How To Grow A Lead Generating YouTube Channel With AI | Jerry Potter

Most YouTube channels are a grind, hours of scripting, filming, and editing. And for what? A few hundred views. But what if AI could do the heavy lifting? Not just speed things up, but actually help you to build a channel that brings in leads on autopilot. I bought on Jerry Potter. This guy knows YouTube, and he's figured out how to use AI to grow channels that don't just get views, but they actually generate business.

In this interview/workshop, Jerry's going to break down his exact process on how he uses AI across the whole content workflow and how you can build a YouTube channel that works for you while you sleep. If you want a system that turns content into leads, stick around until the end because this one's going to change how you think about YouTube. Thank you so much for being here, Jerry. We're so excited to have you.

Yeah, I'm excited to to share this stuff because AI really has made YouTube more accessible for more people in a way. Um, you know, when I started, we couldn't do a lot of this stuff. But I do think one thing that's important to keep in mind with this just realistically is efficient doesn't always mean effective, right? Those are two different things. And so, one of the things I really want to spotlight as we go through the workflow today is which things are worth doing with AI and which ones aren't. Because just because AI can do something quickly, if there's no impact, there's no effectiveness from it, you know, like scripting a video that gets no views, well then what was the point? So that's one thing I want to definitely emphasize as we go is where AI can really make a difference and then where you should just, you know, do it the way that we've always done it.

>> I absolutely love that. Thank you for sharing because yeah, that is the thing. You can use it and rush through and make something, but yeah, it's knowing the right things to do. So I literally love that. And what do you think are the biggest takeaways that people are going to have if they stay and watch till the end of the video?

>> Well, we're going to go through step by step from starting with how to get your ideas uh right through to scripting to recording um even how to use AI for editing and even making your videos better after they've been edited. So, I think the biggest thing that people are going to see is where there are opportunities to either speed up your YouTube workflow, which could lead to you being able to produce higher quality videos in the same amount of time, uh, produce more videos in the same amount of time without sacrificing quality, or if you're not doing anything right now, like you created YouTube channel so that you can embed videos on your website or something, uh, a way to actually approach it and do it quickly and effectively.

>> Oh my goodness, that sounds amazing. I am so excited. I literally love your channel. So, I'm super pumped that like, yeah, we get to have this conversation with you. So, yeah, let's dive in.

>> All right. So, the first thing that we do with YouTube with anything is you need an idea for your video. What am I going to talk about? And this is one of those things where you don't want to use just plain old, you know, chat GBT like, hey, give me some video ideas about knitting or give me some video ideas about how to grow a business or anything like that. We have to remember that on the back end, these large language models are all just based on the average of everything that's on the internet. And so anything that we do that's completely generic, and you'll see this theme as we go through the steps here, anything we do that's completely generic is going to be what anybody else can do. And so if you want to be successful on YouTube, you can't just do exactly what everyone else is doing. So this is one of those things where again, just plain AI is not the right thing. Plain AI is efficient, but it's not effective. But there's two ways that I want to show you how to come up with ideas for your YouTube videos using AI, which can be really, really effective. And the first one, if uh we can share my screen here, is using chat GBT to do some deep research. Now, you can also do this in Google Gemini, but a deep research project, if you're not familiar, is basically where you ask ChatGBT to go out and spend, you know, 5, 10, 20 minutes, sometimes longer, to go all over the internet and gather a bunch of information. Now, this is one of those things that would have taken a human days before and now AI can do it in seconds. And so, the example I've got on the screen here, uh, basically I've attached a document, but you can, you know, get the information in there however you want. And this mine is says YT90 company info, which, you know, I teach people how to grow a YouTube channel in 90 minutes a week, so we call that YT90. But basically, you have to give the AI something about the goal of your channel or your business. And the prompt that I use is the attachment called company info is about my business. I want to know what the top questions ideal clients for my business are asking about the problem that I solve. So, our goal here is to not just guess or start making videos about a single question because somebody asked it once, you know, last week. So, you think, "Oh, well, that's top of mind. Let me do that." We want to figure out what are the topics that people want to know about urgently because those are the ones that people click on. If you're making videos that are like, "Oh, that's interesting. I might want to know that or I might need that someday." people don't click on those videos or if they click on them they don't keep watching. But if we tackle the problems that are literally people are it's keeping them up at night or they're worried they're going to get it wrong or they are worried they didn't get it right when they did whatever that task was. That's going to be really compelling. So we go through we ask it to look for it on social sites uh like Reddit, Quora, other places that they might ask questions. Reddit is incredible for this. uh find as many as possible. Then categorize and rank stack them by frequency. Format the report with a table top at the top summarizing your findings. Then make each category a heading to be followed by a quick summary of what you found for that category and about a dozen bullets of direct quotes. So we want direct quotes because we want to know what the actual person, what our ideal viewer is actually saying. Um the next section here is basically I ask it to only go back 18 months. Um, if you were doing a channel about AI, you might ask to to ask it to only go back three months or six months. If you're doing something that hasn't changed in a long time, then you might have it go back five years. Um, then uh we say to make sure we're on the same page, please respond with a quick summary of what my business is about and before you start doing research, is there anything else I can provide you to help do this task at the highest possible level? So, you want to make sure you have deep research turned on. We're going to hit enter. And by the way, we're not going to wait for the deep research. We'll come back to it, but I just wanted to show you this first part here. So, if it operates properly, which AI doesn't always do. Um, it goes through and uh it says um verifies what the task is and then let's just Okay, looks like it's it's just going to have a start. So, we'll just go ahead and hit start and it'll start researching. So, I did ask it to kind of come back and say um, you know, hey, clarify that you understand the task and and what my business is about. It decided to skip that. Maybe that's a new, you know, chat GPT 5.2 update, but either way, it should come back and give us what we're looking for. So, I really love that if you're starting a new channel, this is what you should do first. Or if you're doing a channel and it's not working, start with this because what you want to do is you want to start making videos for the same person, solving the same general problem over and over and over again because what that does is it trains the algorithm who your channel is for. And once YouTube understands that, they'll actually start to recommend your videos. So, you mentioned in your intro, Simone, you know, a couple hundred views for, you know, was that even worth it? When YouTube knows who your videos are for, they want to show them to those people. And so, this is the key is to get YouTube going, "Oh, now I get it. They are helping." Like for me, he's helping, you know, entrepreneurs gen generate a create a lead generating YouTube channel in as little time as possible as opposed to before when it might have been like, oh, I upload about YouTube and Facebook and occasionally I post something about my kids and, you know, horseback riding and hiking and whatever else that YouTube doesn't know who to recommend to at that point.

>> I literally love all this. This is so interesting. Does this cuz this doc firstly, I used to do this research manually by hand, so I'm loving that it does this for you. Um, but does this work?

>> Yeah,

>> I really did back in the back in the old summit days, like, you know, five or six years ago, I'd be there looking at the the top performing videos to try and just get like ideas and stuff like that, for summits

>> and other like online events and things like that. So, this is amazing to me. Does this technique work for any sort of channel? Like say you wanted to build a personal brand or a face faceless brand or an interview brand. Does it work for pretty much everything?

>> If you're trying to build a personal brand on YouTube, you're going to do best by building it around one thing to start. So, if you're trying to come out and be, you know, Beyonce or Rihanna or Taylor Swift and have a personal brand and you're going to sell music and movies and perfume and fashion and all of these things, the algorithm is never going to figure out who you are. Now, it can work, but it's going to take you a long time. And you're going to probably have to have that like X factor where people really, really like you. So, I think no matter what you're building, you're always better off making specific videos for one specific person, helping them with one specific thing. And so, in entrepreneurship, that's usually solving a problem. But even if it wasn't that, you'd be better off, I think, making videos like, let's say I was uh I just wanted to be an influencer. I'd be better off making videos all about my life as a dad or my life as a hiker or something to start and then expand later once I've started to build an audience.

>> Amazing. Thank you.

>> It's just

>> Yeah. So, the next part too is um you know this is great if you're getting started, you have no information, but here um once you have a channel then you can go into your channel and YouTube has rolled out this for everybody. It's called Ask Studio and it's basically like chat GBT built in to your YouTube channel. And one of the things that chat GB2 can do or sorry, one of the things Ask Studio can do is give you new video ideas. So I just opened the box here and there's a few prompts here and one of them is suggest new video ideas. So I'm going to click that and what it's going to do and we'll look at what it says in a second. I don't obviously know exactly what it's going to say because it's AI, but it actually goes through and it suggests videos based on what your audience is already watching. This is something that obviously chatgbt, not even Gemini can really do because it doesn't have access to the data unless you were, you know, to export your analytics and all that kind of stuff. So, that's what I love coming in here. So, it comes back and it says, I see you're looking for new video ideas to keep your momentum going based on your recent uploads. Your audience is so and so. Um, still your channel's still building its full audience history. I focus my analysis on the common questions, struggles, and topics found across the 117 comments on your recent videos. So it goes through and it shares different video ideas that you can make. Now one of the things that I think is always important with AI is I don't trust anything AI says unless AI can tell me why it chose it and then I agree with that logic. So, if it comes back and let's just say it said, um, you have really popular videos about YouTube, why don't you make some about Tik Tok and I don't want to be known for Tik Tok, right? I'm not teaching Tik Tok. I'm not coaching business owners on how to grow on Tik Tok. So, that would be a terrible idea. But let's say it came out and it said um, you know, here it says make videos about advanced YouTube shorts workflows. Generally, with the entrepreneurs I work with, I say don't make shorts. It's not worth it. We want to focus on the long form videos. But if the suggestion here was a strong enough reason, then I might consider doing that. So I always suggest like it gives you lots of information here, but I always suggest, you know, pushing back. So I could follow up and say, um, which one do you think would be most successful and why? For example, AI can have typos. Oh, I can't help with that. Oh, I haven't seen that. Um, which let's try which one would you make first and why? Unless it was my typo that made it unable to answer. Sometimes with AI, you ask it the same question again and it's like, oh, never mind. I can do that now. But we'll see what happens. But I always like to see with any of these things like why are you telling me to do this? And if I go, oh yeah, that's a good idea. That makes sense. Then then I go with it. So yeah. So now it came by and suggested, you know, the AI powered editing workflow is the top priority. I think it's listening to us right now. Um, it says over 96% of your total channel views came from long form videos, not shorts. So it's actually advising me away from the shorts video. Um, which is interesting. And then it goes through and says why this is the one that you should make. So, so ask studio is a great way to get ideas once you already have a channel and you've got some history for it to work with.

>> Yeah, that's amazing that you can just do that in there and it takes like real time analytics and and talks to you about it. Like I love that. Thank you for sharing that.

>> Yeah, absolutely. Um, so next step is you know a lot of people think okay well now I've got to do keyword research and I think it's important to know that keyword research was once everything on YouTube. I mean, there was a time 5 years ago where I knew that if I could find a keyword that people were searching for and nobody was making a good video about, I could make a video that was going to be a hit. Um, it was, you know, like accurate 90% of the time. Well, because of AI, YouTube's algorithm got a lot smarter in the last couple of years and it learned synonyms. So, basically, if you were going to make a video called how to start a podcast or how to create a podcast, now YouTube's algorithm knows those are the same videos. So keywords don't really matter in the way that they used to. So keyword research, I mean, I'm not going to argue that it's dead. There's still something there. But the next phase of creating YouTube videos is going to be more along the lines of finding out what there is actual demand for the topic itself as opposed to the exact keyword. Now, if you do have access to a keyword tool, you can go through and upload the data and have AI and analyze it. That can be helpful. But what we did in that first step with the deep research like that's where we're going to find where our good videos ideas are that there is demand for um because it's already been proven based on what people are already talking about on Reddit and Kora and public Facebook groups and different things like that.

>> Literally love this. So interesting. Okay, so what do we do then? We've got we've got our ideas. What where do we go from here?

>> Yeah, so um next up is going to be titles. Um, this is one of those things where you can ask ChatGBT for titles for your videos, but I have found a couple of problems with this, and this is true for titling anything. Um, one, ChatGBT is really good at creating titles that sound good to the human ear. They sound good to our brains. Maybe they rhyme, maybe they have alliteration in them, you know, things like that. And so what happens is we go, "Oh, that sounds really good." And early on which at GBT I found myself doing that not only with YouTube titles but workshop titles and things like that and then the next day I would go that doesn't even make sense like that like that's catchy you know like I had this thing that I was going to add to my offer it was called channel booster checkpoints and I was like oh I love the alliteration you know channel and checkpoint it sounds so good and then a couple days later I was like well that's not even what it is I don't why why did I choose that so I usually advise against just getting um title ideas from chat GBT in that way like just hey give me some ideas. But one thing that I think is really effective with um J with any Gemini or whatever tool you're using that Chachip is good at is improving your titles or creating titles if you give it enough information. So I think there's two ways you can do this. Number one is um give it the title you're thinking about and then ask it to add more emotional stakes to it. So, for example, if I was making a video called how to use YouTube end screens, then I might ask chat GBT, hey, how do I add more emotional stakes to this? Um, then it might come back. I mean, I guess we could do it in real time, but I've this is actually a video that I just brainstormed the other day, so I already know it came back and it was like it was like how to use YouTube end screens. Um, and then it added parenthesis like don't get this feature wrong or you know how not to use YouTube end screens. I think was one of the other suggestions. But when we can add a little bit more emotion to our titles, then people are more likely to click on them. And then the second thing that um AI is great for is if you describe the way you want the viewer to see or feel when they see your title, AI is really good at suggesting titles for that. because, you know, I'm a big human psychology geek. But AI is AI understands humans, you know, ironically enough, really well, right? That's one thing that by reading the internet, AI can do really well. And so you can ask it to, for example, say, I want to make a video where I'm showing people how to use YouTube end screens. And there's all these different settings that they you have to look at when you are setting up an end screen. And I I don't want people to feel confused. I don't want people to feel panicked. And I want them after watching the video to feel this sense of relief. So all I've done is I've said, "How do I want them to feel?" And then chat GPT could give you a title based on that. So that that can work really really well. So Simone, do you want me to just keep going through these steps because I don't want to pause at every time, but I want to make sure that I clarify anything if I if I can.

>> Yeah, absolutely loving this. I that alone I just found super interesting then of how you get emotion into the title to stop to stop the scroll basically like the way you worded that then was amazing. So yeah I'm just sitting here like loving all of this. So feel free to just keep going like you're amazing. This is great.

>> Okay. Okay, cool. Because I I know we wanted to get through all the steps and I know, you know, we have a limited amount of time. So, happy to do that. But I Yeah, chat GBT, you know, if I can share anything or Gemini, if you are trying to like reach and connect with people, as as bad as this sounds, right, if we're worried about the AI overlords taking over the earth someday, they really are good at understanding people. So, that's something that it does really well. All right. Um, next step. So, a lot of times when you're making a YouTube video, you know, what the biggest YouTubers on the planet say is they say they come up with their title and their thumbnail before they even create or outline the video because that is the whole, you know, point of getting somebody to click is they're excited about the title and the thumbnail. It doesn't matter how good the video is if they don't click. You know, I I see it sometimes where I used to have a bunch of restaurant clients and I would I had a chef that I worked with on with their marketing and he would say, "Before I decide uh what is going to go in the dish, I decide how the dish is going to look on the plate." And it was a restaurant that had pictures on their website and and some of the restaurants, they had them on uh pictures on the menu. So, he's like, I it has to look good first, then I decide how to make it taste good. And so, I look at YouTube kind of in that same way. And so when it comes to thumbnails, um, AI can give you ideas decently, but one thing that has gotten so much better with AI is creating images that you can use in your thumbnails. Now, there are tools out there where you can just, you know, give it your transcript or even some where it'll look at your video and it'll suggest a a a thumbnail and even create it for you. But thumbnails are so important that I'm very much keeping the human element of it in there. But I want to show you three things that work amazingly with thumbnails. So, it used to be, let's say you had a, you know, if you've been on YouTube, you know, most of the thumbnails is somebody's goofy face like mine making a ridiculous expression, three to four words, right? And maybe an an icon of some sort. But other people who have budgets to go do thumbnails, like they used to go out and do photo shoots just to make a thumbnail for one single thumbnail or, you know, just take a picture just to use on one thumbnail. And now with AI, we can take so many different shortcuts there. So for example, um this is something that uh my wife and I recently put together a workshop on this, but we we were you know teaching entrepreneurs like how do you use AI to make better and faster thumbnails. So this is just a picture that I took um you know against a plain wall in my house and then I went into Google Gemini and I said please revive this this image so it looks like it was taken by a professional branding photographer by improving the lighting. No ID drift, which means do not change the person's face, which AI loves to do, as you know. Um, do not change the face or pose of the man, only the lighting. And a few seconds later, it's able to come through and create this amazing photo that looks like it was taken professionally. And so, if you compare here, I I'll put a little bit on the screen at the same time. It didn't lose who I was. It didn't take out all of my wrinkles or, you know, my it didn't give me perfect skin or, you know, anything like that. Um, and so that's one thing that it can do. That's really, really powerful. Um, another thing, this is one where I made a video and it was about how much YouTube has paid me over the eight years that I've had a channel because people are always curious like, well, how much how much did you make from YouTube? And with that, I wanted to basically the thing I was trying to convey was if I was trying to live off the money YouTube paid me, I'd be broke, right? I'd be living paycheck to paycheck or whatever it might be. So, I uploaded a bunch of branding photos of me. I said, "Please use the images attached to create a realistic photo of the same man on a white background using a fork to eat noodles out of a tin can. The can should have a white label on it and a YouTube logo, but nothing else. The fork should be black plastic. His face should look like he's slurping the noodles while making eye contact with the camera. His eyes should match the attached image with the mug. This one here, be sure to make his sure his head and hands are in the photo and not cropped out. And it came back and it gave me this picture. Pretty good, right? So, I didn't have to go put a bunch of noodles. I didn't have to go find a can in my pantry, peel the label off, print a YouTube logo, wrap it on there, throw some noodles in, and take that photo. Um, not to mention set up lighting, and find a plain wall and all of that kind of stuff. Now, this can looks a little bit weird. So, I came back and I said, "That looks really good. A few changes. Keep the same man, but please remove the facial hair on the man." That it didn't do a very good job of, but I was using a photo from back when I had a beard, and I don't have a beard anymore. Um, but then I said, um, please change the shirt to be a dingy gray hoodie. So, here I was wearing this nice, uh, what does my wife call this shirt? Chamé, I think that's called. Nice blue button-up shirt. Um, also make the label on the tin can reach near the top and bottom so you can see just the silver rim of the tin can at the top and bottom like a can of soup. I've attached an image of a can of soup for reference. Keep the current white label with YouTube logo on it. Make sure the man's full head, hands, and elbows are in the photo, not cropped. Make the full image a 16x9 ratio. So, that's what YouTube thumbnails are. and at least 1280 pixels wide and 720 pixels tall. So, it took this one here with the weird label and me in the blue shirt and it came back and it gave me this dingy gray sweatshirt. Um,

>> wow.

>> You know, soup label can with a YouTube logo, the whole thing. So, yeah. Isn't that cool?

>> I just can't believe you're you're not even doing that in something like MidJourney. You're literally just doing that in your actual wherever you're using chat GBT like Gemini or whatever like and it's turning out really great images.

>> Yeah. And the things to know you the biggest tips I can give with this is one I want to show you one more real quick but um tell it not to change the person, right? So you stay the same person. Um and then the other thing is the two tools that are working the best. This is in Nano Banana which is Gemini's AI image generator. Um, but as of this recording, a week or two ago, chat GBT in uh introduced a new image one as well, and it's we're still testing it, but it's really good as well. I don't know if anyone else can do it in that way. So, that's where I would play even if you're using a free account. Um, this was one where so I'm taking this now and I put it in Canva and then I make the rest of my thumbnail because it's faster for me to do that than to go through and say, "All right, move the man over to the right and add some text and this and that and you know, all these other things." But this one here, um, I designed this one in Canva and I wanted to make two of the three phone screens cracked. And I didn't know how to do that in Canva. So now with this, you can do things you don't even know how to do other places. So I just basically said, "This has three phone screens. Please make it look like the second one. And the third one have a crack in the screen. Don't change anything else." And it came back and did that. So

>> wow,

>> it's really cool what you can do now with thumbnails. You're not limited by budget um or time in a lot of ways. And the right thumbnail is the thing that is going to get you, you know, far more clicks. And the other thing to know about thumbnails too is the more somebody clicks on your video, like 10% more clicks can lead to like 400% more views because if people keep clicking on your video, then YouTube will keep putting it on the homepage, in the feed, in the sidebar, all of these different places.

>> This is amazing and so interesting. Thank you. Yeah. Okay. This is great.

>> All right, I'll keep going. Um, so next up after that is I look to do script writing. And this is something really important to know with this. I never use AI to write a script for me and I don't think anybody else should. And the reason is that no matter how well you train it, it's very unlikely to either a sound just like you and b I don't believe in scripting videos word for word anyway because even if you are really good and you have a teleprompter, I still think if you are ad libing it line by line by line and editing the pieces together later, it's going to come out better than if you read the script. Again, even if you're professional and have a teleprompter and all of that. So, you know, I like to say I don't use AI to tell me what to say. I use it. so I don't stare at a blank screen. So what I use AI for with script writing is um I use it to organize my thoughts. And so let me um here let's look let's open a new chat GPT. Oh, actually let's go back and check on our uh Okay, so the deep research is still going. I was going to say let's check out on our uh our deep research, but that's still running. So we'll leave that go. Let's open a new chat GBT window. And I actually built a custom GPT that I use with my clients. Um, but I think anyone can absolutely train chatgbt to, you know, teach it to write scripts in the way that they want. Okay. So, I'm going to just hit record down here and I'm just going to ramble an idea for a video and then let it do its thing. Okay, just to give you an idea of what this can do. So, all right. So, I am a course creator and I teach public speaking. Um, I want to make a video about why picture the audience naked is uh terrible speaking advice. Uh, my audience is thought leaders and coaches that want to grow their authority and get better at speaking on stage. And so what I would talk about in this video is I'd talk about um where the audience naked advice came from and why we still hear about it. Um, then I would also talk about why doing that actually backfires and creates distraction and awkwardness when you're trying to speak. Um and then I have a framework called the one person method that I would show which is all about speaking as though you're talking to one engaged friend. Um and then I would also talk about um how focusing on connection beats performance anxiety. So um a story I could tell I could tell tell a story about how in my first keynote I tried the naked audience trick and completely lost my train of thought and then later I practiced speaking as though addressing one close friend. Um uh and it worked way way better and it's worked that way for my clients including one of them who got a TEDex talk that has gotten thousands of thousands of views. So um and then strong opinions. Um I would say my probably my strongest opinion on this topic is that most public speaking advice is recycled nonsense that can do more harm than good. So now I'm going to stop that and you'll see the rambling uh jumbleness here in a second. So this is just what I ramble, right? And when I do this, if you're writing a script, I like to think of, you know, three important things. One is what are you going to talk about? The next one is what are um what's what are what's a story or multiple stories you can share to bring it to life? Anybody can give out speaking tips, right? When you add stories and your own perspectives, then it becomes more unique. And the third thing is any strong opinions you have because again, those stand out. So I'm going to hit uh enter here and let this thing go. And I'll tell you what it's doing on the back end. So I I got to study Hollywood script writing techniques with um a consultant years ago when I used to work in radio and she taught me about how you keep people watching and things like that. So it's going to go through here in seconds and it is going to create this outline of my video for me again using what I taught about it on the back end. And if anyone's never created a custom GPT, it's so much easier than people think. You can do it with a you know using chat GBT's help. So, um, so just like that, I don't know how long it was, 8 seconds, and it went through, and this is what I trained it to do, but you can obviously, um, you know, this is what I use with clients, but you can obviously train it to do whatever you want. But it it starts off and says, "Here's the strategy behind this video." So, this goes back to what we were talking about earlier. Don't trust the AI unless it justifies what it's saying. Then it broke it down into segments. We have hook, segment one, two, three, and then I've trained it to go through and basically open a loop throughout each segment. So you can see here um at segment one. This idea didn't start with speakers. It came from performance anxiety in comedy and theater. So if you're watching that, you're like, "Oh, I'd like to hear more about that." So then it goes through um and then he goes to the next part. Um, you know, this brings us to the real problem. It doesn't just fail to calm you down, it actually makes things worse and it sets up the part most people skip. Oh, what's that? So I didn't say any of that as you saw when I was rambling. It's gone through and it's taken what I came up with and it's put it together. Now, I would never say that piece is locked in and it sets up the part most people skip, right? That's very AI, which again is why I don't use it for scripting. But now I just ad lib off of this and I end up with a really strong high retention video from that.

>> I love this. This is amazing. So, you just would literally have that on your screen and like is it obvious that you're reading something or how do you make that look natural?

>> No. So, what I do I actually just released a YouTube video about this. It's on my channel if people want to see it how I record, but I actually have it somewhere else on a side screen and I look at it and I just remember what a couple of the sections are and then I turn to the camera and I talk for as long as I can until I don't remember what's next. Then I go back and then I come back to the camera and I do it again. So, it comes out so much more natural that way. And because we have like transcriptbased editing now, it's so quick and easy to edit that stuff together.

>> I love that. Yay. This is I feel like you're blowing my mind. This is amazing. I just how like fast and simple you're doing the research, coming up with concepts, coming up with like good thumbnail images. Like I'm just like stuff that literally used to take us weeks is literally taking you minutes and I'm like wow,

>> you know, cuz you're one of the people that used to do all this stuff.

>> Yeah. It's in it's just but even all of everything like even just the titles and the thumbnails and the way you're coming up with what goes in your actual content like the whole thing is just so fast but efficient because it's using data.

>> Yeah. And that's that key. It's not just, you know, it's not just efficient, it's effective because of the data. Um and I'm glad you brought up um research too. So, you know, it used to be if I wanted to make a video about the uh, you know, top three microphones to consider for YouTube or best microphones for YouTube, like I used to have to spend so much time researching, reading reviews, in some cases, maybe ordering the product, which I still will do sometimes, but it was so much time and now I can just ask it, hey, what are the top five uh, you know, microphones for YouTube? Um, give me some in different price ranges, whatever it might be, and Gemini or ChatG will just go out and get them for me. Um, so that's another thing that works really, really well. But yeah, this is um, it really has changed things. I would never have taken the time to outline a video that well before because it would have, you know, because I've always been about, you know, getting things done quickly. It would have taken me hours to think through um, like how to open and close loops throughout the whole thing. I could do it. I know how to do that, but it would have taken hours. And even before AI, I would just basically jot down bullet points. think about a hook, hit record, and start recording. But now, because of AI, I do all this extra stuff, and it just takes seconds and it has such a big impact.

>> Yeah, absolutely. Like the fact that your entire workflow from start to finish for a video is so so fast, but like as you said, like back a few years ago, you'd you'd be like, "Yeah, I should be using, you know, open loops and stuff like that. like I should be using all the things you get taught to use when you're doing speaker training and stuff like that, you know, but you just wouldn't sit there and do it. Whereas now it's like you're just naturally doing it as part of your prep in like 20 minutes.

>> Yeah. Exactly. Exactly. And so, and let's talk about recording and editing, right? Because that's the last part and that's the part where people feel like, "Oh, that takes so long." You know, whenever I run a survey, which part are you most worried about? What's going to take the most time? And a lot of times it's recording and editing. And so if we think about uh first the uh you know the recording AI can't record for you. Um, we are getting closer to be able to make these you know like they call likeness videos where it looks and sounds like you. They're still a little iffy. There's definitely ethical questions there especially if you're trying to build trust with your audience. So I just record. I told you I don't memorize anything. I record in chunks. But one of the things that I like to do um afterward uh that is made a big difference is when I am done recording and this is something again that I never would have done before and now I do is I will go through and I will basically ask chat GBT or Gemini to analyze the um, you know open a new tab analyze the video that I just recorded. So here's a prompt that I use. Below is a transcript from a video I just recorded. Please analyze it. Identify spaces that may have poor retention for the viewer. Let me know if there are entire sections that could be deleted without affecting the main purpose of the video as well as individual lines that can be deleted or recorded more succinctly. So, I used a script for recording and editing. And so, I'm going to go in here. I'm just going to copy the transcript. I'm going to paste it below.

>> And then it takes a few seconds and it does its thing. And I do this now if I have time during my recording session so that I can go back and re-record chunks if I want to while I'm still here, while the lights are still on, while I'm still wearing the same shirt, while I have, you know, my hair hasn't grown out two weeks while I'm waiting to, you know, actually edit the video and things like that. And so it'll come back and it'll it's it's a little brutal. You know, you got to be ready for that. Um, so it'll come back and it will give you um, you know, basically feedback on what you can take out. Now, this video that it's given me feedback on, this is already on YouTube, right? Um, so I I can't do anything about this one, but usually I try and do it, like I said, beforehand and it'll tell you what to do. Now, if you don't want it to be so brutal, you know, sometimes or if you don't have time, what I will say is I will just say something like, "Don't tell me to recut any lines because I'm not recording again, but if there's anything I can just take out without losing the impact, then please do that." And sometimes we're too close to our own content and there'll be a whole chunk that could have disappeared and it would make the video so much better, but we just don't know. Other times it comes back and goes, "Well, you might lose interest here where you're asking people to sign up for something to join your email newsletter for or whatever." And I'm like, "Yeah, but that's why I'm on YouTube is to get people to join my email newsletter, so I'm not going to take that part out." So, that's the one that I like to share um about recording. And then when it comes to actual editing, this is where there are some big things that are happening right now. So I'm going to go back to Descript here. And Descript has a a tool in it, if you haven't seen it, called Underlord. And you can basically ask it to edit things. And it will theoretically go through and edit things. Now, I find that for the most part, it goes through and it edits things in a way that it takes me more time than go back and fix rather than in a way that is actually efficient, you know, or sorry, actually effective. So, I don't use it too much. They have some AI tools that I do use for editing, but I want to um show you something else. Actually, my friend Justin Brown, who has a YouTube channel called Primal Video, just shared this um and he uh this is his prompt, so I can't show you the prompt, but I want to show you what it does, but he basically wrote a prompt to have this analyze your video before you publish it. So, over here is the is the finished video ready to export, right? And with this, we're using Claude Opus 4.5. So, that's one of the Clawed uh models. And we're going to paste in this mega prompt and hit enter. And what it does is it goes through and it actually goes through your video frame by frame. It can actually see what you're doing in the video and then

It looks for things that you could correct. So it could include tightening it up, could include typos on the screen. Uh, could include where it seems like you're supposed to be on camera but you're not. Um, you know, all of these different things. And so this one is definitely a little slower process, but if you don't believe this works and you have the script, like go in and just say, "What color shirt am I wearing in the video?" It will actually be able to tell because that's how good AI is or how creepy AI is depending on how you want to look at it. And it can actually go through and do things frame by frame by frame, which is just, you know, unbelievable when you think about where we were, um, a couple of years ago.

Now, my prediction for the future, like what would be really, really great is if we could get to the point where it would actually go through and make all of the changes that it suggested for you. Or you could say make changes 1, 4, 7, and 9, and it would go through and do it. And I think we might get there. We're just not quite there yet.

So, um, all right, looks like our deep research project is done. So, let me show you this really quick. So, this is what we started with. And so, it came back and said a summary of what people asked in the 18 months. Uh, gives you categories about, you know, their biggest questions. So this was about YouTube for my business. So, time and execution, lead generation and sales tools and help, content and positioning. Um, then it goes through and gives you direct quote questions. So, direct quotes are incredible for titles, uh, and to say in your videos because if I said, you know, if I referred to a question in a way that was completely different than you asked it, Simone, then you're less likely to click on that or want to hear that. But if I said it in your exact words, you're like, "Oh, I've got to watch that now." So, it goes through. So, and it breaks them down here by each of the different ones. Um, the different categories that it found.

Now, here it gives the top 20 question types by frequency. So, how basically how often they were being asked. It also shows you like how many times or how many different places they found these things. Uh, gives you all the data that you want. And the great thing about this is once you have this here, then you can come back into Chat GPT and you can ask it to do all kinds of different things like, um, tell me what keeps the people asking these questions up at night. Things they are worried that they won't get right or that they already got wrong. Like this is one of the ways that I like to think about YouTube videos. We talked about this earlier. So now, like you said, we've got this data set and it will go through and it will find those things and like I said, it's really good at sort of sensing, um, audience sentiment and things like that.

Oh, I did, um, here, let me do this again. I still have deep research turned on, which is always slower. So, let's, uh, let's cancel that and, uh, let's try that one more time. Should go a little faster that way. That's one of the things now is these AI tools have so many different like features and you have to remember to turn them on and off and especially if you're doing a program where you're counting credits or things like that. The, the credit counting in AI can be very, very, uh, stressful. So, but yeah, while this pulls this up, um, do you have any questions about any of that kind of stuff or anything I can clarify or or touch up on?

>> I'm honestly just blown away. Like you've literally put together a YouTube channel video in essentially Chat GPT, Keva for the final, um, thumbnail, then you said you're using the script for all the video, like recording, and then even for the editing. Like that's literally three programs and then you just uploaded to YouTube. Like that's craziness. Like how are you even doing this? How long is a normal video even taking you?

>> It depends on, it depends on the video. Um, sometimes it depends on my time. Um, there are weeks where I just think, uh, I don't ever want to make another video again. And there are weeks where I'm like, uh, this is a really fun video. I can't wait to get it out there. And so, you know, with, with the clients that I work with, we teach YouTube in 90 minutes a week. So, and that's one long-form video a week. And so, there's a few things you have to do, right? You have to say, okay, I don't have to be perfect, right? Because perfect doesn't exist. So, first of all, you don't have to be perfect. That's hard for some people, but you don't have to be perfect. And I'll tell you where we emphasize our time to do it in about 90 minutes a week is we emphasize it in the idea phase. We want really good ideas. We emphasize it in what we talk about in the videos. So, that is, you know, what you're going to share, the stories and your strong opinions about it. Um, and we don't emphasize fancy editing. For example, YouTube did this big study where they connected sensors to people's brains to find out exactly how they were, like, why were they enjoying videos. So rather than saying what do you like about this video, they actually monitored their brain. And what they found was the emotional connection with the person on the screen is like 10 times more important than any technical editing or anything like that. So, I always like to tell people if they can see you and hear you and they like what you're talking about and there's an emotional connection there, you can do really, really well. And there are tons of examples on YouTube of people with, you know, hundreds of thousands of subscribers, hundreds of thousands of views, and they're literally just, you know, holding up their phone, talking, walking through the park, you know, no, not even a microphone, no graphics, no music, you know, nothing like that.

Now, that stuff can certainly help and it can enhance it, but if you want to keep it simple, focus on the idea, what you're going to say in the video, especially in a way that is, it's emotionally connecting. And that can be as much as, you know, like if I was making a video for you, Simone, me saying, remember when you used to have to spend days doing research on this? Let me show you how you can do the same level of research, maybe better, no offense, in about 10 seconds. Right? So, I've tapped into the emotion there. The emotion can also come from somebody seeing I'm wearing a Superman t-shirt and going, "Oh, I like superhero movies too," or, you know, something like that. So, there's lots of different ways, but I, if you focus on those things, that is where I like to call it the ROI, your return on effort. That's where your ROI is going to be the biggest. You're going to put in, you know, for the amount of work you're putting in, you're going to get the biggest return on that effort all the way through. When you start putting effort into other things like fancy editing and things like that, the return on effort is just much, much lower. And so it's just not worth it. If somebody says I have three hours a week, I will say write a better outline. I won't say make two videos. I won't say make the editing fancier. Um, I still, you know, say focus on those things. If somebody's got 10 hours, okay, maybe we change the strategy and you put more time into something else.

>> That was actually going to be my next question is how many videos should we be putting out a week? So you're essentially saying just one video a week. Is there a, um, amount of time that video should go for that's optimal or like I'm guessing an hour is way too long with retention rates these days, but should people be aiming for say 10 to 20 minutes or like is there anything like that that the audience should know?

>> Yeah, this is one of those questions that the, the answer is annoying and the video and the answer is the video should be as long or as short as it needs to be. Um, but other things that you can factor in, one, if you're new or you don't have an audience yet, then you, somebody's more likely to click on a shorter video from a stranger than a longer one. So, that's one thing to keep in mind. Now, there are hour-long interviews, podcast interviews, you know, things like that that people are sitting on their couch and watching like TV now on YouTube. So, there is no like maximum length that we need to worry about. So, I think about first, you know, if you're new, somebody will take a chance on a shorter video. And then second, what is the perceived amount of time that someone would think a video on that topic would need to be? So if I was making a video about how to choose the best microphone for YouTube and it was, uh, 60 seconds long, I might go, well, no, I want a little bit more information than that, right? But if it was 90 minutes long, I might go, okay, that's probably overkill. We're talking about a, you know, a $100 purchase here. So that's where I'd want the video. You know, I, I might for that might click on a video. It probably have to be at least eight minutes and probably no more than 20. So that's the other thing that I factor in when I think about how long I want a video to be.

>> Yeah, that's great. That's absolutely perfect. And I totally agree with you. In the context of something like an interview or something, an hour is totally fine. But as you said, in other contexts, an hour or an hour and a half is going to be way too much. Um, I would also love to touch on as well. I know you said you don't necessarily recommend people are doing shorts. Are you recommending they repurpose the content from the long-form content to use it as YouTube shorts, as TikTok videos, as Instagram Reels? Do you recommend that for shorts?

>> So, that's what that's actually something that I, I forgot I had on my list and we didn't get to is repurposing. And that is something that AI can do really well. But I do think a lot of the times it falls into that effect, efficient, but not necessarily effective. Now, that being said, the tools just keep getting better and better and better at this. And there have been some where I'll run one of my long-form videos in and I don't even think there are any shorts in it, and it'll come back with 10 shorts, and two of them I would publish without even a second look or without even worrying about it. Like, wow, this is a great short. Um, and so I think the possibilities there, but like everything else that we do, either as a creator or as a business owner, every additional 10-minute task is taking away from something else. So, I will say that I have been, I have moved away from short-form video. I try not to watch it because I lose an hour like that and I have nothing to show for it and I'm trying to set a good example for my kids because I don't necessarily want them doing that either. But I think that, you know, if it fits into your workflow, I think it's an interesting way to get out there. However, we have to make sure that we are doing it for the right reasons. And so, you have to think about how much is a short view really like worth to you. Sometimes people in my community, you know, they're just getting started and maybe they're getting, you know, 20 or 50 or 100 views for a video and then they'll put up a short and they'll get a thousand views and they're like, "I got a thousand views," you know, and then you go and you look at the analytics and a view on a short, uh, the way they've changed it recently basically means it showed up on the screen. It didn't even have to be a full second. So, that's one thing. And then two, let's say you made a 17-second short and you got a thousand views. Well, how many people actually remembered you three swipes later at that point? And so that's why I really nudge people toward long-form. That the last thing is shorts don't have the longevity that that short-form videos do. But again, if you've got the time, you've got the workflow, you've got the tech, you got the team, I think everything is worth testing. I don't want to just completely, you know, bash on short-form video. Um, if you want a presence on other platforms, it's one of the easiest ways to take what you're doing on YouTube and get it other places.

>> I absolutely love that. Thank you. Thank you. Is there a program that you do recommend for when you're turning those long forms into short forms that you want to just throw out there or?

>> Yeah. Um, Descript does a pretty good job. You can go in, they have a one-click button you can do or you can use Underlord to do it. Um, another one, Opus Clip seems to have all, it was one of the first ones, but it seems to, like I try and test them every few months now. There's too many to test. Like I think there's a hundred different programs that claim to do this, but Opus Clip has always seemed to have a really good knack, like whatever their algorithm is for finding the right clips to put out there. Um, and so those are probably the two that I would recommend. The last thing is if you're already using another tool, like a lot of people around YouTube use VidIQ for various things and VidIQ will do it for you now. It's just built in. So if you're already paying for something or Descript for that matter, then you know, go with that first before you go and pay for another tool.

>> I love that. Thank you. Thank you. Thank you. This has just been so interesting today, like and where, because where can people find you if they want to know more information?

>> Yeah, if you want to learn, uh, well, you can find me on YouTube. You can just search for Jerry Potter and my channel should come up. It's called Done By Lunch. And, uh, secondarily, if you want to learn more about YouTube in 90 minutes, just go to yt90roadmap.com. It'll talk to you, it'll, it'll share basically how to get started doing YouTube more efficiently and effectively, not just efficiently, as well as how to know if YouTube is going to be worth it for you or if it's going to be just a total time suck and in six months you're going to regret doing it.

>> I absolutely love as well, just to, um, go back to this is how you have consistently said, do the thing that's adding like value, you know what I mean? Like if it's going to be a time suck, don't do it. And I love that you've been saying the whole way through, make sure you're spending time on the important things, which are the thumbnail, the title, and the brief as well. So, I love that you even just brought that up then and were like, "Is this valuable to your time? If it's worth it, do it. If it's not, don't." And I love that you teach all of this in such an easy to understand way for everybody because you've literally just broken down the process of making essentially a successful YouTube channel, cuz you've got the data there to back it up, in literally half an hour, you know, or a little bit over. But like you've just made it so easy and so clear, but also so tangible for people to be able to go, okay, this is where I should spend my time. This isn't where I should spend my time. Is this where I should be spending my time? And so just even that is just so much gold for the audience. So yeah, this has been amazing and I just really appreciate everything you've said today and shown.

>> Thank you for saying that too, because ultimately my mission is, I want to make YouTube accessible for anyone who wants to be there. And for so many years it was like, "I don't have time, I don't have time, I'll start next year, I'll start next January," um, you know, it was like eating healthier, I'll later, later, I'm too busy, I'm too stressed, you know, whatever it might be. And I think we are at a point now where YouTube is so much better for smaller channels and giving them a chance. And technology and using AI has made the parts where people used to get stuck no longer an issue. That anybody can start YouTube and even start it as a side hustle. You don't have to wait till you have 10 hours a week anymore. And so I love just spreading the word about that and hopefully inspiring people to take that step.

>> Yay. I love all this. Thank you so much. And yeah, we will share all your links obviously underneath the video so people will definitely be able to find you. But go and check out Jerry, guys, because he is amazing.

>> Thank you for having me, Simone.