How to Create Any Content 10x Faster With AI

0:00If you want to make more content faster, all you need is a better system. And the truth is, this is the secret for how all the top creators and brands are generating so much more traction than everyone else. They've all built winning content systems using a specific combination of AI tools. Now, I'll be honest, most AI tools are absolute trash, and the ones that aren't are usually way too general to actually help you with content unless you really know what you're doing. But here's the good news. I built a workflow that uses AI tools to speed up the content creation process significantly. It's some AI and

0:32some human. And this isn't faceless AI slop content. These are the actual videos that I post that get tens of millions of views every single month. So, in this video, I'm going to break down my entire AI content system. This is my five-step workflow that uses AI tools to generate highquality videos way faster than normal. And by the way, if

0:50you don't know me, my name is Callaway. I have a million followers. I've done billions of views. and I'm one of the biggest creators in both the AI and content marketing spaces. So, this intersection AI and content is all I do all day long. All right, the first step, we'll call it step zero, is to define the content workflow. And it's really very simple. You've got idea, research, scripting, recording, and editing. This is the same five-step workflow that everybody has to go through to make any piece of video content. Now, the secret for winning on social media is to build a system that lets you get through that

1:21five-step workflow as quickly as possible [music] while still coming up with unique ideas, novel angles, winning hooks, viral scripts, and editing that actually holds attention. You can't just build an AI workflow that pumps out bad AI content, or you won't get the results you want. So, you need tools that will speed up the process while also maintaining the

1:40level of quality that will cut through. And that's the system we're going to break down right now. Okay. Step number one in the workflow is the idea. Most people have absolutely no process for generating and filtering ideas. And this first step really is where the content game is won and lost. Now, there are three places you can look to generate winning content ideas. The first is to just generate them from your own

2:00imagination. We'll call these originals. The second is to look at other content platforms outside of your home base. So, if you're focused on Instagram, you might look at email newsletters, X, YouTube, podcasts, linear TV, blogs, really anywhere that people are talking about your topic outside of your home platform. And the third option for finding ideas is to look at other creators in your niche on the home platform that you're focused on. So, if you're primarily focused on Instagram, this would be looking at other creators in your niche specifically that make short form video on Instagram. Now the traditional process for generating ideas, the nonsismatized approach is

2:35what I call just in time ideiation. This is what most people do and why creating content takes so long for them. Every time you want to make a video, which is usually every day, you sit down and start scrolling the feed to find new video ideas. Now, this process can take 5 minutes or a couple of hours. And

2:51people are doing this every single day. This is where you're losing a bunch of time and efficiency. And that's because when you scroll the feed for ideas, you're getting maximum noise and minimum signal. Even though the algorithms are serving you some good videos in your niche, you still have to wade through a bunch of unrelated nonsense. This is all noise. It's a complete distraction from your productivity. And what happens, as we all know, is that every time you open the feed to scroll for ideas, you end up getting sucked into something unproductive or you feel bad about yourself for the progress that other people in your space are making. I know

3:23this is what happens because I did the exact same thing for a very long time. It takes forever to make content when you're searching for new ideas every single day by scrolling. So, here's the systematized approach for coming up with ideas that will reduce your ideation time down to 1 hour per week. You can spend more time if you want, but you're not going to have to if you want to go as fast as possible. All right. First, you have to understand that the highest signal source for ideas is option three that I mentioned before. Option three was looking at other creators in your

3:50niche on your home platform. specifically looking at the videos they've made that have outperformed everything else that they've posted. We call these outliers, and this really is the cheat code for ideiation. Your goal is to find the outliers that outperformed everything else and study why they outperformed. If you can reverse engineer those pieces, you can easily make your own and you'll grow a lot faster. So, here's what I literally do when it comes to ideation. This is my actual process. I do this whole thing on sandcastle.ai. This was my content super tool [music] built specifically for this workflow. The first thing I do is make a

4:21watch list of all the creators on my home platform in my niche that are crushing with short form content. So for me in this example that I'm showing you, it's on Instagram for tech and AI. That's the niche we're covering. Now this can be other creators, it can be brands, it can be faceless channels. It doesn't matter who it's coming from, but you need channels that are doing well, high views, and growing fast. Now, to find them and add them to my list on Sand Castles, all I have to do is search their username or their URL in the channels tab in the search bar. When

4:49they come up, I click them and I add them to my list. It could not be easier. So, I've already done this for Tekken AI content. I've got these 18 to 20 creators. These are all the best in the world for my niche and I just added them one by one following the process I just went through. Now, on the channels tab in Sandcastles, you can already see in the upper left there are over 20,000 of the top accounts across Instagram, Tik Tok, and YouTube already loaded into Sand Castles. You can search those names, click the card, and they're available right now. But Sand Castles is

5:16not just limited to those 20,000 channels because you can literally add anyone else you want just by searching and importing. All you have to do for anyone on Instagram, Tik Tok, or YouTube is just paste in their username or paste in their URL. If they don't come up, click add to channel or that add to channel button on the right side [music] and then import them. Sand castles will automatically index their entire backlog and that person will be included in the network forever moving forward. So seeing this, you have unlimited flexibility to build any list with any amount of people across Instagram, Tik Tok, or YouTube, whether they're already

5:48on sand castles today or not. And I'm telling you, building these watch lists in Sand Castles is the single greatest time-saving content productivity hack you [music] could do. And I'm going to show you why it's so helpful when it comes to ideation and the rest of the workflow right now. All right. Once I have that watch list built, and it's really just a onetime effort, I then go to the videos tab in Sandcastles. On the [music] videos tab, I can filter by the watch list I just made. And this shows me all the videos posted by just those accounts in my watch [music] list. So what I do after I filter by the watch

6:17list is I go to post date and I restrict the posts to the last 30 or 60 days. I also go to engagement rate and I set the minimum to.5%. So I can filter out all the paid brand deals. And what is left are the top performing organic videos from just the accounts in my watch list over the last month or two. This is the cream of the crop for your niche. These are the only videos you need to study if you want to replicate the winners. Now, if you want to surface the top videos from smaller channels in your niche, any of the channels with fewer followers

6:48that are growing faster, you just want to sort this list by outlier score. That tends to favor the smaller channels. On the flip side, if you want to see the best videos at the top from the biggest channels, I call these the proven winners or the thoroughbred horses, you can sort by views instead of outlier score. This tends to favor the bigger existing channels over the smaller. If you want the most recent post by the

7:08last 30 days, you don't really care. They're organized by performance. You just want to see the ones that have posted immediately. Go change the sort to posted date. So, you pretty much have three ways to sort the videos. They're all going to be the same, last 30 or 60 days from your watch list, not including brand deals, but you can kind of order the list to favor small, big, or recent, depending on what you like. Now, all you have to do from here to get a collection of winning ideas that have already been validated in your niche is just watch the videos one by one. You could watch

7:3510, 50, 100, however many in a batch that you want to do at one time. Typically, I like to do maybe 25 to 50 at a given time. Now, this right here, what you're looking at is pure signal, no noise for the tech and AI niche that I've curated over the last couple months because every single video you watch here will be a validated winner in your niche. You don't have to look at anything unrelated, anything weak or

7:57poor performing. It's all right here. Cream of the crop. Now, as I'm watching the individual videos, what I then do is I extract out seven core variables from each video. And I just write a couple words for each of these. The seven core variables are the following. Topic, which is just a oneliner describing the video. Angle, which is a short summary framing the meat or facts or takes, really like what makes up the video. So if topic is a oneliner, the angle or couple sentences describing the thing that was covered in the video. Three is the hook format and that's made up of the spoken hook, what is actually said,

8:30as well as any notes on the visual hook that is shown. Four is the story format or style. This is the way the story flows and how it's written. Five is the visual format. So that's the layout of the visuals. Is it green screen, split screen, full screen? Like how are the visuals actually laid out? Six is the visuals themselves. Are there specific

8:47visuals that really catch your eye? Maybe it's a 3D render or some AI generated thing or some AI B-roll where the background is really cool. And then seven is audio, the music, the sound effects, really like the whole audio vibe of the piece. And I'm just jotting down a couple of words here and there. If I have a format that I know the frame for, I'll write it down. If I don't, I'll just describe it. Now, what this allows me to do is that I can map the components that make up the winning video. Because when you don't know what you're doing, when you watch a video,

9:14it's hard to know why you liked it or why it was interesting. If you break it into these seven components, you can start to understand what went into making up the video that was interesting or good. What this does is it starts to build a supererset of options for each variable. So, for example, let's say I watched 25 videos about Tekken AI and these were the best 25 from the last 2 months ranked by views. I went one by one by one. Now, as I zoomed out at my paper and I see the seven variables for each of the 25, [music] I can start to recognize, oh, for visual formats, there

9:45are only five different visual formats that show up across the best videos. And those were green screen, [music] faceless, split screen, full screen, a-roll, and yap, for example. There's about 40 different types of visual formats. We'll have a doc on sand castles breaking those down, but let's just say those five were what I saw over and over. This then becomes my possible set of formats that I can choose from when I make my own videos. So, if I'm lost trying to figure out what videos will work in tech, I now see those five visual formats work over and over. I should pick one of those to [music]

10:17start with on my own version. And this is super important. The best way to figure out how to make your videos work is to find an existing video that worked. Look at the seven variables and try to figure out which variables to keep the same and which variables [music] to change in your own way. This is kind of like the scientific method approach for content and it makes it so much easier for beginners when you simplify it in these terms. And so when I go to make a video, I look at the top of my list and I see this first topic that I've extracted and I say, "Okay, if

10:45I'm going to make my version, maybe I keep the topic and the angle and the hook the exact same as the reference, but I completely change the way the story is told, the visual format that I lay the visuals out, and the actual visuals that I'm using. So I keep some the same and I make some different, and now I've got a new version. Or maybe I do the flip. I keep the hook and the story and the visuals and the audio the same, but I pick a new topic with a new angle. And so now I've got their format and execution, but my own topic and

11:11angle. See, it's a new combination. This idea phase, this research process is all about figuring out which variables worked in your niche, getting that superset clear, [music] and then knowing which Lego options you have to build your own stack. Now, if you want, I will make another video much slower where I'm literally clicking through sand castles and doing this process idea by idea by idea and walking through how I'm deciding which to remix. I can make that for another video. That'll be a full long tutorial, so I can't include it here. If you want that, just comment something like, "Give me the idea sauce chef" with a little chef emoji with the

11:46chef hat. Just put that in the comments and I'll know exactly what to do. Now, I'm telling you, if you spend just a single hour per week running this validated idea mining exercise that I just walked through, you walk away with dozens of topics, angles, hooks, and formats that are proven to work in your niche. And then you can pick and choose how you stack them up to build your own signature style. So, step one in the workflow, remember we're going through the five-step workflow. We're still in ideas. Step one is to just run this process once per week. Now, if you want to fully systematize this and get

12:15yourself completely out of it like I did, you can hire someone to run the process that I just went through. You could send them this video, train them up on it, and then they do the idea mining process for you. The beauty of Sand Castles is that the person you hire doesn't need access to your Instagram to

12:30see your algorithm scrolling history. They can just jump on Sandcastles and build the watch list the exact same way you do. So, it'll either take you 1 hour per week or someone else you hire 1 hour per week. But we've just reduced your ideation time down from 5 to 10 hours of scrolling per week down to one. Now, I mentioned before those two other paths for coming up with ideas. One of them was brainstorming originals and the other was looking at other content mediums outside of your home base. For the originals, when you do come up with an original idea that you didn't see anywhere else, just jot that down on a

13:00note and just capture it as they come. Most people are not going to be able to come up with a bunch of original ideas every single day. And so, you don't want to build a system that relies on that raw innovation every single time. When you do come up with an original, what makes it challenging is that you're going to have to figure out what all of those seven variables are. When you find an outlier, you can decide to hold some

13:19the same [music] and just change a few. But with an original, you got to build the full stack on your own without having that validation. That's why original ideas are so much harder to make. For the other content platforms outside of your home base, if you want to systematize this as well, you could schedule intentional scroll time for 20 to 30 minutes per week, or you could use a product like ChatGBT Pulse. I also like Feedley, which is kind of like an aggregator of different blogs and newsletters. So, you could structure those kind of aggregation systems to minimize the amount of time you're scrolling on all of those surfaces. But

13:50just to frame this, here's how I think about ideiation. The goal is to make your ideation process like Grey Goose vodka. You want it to be four times distilled. The raw social feeds like YouTube or X, those are completely unfiltered. You're getting some algorithmic direction, but you're also getting a bunch of random stuff. So, that's like the base, least filtered way. Email newsletters give you one layer of filtering because the people writing the newsletters are trying to look at the social feeds and pick out the best stories. So that helps a little bit. Now content creators in your niche are going to be a second layer of

14:18filtering cuz they're reading the email newsletters and other aggregation sources and they're only making videos on the best of the best. Right? So that's two layers of filtering. Sand castles with your watch list is providing a third layer of filtering because now you're not looking at all content creators. You're looking at only content creators that you've hand selected in your watch list. Then when you filter the videos feed in Sandcastles, [music] that's a fourth layer of filtering because now you're not just looking at the videos from the creators you've chosen. You're only looking at the best videos from the creators you've chosen. So when you come

14:46across an idea in Sandcastles that's at the top of the video's feed in the process we just went through. That's four times distilled. And that is how you minimize noise and maximize signal [music] in the ideation process. All right. So at the end of step one, you should have a Google doc that has all of your shortlisted ideas. And for each of those ideas, you've broken down in seven bullets the core seven elements or components that you've extracted. You don't have to go crazy just jotting down a few words or phrases here and there so that you know what the elements were so

15:15you can zoom out and see the patterns. Now, when it comes time to make your next video, you're going to pick the best idea on the top of that list. If you have anything time-sensitive, you can prioritize that. But if you don't, you just pick the one that you're most excited about. But as I've already mentioned, you're not just going to straight copy the topic, the script, the hook, the angle word for word, because that would be corny and unoriginal. So this is where step two comes in, which is the research. Now, the goal with the research step is to try to find some unique take or interesting angle that

15:40can help make your version of the idea a little bit more unique to you. Typically, what makes someone's video original when they're holding the topic constant is by changing the angle, the actual meat or facts or take in the video. Now, the traditional non-sistatized way to approach research is then spending another 30 to 60 minutes deep diving and learning everything you can about the topic so

16:01you can try to build your own take. Again, that is a lot of time. Some people are already 2 hours in the holster on this idea and they haven't even started scripting yet. But here is the systematized based approach that I used for research to get through this piece a lot [music] faster. In step one, when I was doing the idea and I was writing down those seven pieces, I'm watching the video for 60 seconds one time through. Now, the goal, as we talked about, is when you're watching the video, you're trying to extract those original components. [music] But now that I'm in the research phase, I'm

16:27going to go back to that idea and I'm going to ask myself three questions, trying to find a new angle that the original creator has missed. The first question is, what did I find interesting about the original video that I wish the creator talked about more? The second question would be, what unanswered questions do I have after watching the original video? And the third question I ask myself is, what unanswered questions would a beginner have after watching that original video? Now, my goal is to find some additional information, something the creator missed that I can use to frame the idea in a different way. Sometimes, if you're making content

16:57on a topic you already have a lot of knowledge about, you don't actually need to go do additional research, right? Cuz just asking those three questions might pop an insight or a fact, something you pull from your own experience, and that becomes your new angle. Boom, research done. But if you don't have experience or original context on the topic, then you're going to have to kick into this research phase to try to find a new perspective. Now, for me, what happens when I'm in this mode? I just don't have the knowledge or perspective on the topic. I'm going to go to Chat GBT andor YouTube to try to close this research

17:27gap as quickly as possible. First, I start with Chat GBT. And when I get there, I use this prompt every single time. I want to make a video about X topic. Can you build a report summarizing what is going on, breaking down the most interesting aspects that will induce max shock, and sharing a super contrarian take on the implication? Assume I know nothing about the topic, but I want to know 99% after reading your report. All I've watched so far is this video, and then I'll put the

17:50link to the original reference in there. Now, this is really important. The goal here is not to do deep research and spend 20 minutes learning every nook and cranny about this topic. That is procrastination. All you need is a short cliffnotee summary to try to find an interesting piece that the original creator missed that you can frame your video around. Typically, you can close the gap on this in just 2 to 3 minutes with ChatGBT. Again, if you want to systematize this even further, you can build a chat GPT project that acts like a template where you paste those instructions in as a system instructions and then all you have to do is write in

18:21the topic, paste the link, and it will generate the report right away. Now, sometimes if chat GPT isn't enough or it's just like a really complex thing where I need a visual to explain, I will go to YouTube to try to close this gap on the research, but I know that if I go to YouTube, the search is not very optimized. I'm going to end up in a rabbit hole and I'm going to waste a bunch of time. So, I try not to do that unless I really have to. Now, all you're really looking for at this research stage are one to two non-obvious interesting facts and maybe one

18:47contrarian take that the original post didn't have. If you have a unique perspective on what the story means, write it down. If you have a unique take or angle, write that down. Typically, shocking and non-obvious truths that you find that you haven't heard before is usually all you need to make the angle compelling. Now, at this point, you should have spent five, maybe 10 minutes, and you'll have three to five sentences written down next to the angle. Right? You've got those seven bullets. Next to angle, you've got three to five sentences featuring a breakdown of the facts or takes that you found

19:14that you think are most interesting. Now, the AI and the script writing step will take care of the rest, and I'm going to go through that in a second. So, you really don't need to polish it much beyond this. Again, you can go longer on this research step if you want to, but you're not going to have to once you see the rest of the workflow, and it's all about deciding where to spend your time. Now, for me, I have somebody on my team that does everything we've walked through so far for me. He runs through sand castles a couple times a week. He distills the best ideas. He

19:38goes through this research flow in chatbt and YouTube to close the gap. And then we have a call where he pitches me the ideas one by one and asks me if we should make them or not. This is what true content systemization looks like. So if you've been following along, I've reduced my personal time involvement in this flow to zero. The ideas and the research, I don't do them. I have someone else doing that for me following the process I just walked through. For the guy on my team that does this, I call this role a content operator. This person would handle all of the front-end

20:04ideation, the research, and then ultimately pitch it to you, frame it, package it up, so you can decide which ideas to make. Now, if you don't have somebody on your team that can do this, or you don't have the resources for it yet, that's totally fine. You don't need to have a content operator. But this is what full systematization would look like. And that's why I broke down this whole process step by step so that you could do it for yourself if you don't have a content operator. So, so far, the ideation should have taken you 1 hour per week, and you can do that in a

20:26batch. And then the research should take three to five minutes, maybe 10 max per video if you're really dialing it in with chatbt. Okay, now we're moving into step three in the workflow. And this is the script writing. Coming to this step, you should have your idea short list. And each idea, you should have those seven bullets. And for the angle, you should have the three to five sentences

20:44that you wrote down from the research. Specifically, what's most interesting or non-obvious to you about the topic that you want to include? And this can be bullets or sentences. Just write it clearly in a kind of logical way. It doesn't have to be in a certain format. Now, here's where I lean on AI a lot for this workflow, but I also have a human review in the loop at the end. I've spent a long time trying to build a script writing flow with AI. If you follow this channel, you know I have some of the deepest stuff on script writing when it comes to short form and

21:11long form content. I'm so deep on this and I've spent so many hours trying to build this fully automated AI workflow. But the truth is what I found is that even the most dialed chatbt or clawed project just doesn't quite get there with respect to the consistency that I want. And if it does nail the flow and

21:27the format, it doesn't sound like me. And so there's always manual intervention required. But I will say the latest version of scripting we have in Sandcastles is the best AI generated scripting that I've ever seen. It's the best in the world. And the reason Sandcastles works so well, better than Chad GBT and Claude, even though it's built on top of those, is that it doesn't try to oneshot the script. And this is a really critical thing that we figured out. There's a feature in Sand Castles called the Vault, which approaches content script writing in a different way than any other platform in

21:54the world. We kind of pioneered this. The Vault allows you to save components of outlier videos. Those components that we extracted in step one with the ideas. And these components then become templates that you can use when you're writing new scripts. [music] As of this recording, the first two components you can actually save as templates in the vault are hooks and styles. Now, hooks are obviously hooks. is the first few lines spoken in the video. Styles are a combination of the writing style. So, the rhythm, the tone, the pacing, the flow, like the way it's written, as well as the storytelling structure, the actual story flow from section 1, 2, 3,

22:30and how the points go together. We combine those two things, the style and flow into one thing called the style. And so, the full workflow for how you do this ideation and use the vault is the following. You find an idea in the videos tab. You start extracting out those seven variables. Again, as a reminder, topic, angle, hook, style, visual format, visuals, and audio. Those are the seven variables. Two [music] of those, hook and style, can be saved into Sandcastles automatically. And so, what you do is if you come across a video and you really like the way it's hooked, you get hooked, or you like the flow and

22:59structure of the story. All you have to do is click on that video, get to the idea detail page, click that button that says save to vault, and then you can save the hook andor style. All you have to do is name it. And for the hook, you specify the amount of lines in the transcript that you want to grab as the hook. Now, what happens next is pretty magical. When you go to the vault tab, you can see that hook and style that you just saved. Now, the hook, as you can see on the right, shows up at the bottom as the original word for word hook from

23:24the transcript, but we've also automatically translated it into a Mad Lib format. That's what you see on the top. We've substituted all of the core words, nouns, emotions, verbs, adjectives into variables that turn this into a template that can be used for other videos. If for whatever reason the transcript grab or the Mad Lib translation wasn't perfect, you just hit the pencil icon and you can edit it super quickly. This becomes a template that you can use over and over. Now, on

23:49the style side, we do something similar. It's even easier. We take the entire transcript and we save that as a style, which really can be applied as a lens to any script. So, it'll make the script sound like the writing style from the reference as well as flow in a story structure in the same [music] way. And I'm going to show you all this in action in a minute, but I just want to explain how the vault works and why it's so powerful at improving the AI scripting flow. The cool part with the vault is you can also manually add in your own hooks or your own styles. So, if you

24:17have a database of hooks and you don't want to go find a bunch of references in the videos tab, you can just take that database, save it as manual hooks, and then you've got that in the vault. Same thing with styles. If you have transcripts you've already written, save that in, and then that becomes a style preset that you can use. But here's where it really comes into play. Let me show you the scripting. When you want to script a new video, and this is step three, you go to the scripts tab, you press new script, and you get kicked into the flow. The first step, of

24:39course, is to add your prompt. This is where you describe what script you want to write. Now, Sastle does not require you to write a novel. All you really need is like three to six sentences. And what I recommend is just pasting in the topic and the angle that you wrote from steps one and two. You've already got that on your paper. You don't really even need to format it. just paste that into the prompt and press go. Sand Castle will then generate a full research report based on the prompt you gave it. Now, this report is going to drive the meat of what's in the script,

25:04the actual body of the script. So, you don't have to read it word for word, but you do want to scroll quickly and just glance through to make sure it's including the facts and takes that you wanted to. The sources that are used to drive that script are included in that carousel at the top. But if you want to change anything, delete [music] anything, add anything, you just press

25:19the pencil and then you can add that in. Once you're good with the report, you press next. And here's where the vault pieces really come in. Step three in the workflow is the hook. And immediately you'll see that all of the hook templates you've saved in the vault, as well as a bunch that we include by default in Sandcastles are there in that carousel. If you click on one, it will generate two new hooks in the exact format [music] as the template, but referring to the new topic that you're talking about for this script. And this is how we keep the exact hook structure consistently over and over, but give you

25:48the flexibility to use different hooks and generate different options. You can click the same Mad Lib a bunch of times to generate new pairs. You can also pick a bunch of different ones and you can basically build a set of 10, 12, 15 hook options in one spot to pick the best one. This is the fastest way to create new hooks that are in consistently validated formats without having to write them by hand. Now, once you pick a hook you like, you can move to the next step. And similarly, this is where we use the styles, but it's even easier because the styles are listed as

26:14template cards and you just click one and go. This will apply like a lens or a filter on top of the writing so that the style and flow works to match the example. Now once you click go on the style, the script will automatically start writing. It will use the prompt you gave. The body of it will include based on the research report. It'll use the hook that you've generated and the style. This is how you modularly use components that were validated to get the script to sound way better on the first rip. What we found is that this flow results in the most dialed first

26:42generation short form script in the world. Now, of course, the scripting process does not end here because typically I will spend 5 to 10 [music] minutes cleaning this up by hand. The goal is to get this process to be automatic and hands-free every single time. And to be honest, in certain niches with certain templates, you can

26:58probably get there without touching it. For me, since I'm a perfectionist, I still like coming in and tweaking a couple words here and there. To solve for this in Sand Castles, the script editor literally works like a Google doc. You can read it, make adjustments, it'll save automatically. You could highlight sections and then use the prompt below that would just apply the edits to the highlighted piece and leave everything else alone. If you want to apply edits to the whole thing, highlight nothing and write those. You can change the length. You can swap the hooks. You can swap the facts. You can look and review the previous versions to

27:26see the diff mode to see what's changed. We've really built every piece of functionality you would want in the script writing experience in here. Even though the first generation is good 95% of the way out of the box, this AI plus human combo using the AI and preset template hooks and styles plus my own manual editing to go the last mile is my favorite workflow right now for script

27:48writing for short form content. Typically, this entire flow for the scripting takes about 5 to 10 minutes per script. [music] And again, if you want to go full systemization, you could of course have your content operator do this as well. All right, step four is the recording. Now, today I'm not actually using any AI tools in my recording process. I might look robotic, but hopefully you can tell by these hands, I'm a human. And I'm not using AI in the recording because the costbenefit analysis doesn't really make sense for me. Despite not using a teleprompter and still reading my scripts one line at a time, the whole recording process for a

28:18short form video only takes me like 10 minutes. But the alternative, using an AI avatar with Hey Genen and a professional voice clone with 11 Labs, would only save me those 10 minutes. But the cost would be me having a much more robotic monotone delivery in my videos. Now, I know several creators that are fully AI generated when it comes to

28:35short form and it works really well. They don't even log into their Instagram. They have an editor somewhere. That editor has access to their AI avatar and literally just prints videos for them saying the script that they type in. Verun Maya, Rowan Chung, Nick Serev, they all do that and they're crushing when it comes to views in the tech and AI niche. But personally, I haven't been able to get this AI avatar, AI voice workflow working in a way that makes me feel like I'm still delivering at the quality and

28:59trust that people have known to expect. I will say a huge pro of figuring out AI automation is that it allows someone like your content operator to actually do the scripting and recording for you without you needing to log in. So today, I'm kind of a bottleneck in the human process. If I have a horse voice or my eyes are swollen, I then can't record and that slows everything down. But with the AI avatar stack, you don't really have that problem. you can always have the exact avatar delivery that you expect. So, for beginners, I don't recommend offloading the full recording process to an AI avatar. It's a lot to

29:28take on before you have all the rest of the variables dialed, but as you get better, if you really don't like this recording process and you want to save 10 minutes per rep, [music] this is something you can look to include. All right, step five, the final step in the workflow is the editing. And before we get into the editing, let's just recap how much time we've saved in the workflow thus far by using [music] this kind of like AI plus human system that I've walked through. If you're going fully manual without a content operator, you are currently at 1 hour per week on the ideation, 3 to five, 10 minutes max

29:56on the research, 10 minutes max on the script writing, and then 10 minutes on the recording. So, this is about 30 minutes total per rep before the editing, not counting the ideation, which is a lot of hours saved from how you were probably doing it before if you were manually doing everything. If you have a content operator and you're still recording manually, then the only time you're going to have to spend per rep is 10 minutes on the recording. If you have a content operator and you're using the AI avatar, so far you've spent zero minutes in this process at all. And this is really what true systematization

30:24looks like. Now, when it comes to editing, this is the final frontier. I predict we're going to have lots of people that have systematized this whole workflow up to the [music] editing, but they're still going to need a human in the loop to do that last piece because there is not a great AI tool stack or workflow to do the editing automatic in the same way as the other things I've talked about. Now, there's a lot of startups trying to crack this, and there's a lot of pretty average automated editing tools, but none of it will cut above the quality line that I know you think you need. So, if you're

30:49trying to go full workflow optimization, you've really got two paths on the editing. Path one is to edit it manually yourself if you have no resources, but just time, but go with something lower, green screen, yap, fewer cuts, more value, more kind of like longer delivery to minimize the time in. These types of edits tend to work better on Tik Tok over Instagram, especially if you're in a highv value education niche. If you do this, each edit will probably take you 30 to 60 minutes. And obviously, you can scale that time up with more quality if

31:17you want all the whisbangs and flashes. Path two is to hire a human editor in the loop to do the editing for you. The harsh truth and reality is this. No content system is designed so that the creator is editing manually themselves. That's just not how you save the most time. And I totally empathize that everybody watching this is at a different stage. You may not have the resources to have an editor. You may

31:37have tried but have had it churn out. The reality is you're going to pay for the editing one way or the other. You're either going to pay with time or you're going to pay with money to buy back your time. So once you have the resources to buy back your time, you should do it because editing is the thing that you can replace and save the biggest chunk the quickest. Now, if you want to hire my editing team that does these YouTube videos and also every single one of my shorts, responsible for like billions of views on Short Form, I've got a link below where you can apply and I'll

32:02connect you with that exact team. But even if you don't, the last mile of your content workflow to really extract yourself out is going to be figuring out how to delegate away the editing from you. Now, as soon as there is an editing workflow that's feasible with AI, I will be the first one on this channel to tell you what it is, to show you, etc., but I

32:18don't expect that to come for a while. Now, one small tip I will say in terms of working with an editor to help empower them with AI. There are, of course, tons of AI image and video tools that can be really helpful for generating visuals used in your edits much faster. things like Nano Banana, Veo3, Cling 2.1, Midjourney, like these are all the staple tools for AI image and video. My pro tip for you and what we do on our team is have a virtual credit card that you give to each editor that allows them to go freely use and try any of the AI tool subscriptions

32:47that they want. This gives them the freedom to experiment and also builds into the workflow Veo3, Midjourney, Clling, all these tools that improve the quality of your edits and the speed that they can do it. So, it's not an automation workflow with respect to AI replacing them, but it is turning your human editors into 10x machines. So, there you have it. This is the latest and greatest as of today, my best five-step workflow for systematizing content with AI tools and different human processes to speed it up. I went from spending 4 to 6 hours, sometimes 8 hours per rep to spending only 15, 20,

33:20maybe 30 minutes max each rep myself. And then I have a team that's helping to fill [music] the rest of the gaps. Again, everybody watching this is going to be in a different financial position, a different place in the journey. You may not have room for the editor. You may not have room for the content operator. Just know this. I personally spent the first 18 months, the first 300 reps editing myself. I did all of this myself manually and then I edited 6 to 8 hours. I couldn't free myself up to really fulfill my potential until I was able to hire myself out of the editing

33:49and move forward from there. So, as soon as you have the funds to be able to hire an editor and ultimately a content operator, I would do it. But even in the meantime, using tools like sand castles and chatbt will save you significant time, even if you're doing the manual editing yourself. As always, guys, I'm trying to bring the absolute best, the

34:06sauce, when it comes to content, right? I'm trying to be the guy for short form video and long form video that you look to that just gives so much value, you can't ignore the video. If that's you, if you're a business owner and you're trying to use content to grow, you know what to do. Make sure to subscribe to the channel, leave a comment, tell me what you want to see next. And of

34:22course, I mention this in every video. I've got a free community for entrepreneurs and business owners. It's called Wavy World. I've got 65 free trainings like this that are all organized for you. Everything from YouTube, short form, ideas, algorithms, editing, all types of tutorials is in there. And we have over 37,000 entrepreneurs that are all working

34:39together to improve their own content. So, if you want a peer group, if you want more free training, if you want more free access, definitely make sure to join. and it's called Wavy World. And I've got a free invite link for you in the description. All right, we will see you guys on the next one.