. And I think it pretty much is everybody else. Like how do I attract people to just stop, like just stop and read?
Because that's basically the first step that I need. Because, you know, a lot of times, like if it's just a stock image, why would I stop? Like I would just continue, you know, scrolling pretty much. So there's a lot of loss there.
It's just like something that is boring, a stock image, something that is everybody else is using. And this is actually one of the reasons also we got to AI. Nobody's using AI. Nobody's doing like rock stars.
Although, I don't know if you guys saw that the Super Bowl ads like Workday used this literally the same exact, it was trippy, but they used the same exact like terminology is like, yeah, not everybody can call like, you know, whatever a rock star and the used kiss as well. So it was pretty weird. But, well, the thing about AI too is like you could actually put in depending on what day you like log in the mid journey or night cafe or whatever, you could put in the same prompt and get a different image, right? So it's like the originality is infinite.
I don't know if you can say that, you know, I don't know what is infinite or what it's not. Like, you can multiple people use the same premise. You still get a completely original photo or graphic that is extremely difficult to replicate. It's almost like AI, it's like when you're doing the prompts and it's generating the images, it's almost like time, like space and time is involved in that.
Like, you only have these moments where the AI is trained in this specific way to get images like this. And once this time passes, you cannot get that image again. I don't know if you've experienced that. I certainly have playing around, but I think that makes like, I think there'll be a skill and you think Omar that eventually there will be like a guy like you on a team that is just like an expert in AI prompting because it's incredibly difficult and way more involved than just typing in call center rock star and you get like a perfect image.
You know, it like takes so much massaging to get something that's like not a mess. What do you think about that role in the future? Let me add one thing to that before I jump in. This is something Carl and I debated in the past.
I would maintain it's very difficult for someone who isn't a designer to be able to do this effectively versus you being able to do it effectively because you're a designer. But I could be wrong on that. So I'd love you to kind of weigh in on that as well as you answer this question. Yeah, absolutely.
To answer Carl first, I would say no. There's no way this becomes a whole position. Like that's tough. I'm only saying that because once you get it, you kind of you can do it.
I see it more as a tool that we just add in our toolbox. That's it. I see it in job descriptions. They will say, you know, preferably knowing how to use like whatever, whatever is top tool right now, but maybe they'll say like AI prompts like, you know, you have familiar with these tools or have expertise in them.
That's the way I see it. Yeah, Cassidy, can you remind me for questions? Yeah, I think you answered it. It's a tool in the toolkit.
But I as somebody who does not understand design is not going to magically put a bunch of prompts into a tool and all of a sudden I'm a great designer. Yeah. But there's one thing about that. It's same way, like for example, when we run on small teams, like I've been part of right now, I am part of small team, but when you're part of a small team and you want to kind of automate some things as a designer, I create templates, right?
So whatever it's like Google slides templates that way, people don't have to reach out for me to design their templates. I just have everything. I have like their icons ready. There's like an entire resources folder that I created, even for a Zing Tree that has stock images ready to use product illustrations yet to use.
So that way they don't have to reach out for me because I already have a lot of stuff because again, small team. So I see it the same way in mid journey, you can actually do something in settings where basically the way that the prompt works is that you add two dashes at the end after whatever you type like, you know, customer service agent as a rockstar, whatever it is, and then you two dashes at the end. The two dashes after that, you can tell it to aspect ratio. So you can do AR three and then the two dots and then two, right, whatever.
But what I'm saying is you can the same way that they have these AR like aspect ratio, they have something that you can make your own prompt. So for specific image like that, I can make whatever the aspect ratio is, whatever colors it is, whatever the specifics of the prompt like 3d, whatever it is, I can make it to have one word to end up being Zing Tree, for example. So that way I already formatted a term, one term for people to use like anybody can use and they can go in mid journey and they type whatever they want to type and then at the end they type Zing Tree and then it ends up being the same exact style of the things that I designed. So eventually it could be something like that where it is a template, but like also Carl said, it's honestly, sometimes it's a, it's like this game.
It's like massaging it because sometimes the way I think about it, it's a really is like a, if you ever been in like in a foreign country where you know a little bit of words, they know a little bit of the words of yours as well and you're just trying to communicate with the little bit pieces of language that you know, AI is the same way. Some words to them resonate, some words does not. It's a learning process. I just keep trying the prompts and then when I find the word that resonated with them, I keep using that same word.
And also Carl, I know he asked about the prompts I use. I'm happy to share because I don't like keeping anything, but yeah, I can share that later if that's part of, Yeah, we'll just share with me, not the audience because I want to, I want to keep it and I want to. Oh, you just crushed Carl. Carl was hoping his next career was just somebody who did prompting for AI.
Yeah, I mean, you did ran on my parade a little bit, but it's okay. Cassidy knows so little about it that I could maybe still sell him on the roll. So I just can't sell you on the roll. Well, I'll tell you, it's like, it's like, maybe eventually it will be, but not yet.
I don't see. I do think that there's a skill, there's a skill involved. I mean, what you just said about the templates in like mid journey, I didn't know that existed. So yeah, that that greatly reduces or controls for somebody like me who doesn't have a lot or like Cassidy can jump in and actually get a lot more productive in mid journey.
You've defined if a world class designer has defined guardrails. I didn't know that that was a thing. So that does kind of that is a role killer, I think for me, but I still have. I will say, I think there will be some kind of prompt engineering roles, but it's, I see it more on the software engineering side than maybe the marketing and sales side.
Maybe there's like a blend, but on the software engineering side, we're in the process of integrating open AI into our products to add lots of capabilities into it. And to be able to do that from an engineering perspective, there is a lot more nuance involved, but I tend to agree from a marketing or sales side, it's probably going to be a part of a job description for certain roles that will become more popular. And I'm sure like, if you don't have the skill internally, like there's certainly going to be like an agency or like a freelancer ecosystem that's built out of this, where it's like, okay, we don't have an Omar on our team, but we want this done and we don't have the money to pay somebody like Omar, but we could go to like refine labs or some other like professional services firm and be like, hey, can we do a project and it's going to involve AI? And we're like, oh, yeah, we have somebody who can kind of fast track this for you.
So yeah, it's going to be interesting to see what ecosystem is kind of built up around AI. One technical question I had Omar and then we could probably leave this topic is what percentage of the way do you think the ads that we see on LinkedIn from the rockstar ad set? What percentage of the way once you were done with mid journey and you got the prompt? Like how much more did you have to do in like Photoshop or Illustrator or whatever?
What percentage of the way was was the ad done when you were done with on the mid journey stage of the process? Right. So I'll just add it might be helpful to share what the level of effort would take without using gender to AI compared to what it took all in with gender to AI plus. Maybe just a whole like instead of jumping into my question, give us a sort of a brevity like it's kind of like a whole technical process that went through and how it worked.
And then yeah, to grants point, what it would have taken otherwise. Yeah, absolutely. So let's start with that. Like how the process, you know, I started it.
So obviously a lot of trial and error until I got what is the prompt that will produce specifically, like if you noticed, it's like colorful. It's like they're, you know, the things they're wearing, like things like that, it just takes a lot of trial and error. How much though Omar, be specific with me. Like it was just hours, weeks, minutes, like lunch break.
Like how long did it take? Really. I would say like one sitting like, again, I literally did this sometimes while watching TV with my wife, like it's just something that I do because you know how the prompt takes time and journey like takes a little bit of like a minute or two to produce one. So I sometimes give it a prompt.
It worked went through great. I will upscale it, save it. We're good, right? Like it takes me, I would say maybe per image 10 minutes max.
The error is part of it is just like a button that you kind of refresh. You're like, I don't like this refresh. So it's like another two minutes. You just, it's literally waiting.
So there's not a lot of effort. I would say here, that's why I didn't like share time wise here. It's just like literally just, you just keep clicking a button until you see the right image. And then when you see the right image, you upscale it.
And then sometimes still there is the errors of let's say like there's a cup of coffee on the table, right? And that cup of coffee sometimes the table, this happens a lot for some reason that the coffee is not above the table. It's like in between like it's this for some reason. Like AI messes these things up and obviously hands they can not do hands.
Yeah, the store like some stuff is like distorted to a degree. Yeah. Well, the hands they do like seven fingers, six fingers. They just do all out on fingers.
It's kind of like real artists like a hand is the hardest thing to draw. It's like in their own fact and AI has the same issue. So yeah, sometimes when I take these things in, I refresh again, like maybe a couple of three times. And if I notice it doesn't get it right, still the hands are messed up, still the coffee is in between.
This is when I take the process over to Photoshop where I will have to reduce the amount of fingers and I will have to adjust whatever error is going on. And then unfortunately again, it depends on what am I fixing in the image? It will take maybe I would say an hour or two, but the amount of hours saved to Grant's point is insane because if I have to do again, let's talk about the agent on a computer on a laptop, having a headset, whatever they're wearing, like this is a rock star like Elvis, whatever they're wearing, you have to match with Elvis war, the hair, everything, right? Everything you have to match.
It's like entirely a new piece of art. And then you have to do that for all of the rock stars. So that you have an ad set, which is quite diverse in your assets. It's insane.
And then you think how we do it as designers, we take these pieces. So we start obviously with the walls, right? Like how is the scene behind him? What is the color of the walls?
You start with the walls, you separate it, you do a 3D. That's like maybe half an hour. And then you start adding like the desk, finding the right desk and the right angle that you want to put your agent in. And then you put the seat and then you fill that seat with the person.
And then maybe the person, the way AI does it, I honestly like have just amazing like 3D renderings. So you do the rendering of the person talking or I mean sorry, the person like with the headset or whatever, you start photoshopping the features. So you actually bring real images of Elvis or whatever, you put it on top, you have the hair. I would say it would at least take me like 2 days of work, which is like 16 hours.
So it's saving per image around like 16 hours, I would say. Since sanity. Yeah, absolutely. And I mean, that's not like, oh yeah, save a few minutes on your sales process.
That's not like what you just described is not better efficiency. What you just described is like completely ridiculous overhaul of an entire process from the ground. Like this isn't just efficiency, like save a little bit of time. This is like, like you just avoided catastrophic waste.
That's wild. 100%. I thought I was like the AI image would actually only get you like 25% of the way there. And then you'd have to do a lot of other doctoring and Photoshop.
So I thought it actually wasn't efficiency play, but you have debunked that myth. It changed your process and workflow almost entirely. Yeah, it's a lot. It's a lot.
It's a great tool in that direction. Hey, Omar, I'd love to get a little bit into how does somebody start doing this? And so for example, we keep mentioning this tool mid journey. Is that the tool is that there are many of these tools?
How did you pick a tool? Did anything to do with what you're trying to accomplish? I was walking through from like not knowing to like getting to where you were in this example. Sure.
So as you know, there's a lot of different tools. I settled on mid journey because the renderings resonate like they looked most of what I'm looking for every tool out there. So I know like Carly mentioned you're using like cafe. I actually have a new that one that's brand new, but how I started is the very first one that started this whole thing is like dolly.
So I started with that and I was not impressed at all my dolly. It was like, they almost looked like water paint all of the renderings. I was like, this is stupid. And I was just not impressed whatsoever.
And then I started using stable diffusion. And then I noticed that stable diffusion actually stole a lot of art and I wasn't just a fan of that. Like, you know, as a designer, I don't like that. So I did not use them.
And then when Brandon advocated for mid journey, I was like, I'll give it a try. But I never gave me journey. I tried because mid journey, it's one tool, but it lives in discord. So it lives on a server on discord where you go to the server.
If you're using it for free, you only have 25 prompts or 25 images to produce. And basically when you're using it for free, it's a server with literally insane amounts of people. And when you do a prompt, like I said, it takes a minute or two to do the prompt. So your prompt gets lost all the way up.
I have to scroll all the way up until it prompts and then the result, you can see it. So that's why I never tried it to begin with. I was like, whatever until Brandon, our CDO mentioned it. So I started using it and I was like, it's worth a try because the things that he showed me, I was like, no way.
I tried things before it wasn't that good. But I settled with mid journey because it resonated with that. But there are other tools like if you want actual realistic stock photography, I think stable diffusion does it better. But mid journey does it creatively.
Like this is the way that I would describe it. And maybe also car like you had the issues like you try to produce the same image that I could use, but the difference there is actually the data that the AI was fed. So night cafe probably is not fed the same like, like they don't have these same things that they resonate with the prompts that you're giving, but mid journey did because they have different data, different photography, like images or, you know, photography that were fed to it pretty much. Yeah, night cafe is like you, their value proposition is they use like dolly or stable diffusion.
They use like a combination or let you choose one of them if you want to see like the different variations from the different like learning models. So that's kind of like their thing. But wow. Okay.
I will try that then. Yeah, how much did you go in and try it out and then look for the next kind of campaign to run versus we're going to do a rock star campaign. And this is how we're going to execute on the campaign and then you went and figured it out. So I'm going back to this, like somebody who's listening to the same, what do I do?
I just go in and start playing around. Do I try to take an example where we're going to put something in market, et cetera. How do you think about that? I would say so mid journey is more complicated than all these tools because you don't just describe something like you have, actually when you ask the prompt, you do slash imagine and then you say the prompt and then like I said, you can do two dashes to tell it aspect ratio to tell it you can do one dash to say no, like if you don't want people in the image, you can do one dash say no people.
It's like interesting. So there's a lot to learn regarding what you can do with it and then there's settings behind it that you can control, which you cannot do that with any of the other tools. You can control the settings to tell it, like be super creative. There's a mode called my G mode and there's one that is V4 before is the latest one, but there is previous ones that are more illustrations type.
I'm using V4 because it's more realistic. It's like a 3D. It's like the ones you see in the as they're like realistic, but also like you can tell they're kind of like AI generated. But my advice would absolutely be try all the tools first.
I would be which one like the specific prompt that you tried first worked best with it. And then when you go to that tool, I will speak more for mid journey. I watched tutorials online. I would just sit in front of tutorials like how, you know, tell me what it is, but I found the best tutorial because people ask me also about that.
So I have one tutorial I can send it to you guys after we're done. It's super quick. It's like maybe 10 minutes max and he covers everything, literally everything that you need. All the settings that you can go through, all the ways that you can do things, all the prompts that you can give different like maybe sometimes he also like gives terminologies that you could use to make the image like better.
Like sometimes, believe it or not, if you don't use high resolution, it sometimes is not high resolution, which is weird. But sometimes when you use it, it actually becomes better. But yeah, it's interesting stuff that I would start with like just looking at tutorial, trying like the tools first, see which one works better and then see if it's mid journey, you need a tutorial. If it's not, the other ones are just like literally prompts and you can just trial and error, see what works, what resonates, what are the terminologies you could use that AI understands best.
That's awesome. We'll get that from you in and add it to the show notes. One more question for me. I'll let you guys close down.
Most of the time, what I hear about this technology coming from people who do creative work as an example is a lot of doubt. I was going to replace my job or I don't believe this will ever work and know we can write better copy or design something better than a human. And you just jump right in and start working with it. And you have this mentality that it's a tool as part of the toolkit, which I agree.
Is that where that come from? Is that just kind of like how you've always been? Is that come from Grant and the leadership team kind of giving permission to the rest of the organization to be innovative? Maybe it's none of the above and something else, but I'll be to kind of just touch on that.
And here's this all the time, private conversations and on, obviously, social media. Yeah, totally. I would say it started with us. We are just like, we wanted to be creative with it, but Grant, if you want to also jump in, how did we start?
You can go ahead. I think Cassie's your point earlier. There is a growing toolbox. And if people can view it as such rather than their job being replaced entirely, then they will have a much better opportunity in front of them.
I do think that job descriptions will start having things like know how to use mid-journey, know how to use chat, JPT, et cetera. And so being able to stay up to date is going to be much better than feeling like you're going to be completely replaced. Now sure, maybe some companies will hire fewer positions than before. For instance, take SEO for an example.
Maybe there will be less need for SEO-specific writers in the future because content writers can do their job faster, meaning they can produce more articles. But that doesn't mean the job is completely going away. So staying on top of the tools will be important. I love that.
Well, I just got to say, well done, guys. This has been great for us. The work is incredible. Make sure we link it.
It is legitimate. Some of the best ads I've seen on these platforms. And I'm sure compared to your competition, it's night and day in terms of what you're doing versus everybody else in the Colos Center space. So Carl, you want to wrap it up?
I was off to ask Omar privately, what he does with all the time that he saved. I'm sure he doesn't want Grant piling on more stuff onto his plate. So I'm just like, how do I work 20 hours a week but not tell Grant? It's eight more campaigns, man.
You can get more campaigns. It's cranking. Grant's making his d like fine. So I'm going to work 40 hours.
You're just going to produce a second thousand ads a month now. So the personalization, everybody gets their own unique ad. Everybody gets their own ad. Now do it for the sales team.
The sales team is probably asking for so much more stuff, useless assets for marketing now because like, hey, Omar can get me a new one-pager for no reason. So much faster now. We need all the one-pagers in the world. We got to call Carl out.
Every day. I'm like, can we get some designers in our team to do something for us? I'm going to go and say, listen to this. I'm going to set it up.
Look how fast this guy is producing ads. I think I need a logo or one-pager and I need it before lunch. I'm going to start creating my own stuff. This is probably the need that has surfaced my hunger for getting up to date on these tools is I'm going to start making my own.
You're going to get sales people that are now just creating their own content and then people like Grant and Omar are going to go nuts and hate it because none of it is going to follow any grant guidelines. I do want to add something. So to your point, like logos, for example, none of the tools can do it. None of the tools can get copyright and none of the tools can do UI design right.
There's one tool for UI design that actually grants us like last week. It's still like beta version, but you can sign up for early actives. I think they're called Galileo or something like that. That's the only tool that I was like, oh, interesting.
I haven't seen AI being that smart, but with everything else, for example, Super Bowl Party, I tried to design like a quick. We had a party. We were hosting a party here. So I wanted to design a fun poster for the Eagles, right?
We're Eagles fans. So I typed Eagles poster and I swear all that all the words, none of it got Eagles, right? It's like, it was even random words like Calabolos and though, I don't know, Go, Calabolos. I was like, what is that?
I made that joke. So I go, because we didn't understand what looked at. A lot of stuff. And this is why I'm saying, again, I say like, this will never replace designers.
I think there's a lot of designers worried, but maybe they're just like junior enough that they don't understand what's happening or they didn't use the tool, but because I use the tool, I know that it's never going to replace designers. The only thing maybe headcount would matter a little bit because it's a great tool for ideation, great tool for making things faster, but there's still needs a human to operate it. I think I saw it's we basically comparing it to it's like saying a new power drill that makes a carpenter's job faster. You still need the carpenter to do it, right?
You seem to need a human to operate this tool. And it's the same way. The only things I'm worried about, I'll tell you, is Chad GPT. It's dangerous.
That one is dangerous. Because I do copy now on it and it's super, super smart in advance. There's nothing like go like, Grand Set SEO, I do all that now. I can do copy add.
I can do everything. I can do pretty much a writer's job. But again, a writer could be more creative because something about Chad GPT, and I'm sorry, I'm talking a lot, but that Chad GPT actually, eventually, if you kept re-asking the same questions, it will actually give you sometimes the same answer. So it's still very narrow, but this data will grow.
So that's why it is dangerous. I'm... What about the depth of subject matter expertise, though, Omar? Because I mean, a lot of the...
They're still there, fact... So I see two things. It may be close with this thought, but like the design element, the design side of our conversation, the creative side, it requires the hard work of messaging and core positioning and strategic narrative to be done. And I think the same...
So that's on the creative side, right? You can't go get to make your agents call center rock stars. You can't skip the step of we are making call center agents experts. You can't skip that step.
If not, you just have like random creative floating on the universe. And I think the same thing is going to apply to the copy side. It's like, you still need the base layer of subject matter, like stone cold. I know what I'm talking about.
In order, you can't skip that step and really maximize your usefulness of chat GPT. Do you agree? Is that true? Or...
100%. I... But I think the only thing that I'm saying dangerous is because chat GPT is closer to doing that job than design, like design is still like a little behind like the journey and all these tools are a little behind to what a designer can do, but chat GPT is closer and closer. And I'm worried that this is the beginning.
That's... I think this is what I'm saying is dangerous because it's the beginning of both of these tools starting. And this is their start and this is the data that they have. Like even chat GPT is updated to 2021.
This is just... They're just starting. So imagine in the future where they can take in all the knowledge that it can take more than that human can, right? Like this thing can do what hundreds of us can do pretty much.
I think that's my worry that this is only the beginning. So you're worried, can it actually elevate more or this is it, right? Like this is just as much as it can help you or would it be better? I mean, you have to believe that it's still like super archaic, right?
In its current state, which is scary because it's like, you know, when you just got started with like computers back in like the 50s or the 60s and now it's like you have a computer on your wrist and it's only been what? Less than a hundred years. I know to think that this is the starting point for this technology. It's already wildly impressive.
Yeah, I can see how it's citizenry. And then the ethics of it is something that we can talk about all day. But this is something I'm hoping, you know, some sort of legislation eventually that say like how do you... How do you...
They have to kind of show the data that they use. Do they provide... Do they pay people that they use their, you know, whatever articles that the stuff that chat GPT gets this information from? We don't know.
It's very much against all and work, uncredited work, right? And very much the future of it, I also thought about it actually the future of it is we all work for these two giants, right? Like whether we provide the data that feeds into them and we become employees at these AI companies that just automate the work for everybody. So even for designers, we design specific illustrations or prompts and it, you know, people, regular people can just use these tools that designers were behind, if that makes sense.
Yeah, makes complete sense. And that is the topic we should cover at another day. I always think when I see disruptive technology like this, we overstate the short term impact and we always underestimate the long term impact. And this is going to have a big impact at some point over the horizon.
Well, the one thing I did take away from this, Omar and Grant, is that for now our design team doesn't need to worry about Carl taking a replacing their job. So they are safe even though he is going to try. I'm on the journey. So they better watch out.
Alright guys, well, hey, thanks for the time. This has been great. We love to have you back and we look forward to just watching your journey from afar, if nothing else. So really well done.
Talked about him. Super impressed. Yeah, same team. Excited to watch your mid journey.
Uh huh. Get it Cassidy. Get it? Well done.
Yeah, great talking to you both for all those marketing leaders listening or sales leaders. I encourage you go get your team a $20 subscription to mid journey or another and give them some time to try it out. Well said, there's actually no excuse not to do that. Well said Grant.
I love it. Thanks guys. Alright, we're out. Another episode of Second Room.