Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Integrating AI for Seamless Podcast Production
265 Plays1 year ago
In this episode of the AI-Driven Marketer, Dan Sanchez sits down with Lee-Ann Scott to dive into the evolution of marketing tools, spotlighting the transition to user-friendly platforms like WordPress and Canva. They explore the burgeoning potential of AI in automating podcast production, from initial client interactions to post-production tasks, and discuss the role of custom GPTs in breaking down complex projects. Together, they address the productivity boosts these tools offer, the practical challenges of integrating LinkedIn with HighLevel, and the importance of scheduling deep thinking work. Tune in to grasp how AI can transform your marketing game and streamline your workflows.
Timestamps:
03:35 Automated LinkedIn outreach using conversational bots.
09:03 Networking and podcast appearances offer valuable insights.
11:37 AI clipping tools find almost perfect margins. Allows easy archiving, editing, and publishing.
13:32 Editing, production, upload, distribution - simplified with AI.
16:52 Successful show brings financial benefits, lacks video RSS.
22:00 Edited content still sounds wonky with ums.
25:57 Minimize software handoffs to reduce potential failures.
28:45 Customize super prompts for specific industry use.
32:39 AI will handle numerous steps in digital marketing.
33:32 Podcasting needs marketers to drive traffic.
38:23 Podcast production time reduced through automation.
Recommended
Transcript
Introduction to AI in Podcasting
00:00:02
Speaker
Welcome back to the AI-Driven Marketer. I'm Dan Sanchez. My friends call me Dan Chaz, and I am on a journey to master AI in 2024 because it's clearly something that's coming for all of us, including us in the podcasting space, which is a topic that is near and dear to my heart, which is why I'm talking to our guest today, Leigh Ann Scott, about her podcast in a box concept where she's using AI throughout the process in order to automate like the whole podcast production process. So Leanne, welcome to the show. Welcome. And thank you so much for having me.
Custom GPTs in Podcast Production
00:00:36
Speaker
I'd like to start with custom GPTs. Cause that's, that's one of my favorite AI tools and something I'd love to figure out how you're leveraging throughout the system. Uh, how are you using custom GPTs with open AI in order to compliment thing, other AI tools that you're using throughout the process? Well, the big thing about custom GPTs is that it can do anything that you wanted to do as long as you train it. So.
00:01:01
Speaker
You've kind of got to think about training it as you would an extra person on your, um, on your team. And so when we look at bringing a client on board, we make sure that we teach it everything we needed to know about content pillars, about tone of voice, about little nuances on the way that, um, people say things so that it becomes more authentic. If we use the same GPT every client, it would mean that we'd just be spitting out more AI dribble. But if you can learn how to leverage it to actually really act like a client, then you know but the world's your oyster really. And so that's where we're using the GPTs to really get into the heads of our clients, but also to assist and make sure that the quality of content that comes out is much better.
00:01:57
Speaker
That's huge. Um, what parts of the process are using it in? I love custom GPTs because of how, how tailored you can make it. Um, but are using it on the pre-production end or more on the post-production end? Oh, uh, we, we use it right from the start. So from our very first meeting with the client, um, we start using it. So when we put together a proposal, we have taught it all kinds of things on, um, how we, one thing, how we want to present our services and then how we tap into that. Then we look at once they've actually done the kick of call, we teach it to sound like our clients. And then the next stage of that is once we've recorded a podcast is creating the posts and um all the content that goes with it, the announcement content. And then we can use it in some of the analytics as well.
00:02:52
Speaker
That's huge. I mean, I even like how you're using it. It sounds like you're using it for internal purposes too. You're like after that kickoff call, we start to leverage it to streamline your processes on the backend for your own, for your own purposes, let alone for your clients. Yeah. So let's start breaking down the podcast in a box, like step by step. Like what are you doing on the front end? A lot of podcasts are guest based, kind of like the show. You know, we do a lot of, had a lot of guests on the AI driven marketer and I do a lot of solo episodes as well, but like figuring out how to handle guests is its own, its own thing, right? There's so many little steps just to get to the point where you're actually showing up and recording the interview. How are you automating the process and the beginning of the guest relationship?
Automation in Guest Booking
00:03:35
Speaker
So we do a number of things. One is we do some automated outreach on LinkedIn. um So that's where we kind of use conversational bots so that we can get the volume. um not you know Some people are easy to book and others are not. ah one different industry organization Once we've got the conversation going, we can they we then give the guests the opportunity to book either direct with a calendar link or some people would prefer a conversation. ah So the bot can help with that conversation piece, ah getting them booked in, instead of this backwards and forwards saying, well, can you do next week Thursday or maybe next week Wednesday? um So we just help with that process. Once we've got that, we then have the reminder sequence. So like any workflow has got appointment reminders. I think if you think of Calendly, you know, they've it's got all the
00:04:31
Speaker
appointment reminders, but why would you have a separate a calendar and reminder system for that, where you can actually feed that into um ah lead scoring within the CRM and all of that. ah Once we then have them booked in, we remind them and tell them what to expect on the day. On the morning, we make sure that they're all ready. We give them the reminders. Remember to switch off all notifications. um Make sure that you've got your lipstick on, that kind of thing. um And remind them that it's a video podcast as well. ah So from that point, we then have the guest prepped ready to go.
00:05:12
Speaker
um And obviously, like you know, getting them to show up is is half the battle. ah so So I think it's part of the process of finding the right guests, getting them booked in and making sure they feel comfortable and prepared on the day of the interview. Yeah. So are you using just email to send out the reminders before the interview or using text messaging or something else as well? Oh, it depends on the client. So you know, with with high level, you can use email or text messaging. ah So we depending on the client, we can do both or just email. Okay. And it's interesting, you're using a LinkedIn automation tools, there's plenty of LinkedIn automation tools out there. um But you're using high level for the appointment booking, I assume for the actual like calendar
00:06:00
Speaker
embed widget, whatever they're putting on their website.
LinkedIn for Podcast Outreach
00:06:03
Speaker
um yeah How do you get the two to talk to each other? If they give them the option to go to the appointment link, that makes sense because on there you can ask the questions and get what you need. But if you're having the LinkedIn automation tool a walk through the the questions, how do you get then do those tools tools sync together? Oh, now you're asking for all my secrets. and LinkedIn is one of those and um anomalies in that if you think of Instagram and Facebook, they love to talk to high level. They're a pain in the, but ah they do talk to high level. When we look at LinkedIn, LinkedIn doesn't want to talk to high level yet.
00:06:39
Speaker
ah So yes, we need to have an integration and there are different things. So we can use expandy and then import the information in. um We get to a point where we can push them over onto the link on the website. So it's it's one of those things that if I could wave a magic wand, LinkedIn and high level would play very nicely together. Yeah, that is tough. LinkedIn tends to be a black box, which is why all these AI tools use the method of like pretending they're like going through a a digital browser in order to automate a lot of these things, which is kind of an interesting route to go because LinkedIn's been so over, over automated and spanned by so many.
00:07:18
Speaker
yeah where I guess Facebook's okay with it, but maybe the level of automation people hit with LinkedIn is is more. ah Sometimes I feel like LinkedIn's like okay with it though. It seems like they're okay. They're like, we didn't like this. Keep going. Exactly. They must know the tools are there. But I think it's how people use it. they limit They limit the number of connection requests you can make. And I think as long as you use the tools so that you've got a human element to it, They are happy with that. Lots of people complain about being spammed on LinkedIn, but I actually get spammed way more on Facebook than I do on LinkedIn. Interesting. Oh, I get spammed more on LinkedIn, but I also like spend all my time on LinkedIn and not on Facebook. like friends Facebook is literally just my friends and people that I've known from the past.
00:08:08
Speaker
where LinkedIn, I have like 25K connections, so I get spammed yeah more often. More ways for people to find me because ah of the connections. Yeah, for sure. Though I do find when you're doing outreach for a podcast, it's it's much less spammy, right? I mean, because you're asking to like get them on a podcast, highlight them, their expertise. it's I don't know. i I feel like it's not spam. It's something that I've done personally, manually before. So to use a LinkedIn automation tool to like, hey, would you like to be a guest on the podcast? Even if you're automating it, you're automating it and zeroing it in on the kites of people you want and all that kind of stuff. Even if you're using it for prospecting, like, I don't know if you do use that for as a tactic for your clients. You're using a podcast interview as ah as a way to prospect. Absolutely. absolutely i think I think it's one of the best ways to leverage podcasts and it's one of the most under leveraged ways.
00:08:58
Speaker
ah for people to connect and actually get to talk to their dream clients.
Riverside vs Zencastr
00:09:03
Speaker
Yeah. I mean, the agency I used to work for, that was like their main thing. and The founder had even written a whole book on the topic of content-based networking. um And it was such it was so good. yeah ah even And even if people ask me onto their podcast because they think I could be a client, um I'm i yes like, I always take it because I'm like, I don't, there's something to be in on the other side of the mic too as a, as a guest or a host or as a guest rather than a host because yeah people's questions can draw up different insights. You didn't even know you had sometimes and that's valuable in and of itself. Right. Exactly.
00:09:38
Speaker
um So yourre linked you're running LinkedIn automation, you're getting people to book through a calendar, or it completes it itself and follows up automatically, ah which is pretty cool. And you have a reminder system through email, prepping them for the interview, they they then they jump into the interview. Is there a particular software you use for recording? Are you a Riverside fan? Are you using Zencastra like I am? What do you use for recording? So ala I like Riverside. um quality The video quality is really good on Riverside, so I do like it. and Zoom, we found we used to use a lot of Zoom because I'm just comfortable with Zoom, but the quality of video is so poor. um And we just had too many complaints, so we don't use that anymore.
00:10:25
Speaker
What I like about Zencaster is that it kind of does almost everything Riverside does. there's I wish there was a few things Riverside does that I'm like, I wish Zencaster supported like the ability to come in as a producer into the room. like I don't know if you've ever like been in producer mode where your video isn't mixed into theirs. You're just there to kind of like talk people through the process. I miss that because I've used Riverside for that before. And Zencaster, I just have to like turn off my video mic as a producer if I'm doing that for a podcast. Um, the thing I love about Zencaster though, is that it does everything like it records, you can edit, do basic edits, like text editing, uh, like descript, but within the product itself. And then it also does the, it also is the host. So I don't, I never have to download a file. Oh, okay.
00:11:09
Speaker
It's all in one, it's all, it's an all in one thing where it goes from recording, editing, and production, which means I can turn it faster because the process of even just downloading re-uploading, that's a lot of time. It's most time and I have a high speed internet and it's still a time consuming process when you're dealing with like 30, 45 minute like length episodes. It also does good on clips, way better than Riverside's does, I found. and um yeah that That would be enough reason to to change over it's its ability to find the right clip. So you still have to kind of like, you know, for some reason, all AI clipping tools, like miss the mark by like five or 10 seconds on the front end or the back end. For some reason, you're like, exactly exactly there's like this like 10 or 15 second margin on the front end or back end where it's almost perfect. Um,
00:11:56
Speaker
So that does pretty good. And then I think the last thing that Zencaster does is on that clipping side is that it ought you you can literally just take the clip and either archive it because it was no good, edit it or just one click, like put it in the queue and it publishes to all the socials that you need. Amazing.
Podcast Distribution Challenges
00:12:12
Speaker
Because that's another part that I'm like, just like every episode is a couple of clips or maybe up to seven clips to up to download those and then repost those on every s si or separate platform, even do it through Buffer. where you can schedule it all at once is another extra step that's eliminated that I'm like, Oh, I hated that step from i I loved video. The video is a great AI platform that does clips, but I, I hated the process of downloading them and re-uploading them. So yeah, that would win me over for that ah really that. Yeah, that would win me over. I reckon I like video.
00:12:43
Speaker
my biggest issue with them is that if you need to make any changes, you have to do the whole thing over again. And that just drives me insane. but could Really? You know, when the client picks up one tiny little thing. Yeah. Yeah, that is that is an issue. And I do find by but i my past agency, they were still doing the captions by hand because It was so paint because kind of just it's unacceptable if you miss a semicolon, you know, ah like in a 30 second clip, you're like, oh seriously, it's social media.
00:13:19
Speaker
but well ah So that's that. I could talk to you about the quirks of clients for the next hour. Yeah. And quickly this will become like Agency Life podcast. Yeah, exactly. Which is a different show, but it is it is part of part of life. um So you're using Riverside, you're recording there. um What what's happening post-production wise do this is where I get really curious because there's so many little Steps in the podcasting world in order to take it from something like Riverside edit it re-upload it to your your podcast hope and then to do distribution which is honestly like I Find that I always Suffer on distribution because it takes so much work just to get it to the end zone to publish the dang thing You know and to get the show notes written in the title and the cover image and it's like distribution is like
00:14:12
Speaker
It should be 80 percent, but it often only gets 10 or 20 percent of the effort because it's so much work just to get the episode out there. So how are you making that simplified with AI and automation? So I think the thing about us is that what we've done is we've really honed in on what are all those multiple steps. And and that's really the reason why people stop doing their podcasts after 10 episodes. Most people stop because there are so many steps that nobody ever thought about when they first hit record. So we've mapped out over 500
00:14:46
Speaker
um standard operating procedures and processes and optimize each and every one of them to make sure that we're as efficient as possible and we can keep up with a weekly cadence. In terms of production and distribution, so we use um we use obviously custom GPTs for ah the production process. We use ah tools such as video, Opus clip we've tried. um and we We've also loved the script because that way we can do our own clipping, but really quickly. um And so we look at ways that we can process as quickly as possible and as efficiently. But what we always tell clients is that you have to realize that this is just a taster. It's not necessarily about finding the perfect clip in an hour long podcast.
00:15:40
Speaker
So that's kind of where we we we look at using the AI tools. And then in terms of syndication and distribution across the platforms, we use automated scheduling. We use the various hosts that will push out across the different listening platforms. um ah We tried because. which was great because it would upload directly to YouTube, but the analytics were not great. um Then we've tried Buzzsprout and we've tried Captivate. Captivate's my favorite. I think the the analytics are the most detailed ye and they're the most user-friendly. I think it's intuitive where things go and how things go.
00:16:29
Speaker
And quite frankly, I just think we can give as much information and now you can bring your YouTube listenership in there as well. So that's our favorite. I swear I've tried every podcast host. Like Captivate's really good. I remember looking at Captivate and for the price they charge, they had some of the most robust features I'd ever seen in a host. Um, so I think it's like what 30 bucks and you get like unlimited shows and then it's, it's, they cap it based on like how many views you get, which is actually a freaking huge deal because most shows don't get a ton of views. But if you have a breakaway show, then you're going to be paying no matter where you go through the, through the.
00:17:04
Speaker
roof or or through the nose, right? um So Captivate's really good.
AI Tools for Podcast Efficiency
00:17:09
Speaker
I think the one thing I didn't like about Captivate is it doesn't support video RSS, and that was my my hiccup with them. Maybe they do now, though. um So they don't, not not so much video RSS, but they they do pull the YouTube link in. and they do syndicate the YouTube viewers. So that's the most important thing for us is that we're able to... That's interesting. So you can upload video and it'll push video the video to YouTube? No, you've got to upload to YouTube, but it will pull it in. And so when you use your captivate link, um it will also show the video link.
00:17:48
Speaker
Okay, so are you uploading to YouTube and it just scrapes the YouTube feed and ah yeah publishes it automatically converts it to audio to be published to all the app, to all the podcast apps? Or you have to upload to captivate and YouTube? Yes, you've got to upload to captivate and YouTube. And then for us, the reason we do it is that when you use your captivate link, you can show the video link as well. Um, but you still have to upload to YouTube. And then the other side, the most important for us is that it will pull in the listenership. So we know what the numbers are across. That's cool. Manually calculate your YouTube plus your other listeners.
00:18:32
Speaker
Hey, that's something, because ah that's the only place then that I know that actually counts it all together, like YouTube views to downloads. I'm like, that's pretty cool, because right now I still have to watch those separately. yeah that's but That's probably why I stay with Zencaster, because i only ah I only like uploading, or at least it was Zencaster, I don't even upload anymore. It just publishes both, yeah one to YouTube, separate one to all the podcast feeds. so Yeah. Yeah. I'm trying to figure out how to like how how to get this down to the least amount of steps possible. So I'm always curious. Exactly. There's so many steps. Like you just said, there's 500 of them. Yeah. um Yeah. We have 500 from start to finish, and it covers everything. So yeah, it's a lot. And I'm sure there's still more that we're working on.
00:19:19
Speaker
Yeah. I mean, it's this is why podcast agencies exist because there's so many tiny little steps and they're not necessarily like hard steps, but there's just so many of them. it's just the say menu yeah exactly Exactly. Especially when we you start getting up to a hundred episodes, 200 episodes um to keep pushing that forward. And then you've got your monetization aspects that you have to look at. So um yeah, it's interesting. So I have a question for you. Yeah, when you are editing um a podcast episode, do you listen from start to finish right the way through every client's episode to find the bits to pull out to find the bits to edit out? Or do you depend on AI editing?
00:20:00
Speaker
Um I depend on I don't know. It's a combination of both like for the clients If i'm editing their show then i'm also producing their show, which means I actually was there live on the episode Um, and i'm there live on the episode specifically so I could take the time stamps Of things that need to be cut out not so much things that need to be highlighted Um because I find that organic conversation works the best like even in this our show that we just had we had a moment where like your neighbor was like drilling the audience didn't hear it because I've already cut that out. um But those are the things that I'm listening for as a producer. So when I do edit it in like CapCut or something like that, or now just straighten Zencaster, like I can just go back, find those tastas timestamps, time stamps, edit it out. I'll use auto. I'll use AI to cut out the silence, which I think is the most use case. I don't use AI to cut out the ums and ahs.
00:20:53
Speaker
because I find that it's a little bit too aggressive with it and cuts out the natural ums that are actually okay, right? It doesn't know where to cut it and where not to, so I don't use AI for that. And I only cut those if like, all like only if it's like really egregious do I actually cut and edit what somebody said. Or if it's a clear do over or a mistake or or something interrupted it. Every once in a while, there's more like, I don't know, like some some hosts like interrupt too much. And so you have to go in and then separate it out, right? So you have to edit it separately. And that takes a long time. But generally, I'm using AI as much as possible and doing very minimal editing. um But I was also there as producer listening. Yeah, so I had one kind. And she's
00:21:40
Speaker
We had to edit out 900 and something, umms and ahs in a one and a half hour conversation. It was a nightmare and you had to do it because you actually couldn't follow the conversation because there were so many umms and ahs. Well with that many umms and ahs, I find what you get back edited still sounds really wonky.
00:22:09
Speaker
you know, because if you're clipping together get to work, really, because if you're clipping out like an um, and all like after every other word, it like when you actually chop it together, even if you like went through the timeline and hair picked the exact second to allow enough space in between the words. It still sounds weird. And i I don't know what to do about it sometimes other than like, Hey, like, if it's somebody I want to build a relationship with strategically, then I'll go ahead and publish it and be like, Hey, it's the best we could do. Um, if it's something where it's like, I wanted it for the content, then I'm just like, sorry, this, there was just too many M's and awesome. You know, excellent.
00:22:49
Speaker
Yeah, it's rare. I've I've probably hosted and produced well over four hundred five hundred episodes and probably only once or twice have I heard some a guest botch it that bad. Yeah. But sometimes they are that nervous. mainly Mainly when you're doing these kinds of shows, you're talking to executives or directors and up usually or pretty good specialists where they're pretty comfortable in talking about the thing that they're talking about. So I don't know if I'll write into that too often.
00:23:19
Speaker
I think and i think it that you 100% correct there. It comes down to the quality of the people that you're interviewing um that you are, you know, inviting onto your um show. So it does help. you know But when you're reaching out automated through LinkedIn, obviously you're going to get some people that yeah not only like don't have a lot to add to the conversation, um but you know they they don't know how to b or be interviewed either. But it's I don't know. I find most people most people do.
00:23:51
Speaker
Quite well. um So that's your process, getting all the way to episode published. yeah How do you handle distribution? Where did custom GPT start to come back in? And then how are you like automating a lot of the little steps after an episode goes live? Um, so there it is. It's down to using things like capture. Um, we don't use, not capture, sorry, captivate. Um, capture is another great tool to look at. Uh, if you haven't already, it can be end up being quite pricey, but it is really good. Uh, so we train our GPTs to do most of that.
Automating Scheduling and Distribution
00:24:31
Speaker
Um, in terms of, I think,
00:24:34
Speaker
Probably once the production process ends, that's where the AI tools stop being as effective, and then we have to do a lot of stuff manually. Although we do schedule posts, so we do a lot of that is automated, so we will use high level for that. Okay, so use high level scheduling tool. Yeah, it's glitchy and they've come a long way since they launched it. um But what I love about it is that it's got the approval workflow in it. So we can upload posts and we can just send off to the client that it's ready for approval. They can approve the posts we've connected to all their platforms and it's done. um So, you know, I've got a very big philosophy done is better than perfect. And so
00:25:25
Speaker
Although, yes, there are other things like Buzzsprout and HeroPost, Metricool, there may be a little bit more, but I think for what we want, because we keep everything on one platform, the workflows are much smoother, much more efficient, otherwise we're having to play with things like Make dot.com and Zapier, and that's when things start going wrong. So we try and keep it as simple as possible, make sure it's done, um even if sometimes that means it's not perfect. Yeah.
00:25:58
Speaker
I actually so share a similar philosophy. I'm like, the more handoffs you have to do with software, the more chances for things to go wrong. I actually really hate using Zapier and I try to avoid it at all costs. You have as much powerful as it is and is is is becoming with their AI integrations. I'm like, if I can avoid it, I will because the more handoffs, every every handoff is ah is a failure point that you then have to troubleshoot when something goes wrong and things go wrong consistently. Exactly. So that makes a lot of sense. um Are you using high levels workflow AI at all? And somewhere in the process? No, not so much. At the moment, we still think through all the workflows ourselves. um And we we do use some of the AI, but you you can't bring your custom GPTs in yet. So that's where it's coming from what I understand. um okay That could change.
00:26:51
Speaker
changed game. But for the moment, yeah we still have to do too many handoffs to really integrate it. And because you can't determine, it's not repeatable. So it's not reliable. So whereas with a custom GPT, you've trained it, you know what it's going to do. But in high level, you're dependent on it, thinking for itself each time. So you can't guarantee what's going to come through that the pipe. I don't know. I think you can, you can train it and teach it. It's a little, the process is a little bit different because you're working through the API rather than through the, the customer. yeah e Um, I just worked through a project that I was really happy with, which is why I'm now interviewing other high, high level users to figure out what else people have figured out.
00:27:37
Speaker
um I built a ah Course on AI that walks people through like a five-day email course, but each one is based on a video So what I did was I took the transcript from the lesson and on the form you have to fill out to get to the course ask for three pieces of information ah Like I mean other than name and email ask for like what's your job title? What industry you in and who's your buyer? And then I so i give it I essentially prompt it it essentially becomes a super prompt which isn't as good as using custom GPTs because you can AI sucks at taking complex projects and break it down into steps, right? Which is why custom GPTs are nice because you can pre-program the steps into it so it can think through. And if you give it individual tasks, it's really good at thinking through it. You just have to flush it out for it.
00:28:22
Speaker
But it's still not bad at some of those robust tasks if you give it information. So what I'm doing for like lesson one, um please take the transcript below and customize it for the job role. And then I'll give it some more context to about like what this is. It's an email course for them to learn AI. Take the lesson from below and customize it based on these three attributes of the the student job title. yeah industry. but And then it will, it'll like take the lesson on super prompts. And it'll give much more detailed examples of how a CMO in ah in the beauty industry could use super prompts versus an email marketing specialist at a staff startup, selling to, you know, veterinarians or something, you know, it'll actually customize it based on that.
00:29:06
Speaker
And it is reliable, but you do have to keep crafting it and just run ah ah a dummy, you know, through the form over and over and over again to I don't know. It probably probably takes a five to ten iterations per prompt in order to get it. Um, but it is possible, which is why I'm wondering, I'm like, even in the guest follow-up, uh, or the guest reminders, I'm like, surely there's something there in order to customize reminders. Maybe not the reminders, but even in the follow-up, you could probably save a step from your custom GPTs for repurposing and build it into. Take if, if you submit the transcript of the episode into high level, which there, I had to test it, but I'm like, can there, can their forms, that form fields actually handle that many characters? They can. ah really Then you can inject that into the OpenAI's API through the workflow. like yeah Based on this transcript episode, inject transcript. yeah Come up with to ah a single ah or maybe like three different tweets around the content from the guest point of view.
00:30:09
Speaker
Yeah. And inject our at our app mention for our shows, Twitter handle into the end of the tweet or something. Hey, here's a here's a template of what we want these tweets to look like. So you provided a template, an example of what it looks like. Yeah. That would be reliable for a follow up email for the guest. Okay. And then you don't have to run it through custom GPs. It's just, you just, yeah, I don't know. I'm trying to think of like, how do you schedule the email to go out for just thanking the guest and giving them some content in there? yeah yeah You can't do it with the video clips. Cause that's like a whole separate process and it can't create the videos, but like for tweets and stuff, I'm like, ah like it you could do it and it should be reliable. That's something I'm still figuring out. I think there's more opportunities to be had in there. Oh, absolutely. I, you know, for me, I'm in my absolute happy place when
00:30:57
Speaker
I can sit for a few hours uninterrupted and just think through the different processes, think through how we can kind of automate a lot of that. so yeah I know exactly what you mean. yeah yeah It is hard and it's time consuming. It's like that important yet not urgent work that gets cast aside so often. A lot of AI work is like that, right? but It is. And i think I think the secret is that you have to make the time. So you've got to set time set aside time for that deep work, that deep thinking work. And that's the only way you're going to stay ahead um with it. Because otherwise, all you're doing is busy work, kind of work. I mean, there's so much to do. And if you're managing multiple podcasts, which you probably know as well,
00:31:43
Speaker
um is that there's always something that needs doing. um There's another step. There's the clients querying something or other. um you know You need to send out more posts, schedule more posts. so you If you don't block that creative time, that deep work time, um I think AI will take over and and you'll get left behind in the race.
AI's Impact on Marketing
00:32:06
Speaker
yeah I just finished reading Gary Vee's new book. It's probably his most tactical book. And honestly, my conclusion from his book, it was called, a um, Day Trading Attention, I think it was called. Um, I listened to it on a long drive I just had, and it was probably the most tactical book he's ever put out. The one conclusion I had from it is that digital marketing has matured to the point where like, look, we know how to do this, but you walk away from that book being like, wow, that's a lot of freaking work.
00:32:35
Speaker
Yeah, right because he breaks down like the tactics of like all all the channels and like we just talked about with podcasts There's 500 steps to this right or there's just hundreds of them whether it's 500 or 400 is just a lot of steps to doing all these things and then split testing and Analyzing it and then retrying stuff and taking more swings So you figure out what works in order to find and win on a specific channel yet times every channel? You read that book and you're like, wow, who can win? You know, in this environment, it's so hard. um So much harder than traditional was yet.
00:33:08
Speaker
I read the book and I came away with the conclusionion conclusion. I'm like, and this is why people are going to use AI because there's so many freaking steps to do digital well, that AI will eat away at all these tiny little steps slowly but surely. so yeah I mean, this is why I started the podcast because I've seen the writing on the wall that AI is going to slowly eat one task and then two tasks and then three tasks, like it's coming. Yeah, see, I think that's where, um you know, for podcasting, that's why they need us. Because there's, there's one thing being a video editor that sees an opportunity to produce podcasts. And there's a whole other aspect where you have a marketer who sees the opportunity in podcasts. And oh yeah ultimately, it's the marketer that's going to win out because they've got the foundations in marketing. So it's not just about producing video and audio, it is actually
00:34:02
Speaker
How do we drive traffic proactively to download the episodes, to listen to the episodes? How do we proactively drive the content flywheel? And how do we proactively add people to the list as opposed to just sending traffic to Apple and Spotify? They have all the enough traffic of their own. They really don't need our help. um What you need to know is how to bring people to you. And I think that's where we can really leverage podcasts, but where we really add value to a podcast host.
Enhancing Creativity with AI
00:34:38
Speaker
It's huge. And that's why I think marketers are like leading the charge when it comes to AI. um It's been my... A flag I've been waving here on the show, it's kind of like, hey marketers, like you get to set the tone for the whole organization because it's coming for the whole organization. Everybody's thinking about it. Marketing is the first department to really like champion this thing. um But of course, the big companies are slow because they have a lot to lose and it's risky. So the solopreneurs or the small businesses like yourself, agencies, solopreneurs, small shops, that small tech startups are the ones leaning in the most that are kind of like setting the first ah trends of what is possible with AI and marketing. um and I'm hoping that we you know get to pioneer the examples and set the set the trends for marketers for the rest of the organizations. I think it's going to carry carry all the way through. It's a very exciting time.
00:35:29
Speaker
Yeah, I think it's a very exciting time. I think that it's probably one of the most exciting times to be in marketing. You know, I think if you think back to like 2010, where we were just dreaming of things like drag and drop on WordPress, whereas now we can build websites, we can do our own design without having to figure out InDesign, Adobe's packages. um You know, we've got Canva, we've got WordPress,
00:36:01
Speaker
um You've got funnel builders and now you've got high level. I think the magic comes when you can take what we know in our minds and actually put that down into a plug and play type of product so that clients can actually get it without having to do the heavy lifting. Yeah. i remember I remember those days and word when WordPress finally made it possible to like just change just change the menu. like That was the big break in WordPress when you didn't have to hard code the menu and then you could put the blog back on the page five instead of on the homepage and it became a CMS rather than just a ah ah blogging tool. hundred percent
00:36:42
Speaker
and you you like The first time you saw a drag and drop builder for marketing automation, like like I almost cried because we were hand scheduling them through forums. There's all these steps, but now it's gotten easier. right like I remember crying like near crying like when I saw that you could do auto fill in Photoshop. right Because you used to find the perfect stock photo. You're like, I finally found it, but it's not quite oriented, right? Or you're like, I need the right side to fill in. And now you can just be like, yeah, just take take the subject, put it here, but fill in the rest. ah Exactly.
Future of AI in Clip Selection
00:37:11
Speaker
it's For those who know how to like kind of operate all these things, they're your ability to go through and do much higher level higher level work in every area that you know a little bit about goes way farther. AI just becomes an extension to 10x.
00:37:24
Speaker
your capabilities so the more capable you become it only gets extended farther i find with a i so if you're a pro all your ability to fine tune it even more is is through the roof. Yeah, exactly. Exactly. Like, can Canva for me is just like magic. And there's a guy on Instagram, I just can't think of his handle now. And even what he does with Canva, I'm like, wow. And that's because he understands Adobe. And so he just creates absolute magic. So 100% agree with you there. The more you know, the more you can actually do. Yep.
00:38:00
Speaker
It's not a replacer, it's more of a multiplier. Though I'm not saying there won't be job loss, there probably will be in some way or form, how much we have yet to see. And it might I think it's still a little bit of a ways out. um But one one last question for you, what future developments with AI are you really looking forward to implementing into your process, like in the near-ish future, like six to eight months that you're hoping it comes down? six to eight months. So we when we started off, and I think we spoke about this, I can't remember if it was off off camera or on, when we first started it would take you know two weeks to turn around a podcast episode. And we managed to get take it from 16 hours per episode down to about six or seven hours. Wow.
00:38:44
Speaker
um with automation and where we're at now with what I can see coming through the pipe with the new um AI products and capabilities, we will be able to automate start to finish one part of our episode down to 60 minutes. And that will be literally all automated. There will be no human intervention. We're not quite there yet, but I think we're very close. yeah What are some of the breakthroughs you're looking for to make that happen? like What are the biggest bottlenecks that you're hoping AI addresses? The biggest one is the one that you spoke about, which is that AI is good at it'll it'll find the clips, but it doesn't complete the clips. You still need a human to find the start and end point on that. And sometimes it doesn't quite make sense the way it has clipped it. So as that capability is growing, for me,
00:39:42
Speaker
that is the the last piece of the puzzle, which is if it can pick reliably pick the AI clip, then you can automate anything. Yeah. I don't know why, but that seems to be like one of the most annoying tasks of all time, is going through a wall of clips and just adjusting them all.
Conclusion and Contact Information
00:39:59
Speaker
I don't know why I hate doing it, but it's one of those things you have to do. It is one of those things I procrastinate on the most. but Yeah, yeah. Well, Leanne, thank you so much for joining me on the show. Where can people go to learn more about your podcast in a box and from you about this whole topic? Oh, absolutely. So I'm very active on Instagram.
00:40:21
Speaker
Leanne, podcast producer, and I'm very active on LinkedIn, so Leanne Scott. And if you want to go and have a look at the website, it's www.podcast-in hyphen-a-box.com. Podcast in a box. Thank you so much for joining me today. Thank you so much for having me. It's been wonderful.