29 | Eliminate Marketing Bottlenecks with AI Visual Asset Creation with Global expert Rory Flynn shares his step by step process

24:41It's not like I'm actively seeking out Delta 8 for major health benefits. There is a little bit of a feeling that's associated with it when you smoke it, eat it, whatever. So there's something here that we want to do is bring a little bit of life to the emails and something like when in this case, Mood, they had a very colorful brand palette. And we were like, let's see what we can do to just pump some life into these emails because they already had a big list. Basically, what we want to do is A, B test AI art against just their stock photography or regular brand photography and to see what the outcomes work.

25:15Because it's, let's see if we can actually create a little visual aspect. Because in email, really what we're trying to do is just get the click. At the end of the day, that's all we're trying to do. It's not that you can't purchase anything physically in an email. You have to click and go back to site. So we're a medium between point A and point B. So effectively, that's what we're trying to do here is just generate interest, get people to click, move them back to the website, and then we've done our job. As email, that's what you can do. Now, Now, naturally, there's other ways that to think about email, you can educate on there and whatnot.

25:47But again, at the end of the day, an abandoned cart email, you just want to get someone back on site to purchase, right? So things like that can be super, super helpful. So looking at this, we were able to increase click-through rate fairly significantly in testing just based on some of the unique ad creative that we were using. So first, the unique email creative that we were using. And again, going back to your point about feeling something, feeling a mood as aptly displayed here. Okay, what are we trying to convey? We want to convey a nice sunset at the beach with maybe a little transcendental effect there.

26:19But trying to get that or take a picture of a lifeguard stand, which is what is on screen right now with a very neon colored background. You might have to go to a stock image site, find it, tinker with it in Photoshop for five or six hours to get it to look like this, then actually make it fit into the email. So that's not going to be the case here. And on mid-journey, you can also prompt an image and then change the aspect ratio. So if you like something, you get it. You can then change the aspect ratio. So if it's a square one by one, you can then change it to 916. So it'll fit a phone screen.

26:55Very easy. That again would take, without the tools of the day, that used to take hours and it would look messy because you just, you have to copy paste things. You have to transition things. You have to use different types of blending tools. And it's just, if you weren't a professional, you couldn't do that. So it's things like that are super important and thinking about, again, just the concept of this. How can we use it? Where are the opportunities? And this one is a little bit... I want to pause you before we go to the next example. There's a very interesting question from Anton. He's saying, how do you handle the low image resolution that comes from MeJourney outputs?

27:28I would assume that works fine for email marketing, but what about print advertising? Simple. So there's a number of AI upscalers that you can utilize. They use the same technology. Basically, you can get it anywhere from 400 to 800 % of what the initial output is. So that'll typically get it into the 300 dpi range. And then from there, that's basically transferable, right? So that's typically how I look at it, because there are really good tools. It's called IMG Larger. That's one tool that I use pretty consistently to have a free trial. It's great. But typically, I upscale my images 400%. and the AI will fill in the pixelation and absolutely correct the distortion if it gets bigger.

28:11So that's a really good question. And that's definitely one way to do it. I mean, you can do that for print these things as big as 10, 15 feet before. So it's definitely scalable in that sense. There's a lot of tools that you can use. And another question, thanks for, first of all, great answer. Again, there's a bunch of upscalers. Some of them are free, some of them are paid. The paid are obviously better. They're not crazy expensive, even the paid ones. I used Topaz one, which is incredible. And I think a lifetime access to it is like$100. So in the scale of a business, it's completely negligible.

28:43Speaking about that, there's a question from Kevin about the cost of this. I'm not sure I understand your question, Kevin. So if you can, he's saying they hesitate to do$60 a month for whatever reason. Can you explain$60 a month for what? Because me journey is$10. So I'm not sure I understand. Yeah. So I saw something. If you're over a certain revenue threshold, do you have to go to the corporate level? Interesting. Maybe I was wrong. They introduced their, I mean, there used to be a free version of it and then they introduced their tiered section. I think most of it's based on output. That's my general understanding.

29:16I haven't looked at their terms and conditions in a while, but typically I know I'm on a$35 a month plan and I've never hit the time thresholds based on or the output threshold. Now, I know there is a larger plan that they call their enterprise, right? I personally haven't used that because we have all of our designers on individual accounts. So it's a little bit different because we want to keep, instead of having everyone jammed up on one, we want to spread it out from a designer standpoint. So that's where it becomes a little bit more of an ROI conversation, right? Because I look at this as depending on what the pricing is.

29:53And again, I apologize for not knowing the actual enterprise pricing of this. But realistically, if we are doing, let's just say$30 a month, that's$360 a year, it can definitely net you a significant amount of clients. I know it goes up to around $140, I believe. But that's also, even if we're looking at anywhere from$1 ,000 to$1 ,400 a month for a year, then can we generate that type of return off of it? And we've been able to just based on not even some of the tangible aspects of it, the new clients that we've acquired because of it, but just the time that we're saving. It's an interesting conversation.

30:33And I want to touch on two sides of that equation, right? One side is, I think it's incredibly easy to prove the ROI to whoever the decision maker is, whether it is, here's how much designer time we're going to save, or if you need one photo shoot for anything, one is going to cost you 10 times the annual cost of this. So that completely makes it a lot easier to prove, forget about 10 times, a hundred times more, because you can't get a professional photo shoot for thousands of dollars. It's going to be a lot more than that. So that's number one. Number two, the flip side of this is your ability to play with the assets.

31:07And again, that's what we're going through these use cases, right? So the use case that we're going to dive into right now had a 73 % increase in click-through rates over three months, 73%. That's almost doubling the amount of potential business you're getting through the funnel, just because you can generate better assets faster. So both on the saving side of how much designer time or production time you need, and as well as the opportunity costs on the other side of I'm winning more because I can be more original and do more stuff that I couldn't have done before. On both ends of that equation, you will win with your ROI.

31:48So I think it's a no-brainer, but again, I'm not the CEO of your company. So I don't know, but let's dive into the next use case. Yeah. So this one, the interesting use case, this was one of our clients. They sell gold and silver coins. Now they're a mature business. They've been around for a long time. They're not necessarily like super brand focused. They don't have the flashiest advertising. They don't have the flashiest brand colors or brand palette. So for us, most of the time we were sending plain text emails. That's at the end of the day, they work. There's nothing wrong with them, but can we give this a new life, a new flair, a new look.

32:21Because that's really where, again, we live in a world of testing, especially in marketing. Everything is testing. And for us, now we have the ability to A-B test against each other. And we can send things that have been generated by AI versus things that are just the plain text side of things. Now we can actually see what works better. So in this case, yes, we had a 0.73 % click-through rate, which is fairly significant from a click-through rate standpoint. point. But again, all we're trying to do is give a little bit more life to this, bring a little bit more visual aspect because there's nothing wrong with plain index, but can it be better?

32:58That's essentially where we keep looking at different ways to optimize this. So even with that, you see it here. If you can, again, for everyone that's listening, basically we just created a picture of a coin. We just created a picture of a coin. Might not be anything that is totally, completely relevant to the topic, but it's just a visual aspect. Now, the other thing that we have to talk to our clients about this too, right? This isn't a one-way street where we just create and don't tell. That's not how this works. We have to get buy-in ahead of time. We have to be on the same page about this because there are definitely ethical and legal concerns that we have to review together internally and actually come to an agreement on it because if we're going to use it, we want to use it the right way.

33:39And things like this, I think from an email standpoint, think about it in any other sense, social media, or ads or whatnot, maybe not print ads, but digital ads. They're like news headlines. They're here today. They're gone tomorrow. If we're able to pump this stuff out, there's really no way people are going back six months to look at my promotional email that's stuffed in their promotions inbox in Gmail. So it's like, how can we amplify the production there as well? So we're not just so focused on perfection. And I I hate to be that guy, but not everything's going to be perfect. You're going to have more data the more you shoot on goal.

34:15And the more you shoot on goal, you're going to get better results. And then you're going to be able to iterate off of it and just actually continue to tailor. So that might be an interesting perspective for some. That's how we look at it. But again, this just gave us these emails before and after look significantly different. And there was also a production value that was added to this. Awesome. I'll touch on one of the questions from the audience and then I think we'll skip to the how because I think that's what a lot of people are here to see. But the question that was asked by Carla is where is me journey pulling the photos from?

34:47Can you be sure that you're not infringing any copyright or that your image looks different from the competitor's marketing material? So go ahead and answer that. If I have anything to add, I will add as well. So there is a, now I will try to be as clear with this as possible. Now, I'm not a lawyer in terms of how everything really functions from a copyright legal standpoint. What Midjourney says, if you pay for a subscription, you're covered under their copyright. So that's how we operate too. But ethically, there's a few things that I do to make sure that we stick away from any sort of copyright infringement.

35:25One is we use a reverse image search. So anything we're going to do before we put it out, we throw it into a reverse image If something comes up that's very close to it, then again, we nix that right away. That's a no-go. So for example, if anyone remembers the photo from many years ago, I don't know if you do or not, but there was a National Geographic cover. It was Iranian refugee, I believe. That was, she had a redhead scarf. She was very beautiful. She had like piercing green eyes. Amazing blue eyes. Yeah, green blue eyes. Piercing, like stunning photo, won a ton of awards. I looked that up.

36:01I looked up the photo. I looked up the specs. I looked up the photographer. I put that all in the mid-journey and it spit out the exact same image. So it can happen, right? That's why we use the reverse image search. So if you use the reverse image search, there's a number of them online. One is called Duplichecker. Like you can plug it in. It'll search all the image search databases or at least the major ones. And you'll be able to get a good understanding if this image is how closely pulled it was from something that was sourced. My understanding of how it was sourced is there is a lot of training.

36:29They do a lot of training on these images, probably in the billions of images that were sourced from wherever they were pulling them from. And basically, they're just they're piecing together information the same way that ChatGPT pieces information together. It's picking and choosing different things and then molding it all together if I can describe that as as easily as I possibly can. But that's two ways. And then the other thing is really like understanding how because I do this with my team. And also, if you're doing it yourself one way or the other, you'll see a lot of information out there on or on prompting, if you've been on Twitter, if you've been on LinkedIn, anything like that.

37:05People will use in the style of. That to me is a no. I don't want to cross that boundary. I don't want to be in that pool. So that's where you'll see in the style of Wes Anderson or in the style of Michael Bay or in the style of Picasso even. To me, that's just, if we're going to put this out consumer facing, not touching that because that's just crossing some ethical boundaries that I don't even want to be involved in. But also if you really wanted to know what Wes Anderson's style is, you can look that up. You can find the cameras that he uses. You can use his color schemes. Just to me, using someone's name like that is not the way to do it.

37:41So I personally try to stick to our own style. Our brands have very specific styles. So we try to stick to that as well. But again, there's always room for interpretation on both sides there, but that's just how I feel that. Awesome. Let's dive into the how, right? So we touched on multiple use cases. I think everybody who's listening or that is here on the call right now understands now the benefits and the different ways you can use this. And again, we touched mostly on email, but this can be used for any visual asset, right? You can use it for your website. You can use it for print. You can use it for your brochures.

38:15You can use it for basically any visual asset your company needs. you can produce with this in, again, minutes instead of days or weeks. And so let's talk about how you're getting to that level of quality and consistency with your mid-journey prompting. And there was a question here about lighting, but I'm sure we'll touch on that amongst other things that we're going to talk about. Definitely. Yeah. Again, we use this for a number of things. I've used it for a number of different mediums. Anything that you can think of where there's a creative need, there's probably a use for mid-journey. And if it can't do the full job, there's a way to just edit it very quickly to make it work.

38:53So I think that's, again, how we want to think about this. And really, the art is in the prompts. That's where everything happens. The artist is the prompter. Really, what you want to think about is having clear and direct prompts, because that'll give you clear and direct output. If you have ambiguous prompts, that'll give you ambiguous output. What I mean by that is, if you were to write, think about if you're telling someone a story, right? and you say, the man got into the car. Is that a compelling story? And basically anyone is thinking about the man getting into the car, their minds fall over the place.

39:26But if you said it's an elderly man getting into a car very slowly at night doing this, the story builds, it starts to feel some emotion, but also you have more control over it. So what I mean by that is an ambiguous prompt. If you just said man in the car onto mid-journey, press enter. You're going to get something that's all over the place. Every image is going to be different. It's not going to be any sort of vision. If you have a clear and direct prompt where you're saying, the man's getting into the car, it's night, it's a Nissan Altima, he's an elderly man, the light thing is dim, and that's your vision, and you prompt it that way, that's what you're going to get.

40:03So that's kind of the way to think about how you want to prompt it. You want to tell it a story. You want to set the scene. Realistically, we need photorealistic images. We don't need a ton of futuristic stuff unless that's what you are looking for with your brand or product or service. But the best way to do that is to really put photorealistic elements into your prompting. That's the number one thing that I've seen has worked the best. There's a lot of tools out there also that you can utilize to help on prompting that. You can use ChatGPT. You can use some of the other prompt engineering tools that are out there.

40:36But typically, you just want to have some photorealistic elements in there. So this is essentially how I structure majority of my prompts. There's the number. There's about eight different categories that I like to list in here. You don't have to go crazy. You don't have to write in a novel. It can be just keywords. It can be very powerful keywords. So typically, I like to look at subject. that's whatever we're prompting right that's the person the car the dog whatever that might be you want to always illuminate that somewhere towards the beginning of the prompt so mid-journey knows to figure to really like trigger and understand that the other things to style and technique that's really again are we doing this in the style of something that's more is it surrealism is it actual like photo photorealistic is it a fashion photo shoot is it car photography that stuff can really help trigger it again to understand where it needs to go.

41:34Mood and atmosphere, definitely important that we want to add these things to. So mood and atmosphere, again, tell the story, set the stage, what's going on? Is it a happy scene? Is it a sad scene? Are we excited? Is it vibrant? However you want that to be, definitely translate that into your prompt. Some details modifiers, this is really where it gets intricate. It's where you want to have details, right? So if you want this at sunset and you want a lens flare or you want all this stuff, you have to add it or else it's not going to do it itself. So those are important, but really specific cameras and lenses work well too.

42:11So if you wanted to add, you want to be super professional, you want to use like a Leica M10 in there. It's a really great camera. It costs thousands of dollars. but you can tell if you go and sift through different camera options on mid-journey and by that like actually putting in different camera options and trying it they have a different look and feel so if you put in a leica m10 versus a polaroid it's going to look significantly different the polaroid is going to come out looking like a polaroid picture would have back in the day and m10 is going to be super high detail high contrast like very good so the cameras they work very well i know there's a big debate about it online i tend to use it it typically works better for me so i'm just going to say that it's that it's great aperture this is something again i don't i didn't know how to properly function in aperture until i started doing measuring that's actually amplified what i want to do in photography but either or aperture is just depth of field so if you take in if you have an iphone you've taken like a an image on portrait mode for example like how you your face shows up super clear and the background's blurred like there's a number different settings in the aperture that can help you do this.

43:21So I've sent away, I put a little cheat sheet here. I can send this to every, hopefully I can get this out to everyone at some point so you can utilize this just again in your prompting. That'll give you some different apertures and what the usage is for. Yeah. I think what we'll do is we'll somehow post this or link this in the show notes. So anybody who will check it on the podcast will have access to this cheat sheet, which is awesome. Totally. Yeah. It's a lot to learn, but again, this will just make things super quick. Then lighting, big deal, right? Like lighting changes everything. If it's something in midday, great.

43:50Feels totally different than something at an ominous like sunset. There's a lot of different lighting conditions. There's lighting techniques, there's lighting styles, there's different environmental lighting. So this can give a whole new look and feel. And again, what we're trying to do is get some emotion out of these images, right? That's really what we're looking at. And then composition, perspective guidelines. Composition, perspective guidelines. Again, rule of thirds, if you're not familiar, things like that, like basically setting the image up into three different sections vertically and horizontally so that there's a good composition within the shots.

44:21You don't want someone to be like here. If you can't see me, I'm ducking down in the corner because then that doesn't make sense. You want someone, if you want a picture of someone in the middle of the screen, you can basically prompt it to do that. If you want them on right justified, left justified, side profile, these are all different things that you can control. And that's really what we want to get to is a level of control so we don't have to waste time tinkering on here. So I want to pause you on a few things you said to help people out in this process. One thing is we talked about a lot of photography stuff, like what cameras, what lenses, what's aperture value, speed, the right lighting, and so on.

44:59It's very helpful if you know nothing about photography to take a quick review of what these things mean. just take basic photography details on Google and you'll find a gazillion videos and so on to give you an idea of what these things mean. The other thing, which is maybe the coolest thing about Me Journey is Me Journey, and it's the coolest and the worst at the same time bundled into one, but Me Journey is not just, you don't just go to MeJourney.com, you got to go on this chat interface, right? And that's the only way to use Me Journey right now. They said that Me Journey 6 will take that away.

45:37We'll see. I don't know if Me Journey 6 is planned, but that's like the rumors right now. But that means that when you look at the chat, you can see literally every single image that you see, what was the prompt I produced it, right? Nowhere and anywhere on anything in history, there was the ability to open the kimono of the outcome and see how that outcome was created. No photographer, unless you're in like a geeky photography group that tells you, okay, I use that lens and that camera and that thing, and they share all the details. Usually you see the outcome. You see the ad, you see the poster.

46:15Oh my God, how did they create this? Nobody will tell you that. But on Me Journey, it's out in the open. You can see every single prompt of every single image on the server you're on. And so it's an incredible way to learn beyond trying it yourself. So Rory said, okay, try it. Run the same exact prompt and change cameras and see what happens. Change lenses and see what happens. Change the lighting and see what happens. But in addition, you can look back and say, oh, I like this one and this one. What did these people prompt as far as cameras, as far as lighting, as far as mood? And you can learn from that and copy.

46:47And then you're not starting at zero. So you're starting at step six because you can literally copy people that you like the style of what they're doing. So that's a few small tips on how to get started if you haven't done any of this before. It's a great point. It's a totally great point is learning from the community too. There is just a lot of good output in the explore section of MidJourney's website. And there's a lot of good elements to pull from that you can build and put into your prompt database. So that's a really good point. The other thing that I will get into now is the prompt structure.

47:19Now, basically, I gave you the outline with what I'm doing. And here's a prompt that I will, in the same structure as I typically use, what we want to do is use powerful keywords. There are, you can use more natural language. I tend to use just powerful keywords to try and get my point across to what, to mid-journey and how I want it to look and feel. There's a lot of prompts, tools out there. They spit out paragraphs. They work. It's cool. But if you want to be able to control your images too, I suggest that you start here and build from it. So this example that we're going to run through, and I apologize for anyone who's listening.

47:57I'm going to show you some images on the next slide. But BMW, we're going to look at this prompt as the BMW M3 in the style of matte black, a bustling city night, intricate reflections, Dutch angle, 35 millimeter camera, F1.4 aperture, dramatic lighting, leading lights. So we've set a scene pretty well, and we told it exactly how we want it to be shot. We told it the angle that we wanted it at. We told it some details. We wanted some intricate reflections. So now we can see what it produces. And now you have these images. Now, the image number one, I'm going to talk about prompt structuring because all of this is important.

48:34Image number one is basically the exact prompt I just read off. It's great. And it comes out, and it looks exactly like the car that we described. It's a BMW M3. There's a city. There's reflections. It's matte black. Right? It looks good. The emphasis is on the car. Now, when I talk about this, I try not to be too confusing. Mid-journey weights the front of your prompt more than the back end of your prompt. So the front of your prompt, whatever you put first, that's what it's going to focus on. So you put intricate reflections first. Mid-journey is going to weight that heavier than if you put intricate reflections at the end.

49:10So think about it in terms of a sliding scale. Right. And the beginning is going to be more important. The end is going to be least important. So what I did here is you can, I took different elements of the prompts. I just switched up where they were. So I put in the second image where it says intricate, intricate reflections. I put that at the, as the first keyword. So it was intricate reflections, BMW M3 in the style of Matt Black, so on and so forth. Same thing with bustling city. And then same thing with leading lines, just to see how it would affect the outputs. Now, in Image 2, where intricate reflections is prominently displayed, you can see there's a lot more focus on intricate reflections.

49:47You have it on the card, you have it on the ground. There's some in the background too. It took that way further than it did in the first prompt. In the third prompt, we have Bustling City. We listed that as first. Now, the card pulled back a little bit. You see a little bit more of the city. The city becomes more of the focus of Mid Journey, and you're able to now control this. right if you wanted done if you wanted to say i want more of the city all you had to do is just switch it to the front and you got more of the city so like that's again conceptualizing and these prompting these little small details make a lot of difference and then leading lines if you're not familiar with that technique it's basically just horizontal diagonal vertical lines that guide the viewer's eye to the image or subject that you want them to so again if you can't see image, I apologize.

50:35There's lines going vertically, horizontally. Everything is basically focused on the image, which is right in the center. So all this was taking one element of the prompt, putting it right at the front, and then letting mid-journey do its thing. So again, this is just an example. When you build these prompts, what to look for. If you're not getting something that you like, maybe it's like, I asked for intricate reflections, didn't get intricate reflections, put that in the front, and then see what happens there. So that's, I think, really super, super important to this. Details matter. And this will definitely increase your prompting capacity to get better results.

51:11So that's my little soliloquy on prompt structuring. Oh, it's awesome. And I want to answer something for those who are not in the chat. The question was where, because I said before, you can see other people's prompts in the history. So for those of you who haven't used MeJourney, MeJourney runs on a Discord server, which is just a third-party chat platform. So you need to go to Discord and sign up for Discord and then look for a Me Journey server. It's not that complicated. You can just Google that and you'll see exactly how to do that. It's not that complicated. So think about a chat interface where what you see in the chat interface is the prompt that somebody wrote and then the image that came out.

51:51And then the prompt that somebody wrote and then the image that came out times the number of people that are active in that chat that time. That could be hundreds, that could be two, depending on how busy the server you're on and there's multiple servers you can go to. It sounds complicated. It's really not. Again, if you just Google that, you'll see it's really simple. So that's where you see that. And now there's another question that says in the BMW examples, some of the cars are facing right and some of them are facing left. So in this particular case, it's not in the prompt where you saw the actual prompt.

52:20So MeJourney decides that. If you want it facing right or facing left, you can say that. And in a lot of portrait stuff that I do, I tell it, I want a slight left portrait and it will give you a slight left portrait. So it gives you, now you can always then take it to whatever other editing software and flip it around, but why would you do that if you know what you want and it fits exactly what you need on the website or on the brochure or whatever, you can define any detail and that's what you're going to get 85 % of the time. So we need to remember that's at the end of the day, it's a statistical model.

52:52And sometimes you won't do exactly what you told it, but then you can play around with the prompt and get probably a better outcome than you did the first time. That's a perfect example where I said before, where you leave room for ambiguity, mid-journey will fill in the blanks. So because I didn't define that because I wasn't necessarily looking for it to go right or left, that's where it filled in its own blank and was like, I'm going to do it however I want to. So that's why I'm saying, if you go and do a shorter prompt, it's just going to take creative liberty and do what it wants to do. So if you want to take control of it and you want to have very defined outputs that look exactly like what you're envisioning, just be as specific as possible.

53:31And the other feature that's in there that will help you with this too, it's called the shorten command. Basically what it is, you type backslash shorten, S-H-O-R-T-E-N, and you copy paste the prompt and you run it. And when you run it, essentially mid-journey will tell you what the system picked up from it and what it placed the most value on. So if there's certain words that you put in there and it didn't really make an effect, get rid of those words. That doesn't affect mid-journey. Or if you want to switch the prompt up in the structure of it like I was talking before and then run it through again and see what happens, it'll effectively tell you what it focused on and what it didn't.

54:08So you can use that to just refine your prompt a lot better. I wouldn't say you have to go and use all of the options that it produces because it will technically actually shorten your prompts too and make them shorter and give you different options. Sometimes it leaves out key details. So to me, looking at it and finding the actual weight of what it focused on is more important to me than actually just redoing the prompts for you. So that might sound a little technical, but it's a good tool. It's a good trick to actually get through the prompting exercises and get to the point of where you want it to be.

54:40Brilliant. I want to do a quick summary and then we'll open it to questions. First of all, amazing, Rory. This was really fascinating. I want to go back all the way to the beginning. The first thing is there's a real clear ROI across any use case you can think of that requires actual visual resources. And so learning how to do this in a business context of, I need to create images for literally anything you can imagine, you will get an ROI A in the production time, B in the life cycle time. So I want to send a quick draft to my VP or my CEO, whoever this needs to go to, my client, you can do that.

55:22And it doesn't take weeks, it can take minutes. And you can go through all of that very quickly. And like we said, on the other hand, because it gives you a lot more creative liberty very quickly, you can get better results with whatever it is that you're trying to do by A-B testing stuff that otherwise would have been very difficult or cost prohibitive to do. So that's on the business side of things. On the production side of things, it's about really having a very detailed vision in your head. If you can have a very detailed vision in your head, now it's about technique that Rory just shared with us.

55:59It's about how to translate your vision in your head into a very detailed prompt using specific keywords, focusing on lighting and cameras and mood and all the stuff that Rory went through, which I'm not going to repeat. But if you can do all of that, you get to a very advanced state very quickly. And from there, you either iterate or Rory said you can grab it, now put it in Photoshop, make a 20-minute session instead of an eight-hour session, and you're good to go. So it doesn't replace the full production process, but it takes you to a very advanced state significantly faster than literally anything else we have seen before with a lot more capabilities.

56:42Because let's say that BMW image, and those of you who are listening to a podcast, it's a BMW image driving through a city at night. That requires blocking a street and putting lighting in the middle of the street in a bustling sitting, which is not something that's easy to do. You got to book it in advance. It needs to be policing. There's a lot more to it than just... So if you wanted to do this in a real photo shoot, you either take an image of a BMW and then do a hell of a lot of Photoshop work, planting it in an image of a city that you have and try to get the reflection right and the motion and so on, or doing a real photo shoot, which like we said, is probably cost prohibitive unless you're Nike.

57:23And then you're good to go. Now you can play with those around and you can change it and you can add it and you can add text on top of it. Talking about text, by the way, it's the only thing that Meet Journey currently doesn't do, which means all the text you want on top of that has to come on like an after step of putting it into Photoshop or whatever, either editing software. Anything to add to the summary? And I'll open this to questions if people have questions because we're at the top of the hour. No, man, that was pretty good. You did a really good job with that. Awesome. Thank you. So James is asking, what was the program that you were using to create the layouts of the post emails and landing page at the beginning of the Zoom?

58:03What was the program that you were using? So for the presentation or for the actual like emails, because for the emails, I can answer that pretty quick. We use Figma to lay everything out. So if you're not familiar with Figma, it's a web-based Photoshop platform. Essentially, you can share your images across multiple workspaces. it's really clean. It's super easy and it's an outstanding tool. Yeah. And it's a freemium kind of thing. So you can start for free, play with it. And if you're getting to professional level, or I don't even remember when you start paying, but at a certain point you start paying, but it's just like any other SaaS tool.

58:37The payments are very reasonable if you're a business. Okay. First of all, I really want to thank you, Rory. This was absolutely brilliant. I follow your work online all the time. You don't stop to amaze me. So every time I see one of your images, like, oh my God, like, why can't I do this? But it's all about trial and error. And we're just giving you all the basic. If people want to follow you, learn from you, work with your agency, take courses, if you're doing, what's the best way to connect with you? Yep, man. So LinkedIn is the easiest way. I'm on there, Rory Flynn, R-O-R-Y-F-L-Y-N-N. That is where I live most of the time, as you can probably tell.

59:15Super easy. I'm pretty casual about this. If anyone has questions, you want to reach out in the DMs, I'm always down to problem solve or kind of troubleshoot. As long as it's not super labor intensive, it's going to take me 10 hours to do it. Yeah, I'm always down to help. But yeah, if anyone, I'm always around and always looking for new ideas. And if people need help trying to inject it into their businesses too, totally down to wrap about it there, see what we can do. Yeah, man, it was great. I appreciate you having me on and talking about robot art for about an hour here. So it was, no, it was awesome.

59:45I can't wait to do it again. And we will, I'm sure. I'll say one thing just as an interesting add-on for people who want to go a step deeper in this or figure out what's coming next. So there's a tool out there right now that a lot of people are playing with. It's called Runway. What Runway enables you to do is it enables you to create text to videos, which are okay. It's interesting, but it's not there yet. But what it's doing really is it's taking an image and creating a four-second video based on that image. and if you learn how to create images that quote-unquote have motion in them. So if I go back to the BMW example, if you see the car moving and there's those leading lines and those blurry things behind them, then it gives MidJourney runway a very good cue on what motion is supposed to be in there.

1:00:34And people are creating insane ads and videos right now just taking the output from MidJourney and placing it into the next tool, which is Runway, to create these videos and then stitching them together in an editing software to create a final product. So the next frontier that I'm sure that is running is video that will follow the same exact track of here. So any video assets you'll need will be able to be there as well. And so just like opening these people envelope even further on the possibilities of generating AI beyond text, beyond images. But that's it for today. this was an amazing session.

1:01:12Again, thank you again, Rory. Thanks everybody who's joining us live. And like you said, we'll do it again. Thanks. Bye everybody. Wow. Absolutely. Wow. The things Rory is able to do in his business and for his clients, creating assets that convert better. It's just mind blowing. If you want to also watch what Rory was sharing, feel free to also go to our YouTube channel. It's called Multiply, spelled with AI in the end. So M-U-L-T-I-P-L-A-I. So Multiply.ai is the YouTube channel. And over there, you'll be able to watch everything that we're doing if you would like to also see the demonstration that Rory was sharing.

1:01:56As I mentioned, if you want to join future live sessions that we're doing with experts like Rory, and we have a few exciting people coming up, then just follow me on LinkedIn and look at the events that I'm posting, and we would love to have you join our events. And now to this week's news. Now, I know this episode has been already pretty long because our live sessions are a full hour, and they will keep on being like that because we want to give the opportunity for people to really learn in depth. But to make it the perfect storm, Murphy's Law set up that there were a lot of exciting news this week.

1:02:27So what I'm going to do, I'm going to go just over the main ones, but I'm going to add the ones I'm not covering to the show notes plus links to learn more. So if you're interested, you can dive into that just by looking at the show notes. The news that I do want to share here are obviously the biggest and the most impactful to businesses. And so the first one is that Salesforce announces Slack AI. It is a full integration of a large language model capabilities into Slack. It enables you to do stuff like channel recaps and thread summaries and provide AI-generated highlights and summaries to help users catch up to recent activity and conversations.

1:03:03You can search for answers using natural language questions across multiple channels that will query older Slack messages, et cetera. So basically, a lot of amazing AI support to the already very capable Slack capabilities. They are planning to start rolling it out in beta early this winter, and then roll it out to everybody in 2024. My take on adding AI to Slack is it's not a surprise. Every large platform out there, and probably mid-platform as well, will add AI capabilities, which will make it more effective and more efficient to use. And we're going to see that across the board. And that's going to be one of the reasons the big players are going to win this battle against a lot of small startups trying to develop similar things.

1:03:51Another big piece of news coming from one of the giants is Microsoft announced that Microsoft will assume legal responsibilities if its customers get sued for copyright infringement while using co-pilot AI services. Now, this is obviously huge. Why is it huge? Because if you think about what are the biggest fears that companies have when adopting AI. One is obviously their data security, and that will definitely be addressed by Microsoft across everything that they're doing. And the second one is probably the potential infringement of copyrights. And by Microsoft providing an umbrella protection to its clients, they're taking that aspect of the risk as well.

1:04:36This comes only a week after OpenAI announces the enterprise launch of ChatGPT, which also addresses security and data issues, which, as we know, is also the backend of some of the stuff that Microsoft is rolling out. So the combination of these things come to address the biggest concerns of large corporations and enterprise clients, which, going back to my point from my previous piece of news, it gives you another reason why the big players are going to win this battle. Smaller company that Microsoft cannot take the legal exposure of all its clients, but Microsoft can in its push to get enterprise clients to start using its AI products.

1:05:16Another big company that have made a big announcement when it comes to AI additions is Zoom. Zoom just announced AI Companion. It brings a lot of really cool and efficiency building AI capabilities into the already extremely popular Zoom platform. Some examples of what it will be able to do, it will be able to help you chat more effectively and generate chat responses. It will help you catch up if you jump into a meeting, mid-meeting. It will give you a summary of what happened so far. It creates summaries and action items from the meetings themselves. It allows you to take notes while you're doing the meeting, while analyzing things that are happening.

1:05:57It can translate things into chat in real time, etc., etc. Lots of goodies come in this AI bag from Zoom. Not surprising, especially after last week, Google announced that Duet, which is their tool for Google Suite, will have similar features in Google Meet. So expect the future of digital meetings that we all do way too much of to be way different than it is today as we start learning how to use these really amazing tools. Now, will that cause us to have less meetings or more meetings? Time will tell. I really hope it will be less meetings or at least shorter and more effective meetings, but I definitely cannot guarantee that and I'm not going to bet on what's going to be the outcome from that perspective, but definitely it will enable us to have more efficient meetings.

1:06:50And the last huge piece of news of giants from this week is that McKinsey is teaming up with Salesforce to deliver AI-powered growth to companies. The collaboration will combine McKinsey's proprietary AI models and assets with Salesforce CRM technologies, including their new AI platform, which is called Einstein, and including Data Cloud. The goal of this is to obviously also use the huge amount of data that McKinsey has behind the scenes of companies in different situations and allow clients to analyze relevant structured and unstructured data in a way that will improve business processes such as sales productivity, customer experience, marketing, personalization, et cetera.

1:07:32So another example where the giants are going to provide extremely high value using AI tools to their clients across probably every SaaS platform that we know. And the last piece of news that I want to share on the air, and the rest, as I mentioned, is going to be in the notes, is that Anthropic has launched a paid plan for its chat about Cloud2. I love Cloud2. I use it daily. I find that there are some things that it's doing better than ChatGPT. That being said, I find that the$20 a month of ChatGPT are completely worth it because of A, the advanced data analysis module that it includes, and B, some of the plugins that are incredible.

1:08:15And as of right now, I'm using Cloud2 without paying the$20, and I haven't hit anything that will cause me to pay the$20 a month that it costs to get the upgraded version of Cloud2. I will keep updating on that if I have new insights moving forward. If you enjoyed this episode and if you're enjoying the podcast, please share it with other people who can benefit from it. That's our number one way of growing and helping other people like you to learn and get better with AI for the benefit of all of us. And also I would appreciate if you would pull up your phone right now. Yes, right now. Pull up the phone.

1:08:51Unless you're driving, don't do it then. And open your podcasting platform and give us a five-star review and write whatever you want to write about the podcast if it's helping you and how it's helping you. That also helps us reach a bigger audience. And until next week, explore AI, try different things, share what you find across social media, share it with me on LinkedIn, and have an amazing week.