297 | From Prompt to Powerful: The Secret Formula Behind High-Impact AI Videos with Eve Whitaker

Hello and welcome to Leveraging AI, the podcast that shares practical ethical ways to leverage AI to improve efficiency, grow your business, and advance your career. This is Isar, Meitis your host, and we have a really fun and impactful episode for you today. Creating great videos that can connect with your audience on an emotional level is something everybody wants to do, but very few people actually know how to do it. Now, also, most people, probably not everyone, but most people know you can create amazing videos with AI today, and yet very few people know how to do that.

Our guest today, Eve Whittaker, knows how to do both. She knows how to create amazing videos that connect with the audience, and she knows how to do it with AI. Now, she spent over fifteen years of her career on TV in multiple different roles. Her latest role as an executive producer, which means she understands all the details and intricacies of what a good video looks like, something that the audience will want to watch.

But she also knows all the nitty little details, such as where the camera should be positioned, how the shot should progress, how long should the shot be, and all the things that make a great story when told through video. So this gives her an amazing background in what a good video looks like. But in addition, in the past few years, she's been all about AI solutions, and she has developed her own process that allows her to go from an idea in her head or somebody else's head to a professionally fully running produced video of that idea.

And in today's episode, she's going to share with us her step-by-step process and the entire way on how to go from A to Z with creating solid, effective AI videos, including all the details that a professional executive producer knows and most of us don't have a freaking clue about. Which, in my eyes, is very cool and very exciting. Now, since video is currently the best format we have to connect with our audiences, and it's also the best format from a shareable perspective, and it is the most viral content that can be put out there, I find this skill very exciting, and hence, I'm really excited to welcome Eve to the show.

Eve, welcome to Leveraging AI. Thank you. I'm glad to be here. I'm very curious about this one.

I saw your session. You and I both spoke at, Social Media Marketing World, and that's where we connected. and so I, I guess we both need to thank Michael for this. so thank you, Michael, if you're listening.

But, it's a really cool thing, right? The ability to create the right kind of video. Everybody can create a video, like especially today. But creating the right video that will actually produce results, connect with people, and lead somebody to an outcome is not easy to do, and hence I'm really excited about this.

Yeah, no, I find the, this new world we're in of AI to be probably the most exciting time for solopreneurs, entrepreneurs, small business owners, because we really do now have the tools that we need to really promote ourselves and take our business to a level that we never could before. and I, obviously being into TV, have always loved video, but even if I wasn't, as video usually is the king when it comes to content. And because now we have so many platforms we can put it on, having your messaging be correct, it really does give us an advantage we've never had before.

I love that. I will just add one more thing. I don't think it's just entrepreneurs and small businesses. I have several really large clients.

two of them are in the apparel space. Two of them are huge, gigantic companies in the apparel space, and they, until not too long ago, needed to do photo shoots for every new style, for every new season, for every new everything. and then that's very expensive because you need the talent, and you need the location, and you need the lighting, and you need the photographers, and you need all of that. Yeah.

And you need to do this every few months, and it's crazy expensive. And now you can do this without the talent, and without the location, and without the clothing. you don't- need any of that, to make it happen, and you can change it every day, and not just once every other month when you switch to your next whatever. And so I think it's true for anyone.

if you... Regardless the size of the company, if you want a way to effectively engage your audience through video, as you said, this is the best time in history to start doing that. 100%, and it really is for everyone. I just always, give a shout-out to the small guy- Yeah because I don't think they would've ever done this before.

Agreed. Where the larger companies would've, they just would've spent a lot more money. So the advantage is for anyone that wants to use it. You will save time, you will save money, and you will be able to give more value to your customers or clients.

100%. So let's dive in. Let's see, let's go. Let's go do this.

Okay. So basically, my process is, I did not create this. This is from all of my time working in LA, Los Angeles, on television shows, that a very simple framework to explain to the person that has never made a video is there's three phases. There's the pre-production, and that's where we get our script, we gather all of the things that we need to actually produce the video.

And then there's the production, where before AI, that would be getting your photographers, your cinematographers, actors, a location, and going out and actually the production day or days, shooting, et cetera. And then there's the post-production, and that's where we take everything back into an edit bay, and we put- Put it all together. and so what AI has done is it's taken the whole entire process and simplified it. Now, the production actually got the most simplified.

That is where we literally put a prompt into an AI video generator, and we generate our video, which still blows my mind today. post-production is getting better with AI, but we still need to use the same tools that we used before. That is one of the biggest misconceptions, where people are like, how do you make such a long video, and why do yours turn out great and mine, it doesn't always turn out?" And so that's where I'm always very, I like to get ahead of that to say, "You still need to edit your videos right now."

We're not at a place, and we could be tomorrow, but today we're not at a place where I can just tell the AI what I want and it just makes it perfectly. so that's where the post-production. But where I get the most excited, where I think I can help the most people, is once they understand the pre-production, and that is coming up with your message. Because you can have everything work out perfectly in production and post-production.

If you don't have your story down, your message down, it's not gonna do anything. And especially today with all the content coming out there, you have to understand how to stay ahead and have the correct message and say it to the correct people That's fantastic. Yeah, I agree with you 100%, right? At the end of the day, whatever video you put out there has a goal.

The goal might be just to entertain people. Maybe you're not selling anything. The goal might be to promote a nonprofit thing that you're going. it doesn't have to be financially driven, but there's always a goal.

Otherwise, you wouldn't be spending the time to create the video. and maybe you would if it's just for fun, but then you wouldn't post it anybody. if you're posting it for other people to see, then it means you're trying to get people to interact with it, engage with it, impact them somehow. And that means you have a goal.

If you have a goal, there's a way to get there better than just, "Oh, let's shoot a video and see what happens," which I love the direction that you're taking this. I will say two things about the post-production, thing, because first of all, obviously I agree with you. That's the part where it's the least impacted right now by AI. Two things.

One, I have been editing all my videos, so every content you guys ever consume of me is produced using a tool called Descript that has been around. We - I've been using it for probably five years in podcasting, so it existed way before the whole AI craze. But it is an AI tool that allows you to edit and produce videos significantly faster in the post-production aspect of it. And we're now adding more and more layers of like external AI that works with it, which is really cool.

So this is our latest, testing with that. and the other thing is, Gemini, so Google just came with Omni this last week, and Omni is actually really good at editing videos, any videos, like either AI videos or real videos with just prompts. It still doesn't do all the things that Eve talks about. yes, you can remove the background, and you can remove characters, and you can dress them up in different things, and you can do things that are absolutely magical that you needed a high-end production studio for weeks to do the stuff you can just do with one prompt.

But it still doesn't know how to take six videos together and stitch them perfectly so they look like a one video or stuff like that. So there's still work to be done manually if you want this to be the perfect output. But we're slowly closing more and more aspects of this. We are getting closer.

We are definitely getting closer. So the first question that I always ask my audience when I start talking about if there's so many different frameworks, so let's focus on the person that wants to market themselves using social media. And so I always ask them, "Do you know who your audience is? And do you know what problems do you solve for your audience?"

Now, most of the time they ask - or if I ask them that, they tell me what they do. So I always use the weight loss example because it's really an easy one. Most people have some relationship or experience with either someone trying to lose weight or they've lost weight, or they've tried. And so let's say they're a gym owner, and I say, "Okay, what problems do you solve?"

And they're, I help people lose weight. I have a gym. I have this equipment. I offer classes."

I'm like, no. What problems do you solve? Because there's a reason people can't lose weight, or there's a reason people don't go to the gym." So I also say, because we're putting out so much content We all know and we've heard of, hopefully most of the people listening, your ICA, your ideal customer avatar.

But you have several, right? not everyone's gonna fit into that box. Once you have your ICA, you have your male, your female, the age groups. Everything is very different with that.

So I tell people in the pre-production phase, we're gonna talk to one, one person, and we're gonna solve one problem. And so I created a custom GPT. Now, you can do this in a Claude skill. You can use Perplexity.

It doesn't have to be done in ChatGPT. That's just the first LLM that I used that I still love. Everyone's "When are you gonna switch over to Claude?" And I'm like, "I have.

It's just easy for me to share my custom GPTs." And so a little, foreshadowing, I am sharing this custom GPT with your audience. Awesome yeah, so this, 'Cause it's the hardest part, I think, in really getting this part down. And once you understand it, it's like it clicks, and then you're like, "Oh, I made it harder than I thought."

But once you get it, it really does help. So if I can, I'm gonna share my screen, and we'll just walk through one of these and just show the audience and verbally say to the audience kinda how this works. Awesome. Fantastic.

Yeah, and those of you who are, not watching, you can watch this on Spotify right now, if you're not driving, or you can watch this on our YouTube channel, which you can find the link on the show notes, and there's gonna be a link to the GPT that Eva was gracefully sharing with us. So that's gonna be a link for that as well. and if you're, if you can't watch this, that's perfectly fine. We're gonna walk you through everything that is on the screen, so you can keep on listening if you are driving or walking your dog or mowing the yard or doing something else that doesn't allow you to watch this.

So I call this the problem identification assistant. So I help you identify problems you solve for your clients or customers. And the reason we're gonna do this is each piece of content you produce is gonna solve one problem. So if you can solve one problem in a video, and this can be artistic too.

Like you said, it doesn't all have to be for like saying, "Hey, I'm a gym owner. Here's my gym. Come on down." So we'll just say you just literally hit start here, and this is just so the average person, if they have no idea what they're doing, this is what you can just start, and the tech is gonna go a little slow, so we'll wait for ChatGPT to think here.

And, while we're waiting for that, I will say one thing. Those of you who do not know what custom GPTs are, which this is what this is, a custom GPT is basically a set of instructions with some additional context if you want, files and other references and so on, that does a repeatable process. And you can save those and then reuse them again and again, whenever you want. And as you've said it, they're very easy to share as well.

So there's a literally a share button at the top right corner of a custom GPT, and it allows other people to use it. and inside of it, like I said, there's just a set of instructions. So let's continue. So it's asking you, "Let's get clear on the problems you solve.

Please answer these two questions. Who are your clients, and what is the main problem you help your, you help solve?" So I'm gonna say I own a gym. and I help people weight.

And I'm keeping this pretty basic just so you can see how it's gonna peel back the onions. and so I'm just gonna even just leave it there and just see what it's gonna come up with, 'cause it's, it is programmed to now say, "Okay, based on what you shared, here are three core problems your clients are dealing with. They struggle to lose weight consistently, even after trying diets or workout plans before, or workout plans. They lack accountability and motivation, so they start off strong, but then fall aft- fall off after a few weeks.

They feel overwhelmed by fitness and nutrition and don't know what actually works for their body or lifestyle." So then the user's gonna choose one of these, right? So I'll let you pick. Which one speaks to you in your experience of knowing someone like yourself?

I would go with number two. Perfect. And I would put it in GPT. And you literally just have to...

Yeah, you just have to put in number two. You don't have to type it out. So they lack accountability and motivation, so they start strong, but fall off after a few weeks, which I do think is very common in the weight loss world. So then what the GPT is gonna do, the little AI tool, is it's gonna break it down into five specific problems as to why.

That's what I'm always asking, why? Why? So why do they start off strong and then fall off? And those are the topics that we're going to actually use in our content.

So they can't stay consistent with workouts because life, work, and family constantly interrupt their routine. They lose motivation when they don't see results fast enough, so they quit before the real progress happens. The third one is they know what they should do, but they struggle to hold themselves accountable when nobody is checking on them. The fourth one, they start and stop fitness programs repeatedly, which makes them feel frustrated, discouraged, and embarrassed.

And the fifth one is they feel intimidated or uncomfortable in gym environments, so they avoid showing up consistently. So out of those ones, which one should we pick? I love this for one very interesting reason. I think if you would've just gave this to a random gym owner, they would've said all of the above.

And I think what you're doing is that you're forced to speak to one problem. Yes. And then you can do another video for the other problem, and then a third video for the third problem. But in this one video, tackle one audience.

And the reason I love this is because the person that's gonna watch it, one person is gonna say, "Oh, this is exactly my problem." Exactly. versus this is, eh, my problem, but not exactly. And so I love the focus and making it very specific because these kind of sound the same, but they're not.

Like you're saying, they're, they address a slightly different pain point or trigger at a specific individual, they will say, "Oh, this is exactly me." so let's go with, the first one. They can't stay consistent with, workouts because life, work, blah, blah, blah. Okay, perfect.

So all I'm gonna do is I'm gonna cut and paste that. And just to go, or just copy it for now, to exactly what you said, that's exactly the plan, is if you were doing this and you were building out your content strategy, I always talk about it in pillars. I like to work in quarters, but you could do it by month. It doesn't matter, right?

And so this month, let's say if you were doing it in a month, here are three pillars, which is once I put in I'm the gym owner and I help people lose weight, the AI gave me three pillars that I can produce content on. And then with those three pillars, it gave me five for each one. So I could go in and say, now let's tackle number one. Now let's tackle number three since we picked number two."

That's now 15 pieces of content that we can produce. So you are exactly right. We are just gonna start talking to that one specific problem, because we've heard this before. If you speak to everyone, you speak to no one.

Yeah. So let's get very specific. All right, so then I'm gonna go into a different GPT. Now, with this, you don't have to have my GPT to be able to do this.

This is just how I move faster, right? So I would go into a GPT that I created called Ad Spark Engine, and what I do with this is I begin to look at where is my audience in relation to me? we've heard about funnels, top of funnel, middle of funnel, bottom of funnel. If you haven't heard of funnels, let's do have they heard of you before?

Are they familiar with you? Are they ready to buy for you, from you, or are they brand new, right? So we have the cold audience. They're the ones that have never heard of you, top of funnel.

Then we move into the, they've seen your stuff around, but they haven't yet really engaged with you. That's your warm audience, middle of funnel. And then you have the bottom of the funnel, where they're ready to buy from you. They just haven't hit it yet.

So with this, I have just started this to where I'll put this in, and what this does is this is gonna create an ad for me that I can iterate on and go back and forth using AI to do this. Now, like I said, if you don't have my Ad Spark engine, you don't h- you can use AI. Just go into any LLM and answer these questions, and you will get a result that will work for you. So it basically says if you already have a specific problem, with your audience, from one of Eve's other G- GPTs, you can drop it in here.

If not, no problem. I'll still build a strong commercial. And it literally goes over what I just said. Now give me, these four so I can build the right ad.

Where is your audience right now? And so I'm just gonna say cold and just pretend that I just started this gym. what's the one emotion we should hit in the first three seconds? Talking about hook.

do we want urgency, aspiration, curiosity, trust, protectiveness, surprise, or inspiration? I'm gonna go with inspiration just because I want to launch this campaign maybe a couple months before my gym's gonna open and start getting the word out there. what does your product, primarily do? Solve a problem, a painful problem, create a transformation, protect something, or build a skill, or remove confusion?

So I'm gonna go with transformation where will this ad run? Now, this isn't really as important because the only thing that - the only reason it's asking this is because it will then create, an ad that will fit these formats, whether it's, TikTok wants forty-five seconds, Reels, et cetera. So I'm just gonna say Facebook right now, just to be simple. You can say all of them.

And then I'm just gonna cut and paste my, problem in. So again, if you don't have this GPT, you could literally just go into ChatGPT, Claude, Gemini, and just answer these questions and say, "Hey, I want you to create an ad for me," answer these questions, and give it the problem, and it's gonna come up with something that you can definitely use. So it says, "Here's the problem we're building this around. Your audience keeps starting over with fitness because their routine only works in perfect conditions.

The second work gets ca- the second work gets chaotic, kids need attention, or life shifts unexpectedly, the entire system collapses. This is not really a motivation problem. It's a fragility problem." That's strong because it's emotionally true, visually clear, and easy to dramatize through a metaphor instead of a generic gym frustration.

So based on your audience, which is cold, the desired emotion, which was inspiration, product outcome, transformation, and platform, Facebook, I recommend the BAB plus pattern interrupt hybrid framework. Here's why. So I'm not even gonna go into this because this gets into, the backend, the GPT, and it's frankly boring. We'll just get to it.

So the emotional outcome is maybe fitness doesn't need more discipline. Maybe I've just been using a system that breaks under real life. would you like to proceed with this framework? So I'm just gonna say yes.

So this particular custom GPT, I put in a lot of backend information about different ways that people will relate to ads depending. So now it's gonna do my commercial script. It's gonna do a duration of thirty seconds. Again, the cold audience, the platform being Facebook.

It's gonna start off with a three-second hook. A man is jogging on a treadmill inside a house that violently shaking like an earthquake simulator. Kid runs past him, phone alarms explode, laundry falls from the ceiling, a laptop slams open with notifications, but he keeps trying to run. Then the treadmill suddenly snaps in half.

Silence. your fitness plan only works when life doesn't happen. That's a text overlay. Then it's gonna- That's pretty good.

Yeah. It's fun. And- I can - First of all, I can visualize it in my head, and second, I think it's a really good tagline. I will say something else that I really like about your approach, is the fact that the question was what emotion you wanna invoke in the first three seconds.

Yes. This is... the hook has to generate a strong emotion for the people to stay around. It's not about what you're selling is about what it's gonna make the people feel connected emotionally going back to I'm solving your particular problem.

If I'm not, don't watch this video because I don't care. I may care, but this is not the role of this particular video right now. So I, I totally see how all these things connect and absolutely love this. So let's continue.

Awesome. and I'll throw in just a little fact that I read that Google did a study, probably about six months ago and came out and said that for someone to actually take an action, and purchase from you, they on average need to be exposed to your content for up to seven hours of content across four platforms with at least seven touch points That's a lot, right? And so that's why the name of the game is content, right? And we're going back to what we talked about earlier, AI allows us to do that in a much more simple way.

Now, those touch points and that content doesn't always have to be video, just to be clear. Yeah. If people are like thinking, "Oh my God, I have to make seven hours of 30..." No, it could be a blog, it can be your website.

It's all different touch points, right? All right, so moving back to the body now. So we just went over the hook, and that was just in the first three seconds, right? the body is, then it's gonna divide it into beats.

So just think of it as little scenes. Quick visual cuts, meal prep containers slide off the table in slow motion. A calendar fills with chaotic overlapping events, running shoes buried under toes and work folders. A giant maze forms underneath a dumbbell in the center.

This is not shown or this is not shown as stress. It feels like a broken system swallowing him whole. The text overlay again, "Work interrupts, family interrupts, life interrupts." So again, we're just drilling in that this is my life, this is what's happening.

Then the next beat is the pattern shift. We now see another person in a totally different environment doing pushups while pasta boils. resistant bands attached to a park bench during soccer practice. Short workout in or short workout beside a sleeping baby monitor.

Walking, meeting with a - walking meeting with a weighted backpack. I use one of those. life is still chaotic, but the system bends instead of breaking. The text overlay, "The goal isn't perfection, it's adaptability."

Then the third one, this is taking us from 16 seconds to 22 seconds, is the visual metaphor payoff. The earlier collapsing maze wall suddenly stops moving. The path simplifies. The person walks straight through effors-effortlessly.

Cut between smiling with the family, quick workout moments, visibility phys-visible physical progress, calmer energy. Text overlay, "Consistency starts when fitness Fits real life. Resolution, and this is plus the call to action. The product finally appears, which if we said this was a - we didn't tell the AI that it was a gym.

So if it says product, we'll pretend that this is for a fitness coach, and that's part of his gym is also mentorship. so simple app program, interface showing short, flexible workouts, adaptive scheduling, realistic habit tracking. Final cinematic shot. The original man now jogs calmly outdoors while chaos still exists around him.

Kids playing, phone buzzing, life moving, but none of it controls him anymore. Text overlay, CTA, stop restarting, build a routine that survives real life, start today. And then it just goes over the tone, the sound direction, and a fill-in-the-blank template to begin to take that. So that's where you could then say, "Now build me a prompt for whatever AI video generator you're using."

So I'm a big fan of Seed Dance. I'm loving Omni. Omni just came out, like you said, last week. And, a lot of if you're doing something for either a presentation, meaning it's gonna be blown up on a big screen, or if you're actually making a video for television, you're probably using Google's Veo 3.

And so all of those different AI video generators need to be prompted slightly different. the good news is there's enough stuff online that you can just ask AI to create the prompt for whatever AI video generator you're using, and it's gonna know automatically how to do that. and I wanna pause it for just one second because I wanted to comment exactly that on the previous step we just did, right? So you're all thinking "Oh my God, this is amazing."

we just gave it a few small things, and it created us a script for an ad that sounds really cool and really fun and really, amazing. But Most people don't know how to explain to the AI what to put into the custom GPT in order to be able to spit this out. the reality is you don't have to. Does Eve has the benefit of doing this professionally for 15, 20 years?

Yes, of course. but can you go way beyond your current knowledge? 100%. if you tell AI, "Hey, I want to get from an idea to a script that will be 45 seconds to a minute, depending on the platform, and you will go and research which platform.

I want the script to be able to connect emotionally in the first three seconds. I want it to be funny." I want whatever you want it to be. You tell it, and then you tell it to go and search best practices and to create the input and output that are required in order to make this work.

It will give you a very solid starting point. If you take the transcript of this podcast as a starting point, as an example, and say, "Oh, I want something like this as an output. help me create the instructions to do it," it will help you do it. So the starting with a blank sheet of paper, which is always the hardest, especially when it's a topic you're not familiar with.

I don't know how to write scripts, right? I've never done this, not even once in my entire life. Like, how do I even begin? How do I - how many sections does it need to have in 45 seconds?

Is it one? Is it four? what do they need to be? Like, all these things the AI can research Test, implement, test again, reiterate.

So you can... You don't have to do the work yourself. You just need to know what to ask for. You need to know where to find good references, and then the AI can get you to a point that, again, may not be as good as Eve that's been doing this professionally, but is like a hundred times better than your current knowledge on that topic.

In this particular case, how to write a script for a good ad. You can - And e-everything you just said is so true. I learn AI by AI. So if something new comes out, I have AI teach me how to do it.

And, Gemini just took a major leap forward, and I love this, where you can actually put a video into Gemini. So let's say you have a commercial of a competitor or, you're researching on YouTube videos that have done really well. You can take those, you can either take the transcript or if you can download the video and put it into Gemini, and you can give Gemini all the information about your business, and you can ask Gemini to create a video script for you based on the framework of the video or transcript you put into it, but use your business.

And it's a launching off point. I tell people this is where I use AI as, an assistant, a brainstorming assistant, an iteration assistant, let's get better. And one of the things that I always prompt it to do is after it gives me something, I always tell the AI to critique itself. I'm like, "Critique the above and optimize whatever the task was."

And it always gives me a layer above. Now, I warn you, don't keep saying that because it will trip up the AI, and it will keep critiquing, and all of a sudden you'll be going down a rabbit hole. But if you do it once, it does a really good job, and it will refine it. So you could even take someone else's or a YouTube video that is - you're inspired by, put it into Gemini, ask it to create the framework of your video or of your business, and then say, "Hey, critique it and optimize it," and give it the goals that you want.

I, I want this video to be more for inspiration, or I want this video to be m- used to generate leads. I want them to actually go to my website with this video. And it knows what to do. Everything you just said, you don't have to know this stuff.

It will do it. It will do it for you. Brilliant. I love everything that you just said.

what's the next step? so you in the - Then you take it, you ask it to create a prompt, and then I assume it's not one prompt, or maybe it is. Like, how many prompts does this generate in order to create a solid ad? do you take every section and make it into a prompt?

Which I assume, or maybe not, like this is - Tell me what the next step is, basically. So the next step, and I'll walk you through it instead of, showing you on this one- Yeah unless you want me to show you. But, is it depends on the AI video generator. So let's say we're gonna use Seedance.

Seedance I like for what I'm gonna say the consumer versus the professional. So a professional is an ad agency or someone that is producing something for television. So we're gonna talk to consumers right now. So I would use Seedance or Omni, and both of them are very similar.

The difference is Omni just came out and, it allows you to produce up to 10 seconds of video at a time. Seedance allows you to produce up to 15 right now. So that gets back to where we talked about the post-production, where we have to string it together. It's gonna have to be strung together.

So with that, what I would do is I would put it in a ChatGPT or Claude or whatever you're gonna use, I would take what my GPT just did, and I would tell it, "Hey, I'm creating a 30-second video in Seedance. Will you create me the prompt and break it into two sections, each one 15 seconds, and make sure that the s- that they can go into each other seamlessly in an editor?" And I'll even be more specific and say, "I want you to end the first prompt on a scene that then cuts to a completely different scene in the second prompt."

And that way it's gonna be an easy cut. You're not trying to match frames or match up something that's gonna cause a little bit of a glitch or anything, and it does a pretty good job. Now, that is where AI can AI, and it might not come out perfect, and that's where we go into Final Cut Pro or CapCut or whatever you use to edit Descript, and you can edit it together there. And I tell people, "You're gonna generate...

You're gonna take that prompt, and you're probably gonna run it a few times because AI is gonna mess up. It just does." And so the other part of my process is, again, Omni just getting launched or dropping like a week ago, puts a couple caveats to what I'm about to say, but let's say you're in Seedance, and you put in your prompt, and it renders out a video, and there's a couple things that just aren't right. I give the example that I did a video right before Social Media Marketing World.

It was just an animation of a girl that walks out on stage, and the audience becomes a monster, and the stage begins to turn into a cliff, And then she comes back, and she's "Oh, I'm okay," and then a boulder falls next to her. And it clearly said that in the prompt, but when it generated it, instead of a boulder at the end falling next to her, it was like half of a boulder, and it fell on her head. Which was funny, and that could be what I call a happy accident, and I could like it, but I didn't.

And so what I did was I took the prompt, and I brought it back into ChatGPT, and I said, "Hey, this is the prompt, but Seedance, instead of having the boulder fall next to her, is having the boulder fall on her head. Can you rewrite the prompt?" And it just rewrote it in its way of, really stressing that the boulder falls next to her, and then the next iteration was perfect Yeah, I wanna add a few things. I love this process, right?

It's an iterative process. You say, "Oh, that sounds a lot of work." I'm like, "Try shooting an actual video, and you'll see what a lot of work is," right? Then people say, "Oh, but every time I write it in CDance on the full resolution and for the 15 seconds, it's gonna cost me $3 instead of, free or whatever."

I'm like, "What are you talking about? doing the actual video would've cost you $50,000, and that's without dropping boulders on the stage." So- Yes it's... Yes, there's gonna be a cost to generating good videos because, yes, you may need to render the one scene 10 times until you get it right But it's still negligible, both in means of times and means of money compared to the - doing this the old school way, which is actually having a videographer and teams and locations and so on.

So this is one thing, be ready to run this multiple times. The other thing that I will say that has to do with, connecting different scenes. I do this a lot in different tools, but in many cases, you want to connect two scenes seamlessly to one another. And most of these tools today have a first frame, last frame input, meaning you can say, "I want the first frame to be this, I want the last frame to be that," or, "I want the first frame to be this and the last frame to be that."

You can decide. But the cool thing about this benefit is that you can go and grab the final frame of the previous video and use that as the first frame on the next video. And in the more advanced tools today, and SeeDance is definitely one of them, you can use the previous video as one of the inputs explaining to it this is gonna continue what the first video has. And then it has the exact first frame, and it has the transition and the lighting and everything that happened, and it takes that into consideration when it is rendering your new video.

And you can get to a completely seamless, connection between video A and video B. The one thing that I found that these tools do not do well, and I'll tell you what my solution is, and then you can give your solution, to the audience, is voices. if you ask them to speak, and they will. Like they speak and you can even prompt the voice.

It's a deep female, it's a whatever - W- like whatever voice you're gonna say. It's never exactly the same voice. Like it's similar because you define like the type of voice you're looking for, but it's not gonna be exactly the same voice. And so it looks a little weird, right?

you run section one, section two, section three, and it sounds like a different person, spoke. And the solution that I found, and again, I'll let you tell yours, is I then take that back into ElevenLabs, and ElevenLabs allows you to switch the voice, meaning I'm gonna replace all the voices, not just one of them. And it knows how to replace the voice with the exact same tonation, with the exact same wording, in the exact same milliseconds. But it's just the same voice, which takes us back to post-production.

Now I go back, remove the original voices from the original soundtrack that had three different voices, replace it with what ElevenLabs generated that fits the lip sync and everything perfectly because it's on the millisecond exactly what it was before, but it is the same voice on all three. So there's all these little glitches that are still not perfect, but they're still solutions that are still way, way easier than doing it the old school way. That is exactly my, solution also.

Okay. I did hear that, in Omni you can now connect a voice, you can choose a voice from ElevenLabs. Yeah. So that is something that if, I haven't tested it myself, I'm diving into that this week.

and what I always tell people, because there's always that what tool do I use? This one just launched this one just launched this. Whatever you're using, if Omni, Google's Omni came out with that feature, I guarantee every other AI video generator is gonna come out with that feature. So it is, because it's kinda dumb that it hasn't.

Like I've always said that can't be that hard to be able to choose. Maybe right now you only have a small library of choices, or if you wanna clone your own voice or you have something like that. but Should be there. The technology is there.

That's not like us thinking that we should be able to just get a perfect prompt right now. so that is hopefully the solution is, is gonna be easier than taking that step into ElevenLabs. But that is- Yeah, and I agree with you 100%. The one thing that I will add, there are cases where you're gonna use several different tools, and that's perfectly fine as well.

Yes. Like the editing we're not gonna do inside of the AI for now, right? Like you said, you're gonna do it in Final Cut Pro or in CapCut or in whatever editing tool you wanna use. It's still not in the AI.

You started with ChatGPT to create the... we're going to use multiple tools for the foreseeable future, and we just need to find the right tool for the right step in order to make the overall process the easiest and the outcome the most relevant, professional, capable that you need, not necessarily related to video production. Exactly. Eve, this was fantastic.

I'm sure people learned a lot. If people wanna follow you, learn from you, work with you, what are the best ways to do that? I have a school group, that I just started, called Video Spark, and that is the best place if you wanna work with me. I do, ask-me-anything calls twice a month, depending on what tier you're in.

I'm usually teaching either a two-day or a three-day workshop once a month. The premium tier get that, included. So that's probably the best place. I am on all social medias.

I do, usually find myself more on LinkedIn and Facebook are my sweet spots, where I just find I spend the more time. It's probably 'cause I'm older and I'm not on TikTok as much as some people. but, so that, that would be where I would suggest. And, because this has been so fun, I decided as I was showing you that second GPT, I'm gonna give your audience both GPTs so they can go and they can do the what's the problem and then take it in and iterate, just so they can have that and then, you can come find me if you wanna learn how to create the videos better.

'Cause there are more steps, but I think that is just such a good way to get started. it just is. And I'm always updating them. so hang onto them, and you might notice that there's a version two, version three, stuff like that.

and that's just usually me updating it. Awesome. First of all, thank you so much. I'm sure a lot of people in the audience will want to play with that.

I really appreciate you taking the time, sharing your information, doing the cool things that you're doing, and sharing with the world. thank you. Yes. Thank you.