Name: It’s Alive! How to Infuse Your Brand Into AI-Generated Content
Uploaded: 2024-10-23T21:15:37.743Z
Duration: 1 h 32 s
Description: It’s Alive! How to Infuse Your Brand Into AI-Generated Content

Transcript for "It’s Alive! How to Infuse Your Brand Into AI-Generated Content": Everyone, it's so good to see you again. I missed you last session. I hope that you had a wonderful time with Kelly and Rachel, and tell me if your mind was blown about AI search because I know mine was the first time I saw that feature. I'm very spoiled with the fact that I've had access to it for a while because I work at Goldcast. But, for those, again, who want to see more about AI search, that button up top is your gateway into seeing how powerful that new feature is. And, also, of course, if any of our other AI tools have kind of caught your fancy, that AI tools button at the top also will direct you to some more information about that. So real quick, if there's anyone just now joining us for this last session, I can't believe we're already in the last session of the content resurrection. It flew by y'all. Do keep sharing your comments and engaging in the chat with us. We love to see, what you think of the content, and we love to kind of interact with you throughout the session, so keep that up. Y'all have been great today. Drop your questions in the q and a tab. We do actually have reserved time in this session for q and a because I know y'all have been on fire with the q and a. So this time, please do take some time upvoting the questions so we know which ones we should answer first in priority. The on demand recording will be available after the event, but you want to stick around this whole session so you have that chance to win that $500 gift package. Remember, there's docs in the tab, and they're actually different for each session. So don't think because you've checked it once that you already have the goodies. Nope. There are fresh goodies waiting for you over that docs tab on the right. And, again, if you wanna learn more about the AI search functionality or brand voice, which we're gonna dive more into this session, hit that AI search button at top. So as you probably likely know already, we have a specific giveaway item for this session. So today, 5 lucky audience members will win a smartphone microphone with tripod. Super helpful if you're wanting to do more of the, you know, videos on LinkedIn that are everywhere now, and LinkedIn's given priority. If you wanna join that train, definitely wanna win this. So how you win it? Engage in the chat, ask a question in q and a, and vote in our polls. And final reminder, we have the big grand prize we're giving away. So stick around to the end to be sure that you are entered into this. We're so happy you've spent time with us today, and we hope you get great, great content out of this final session with Ashley and Scott and I. So without further ado, thank you thank you thank you for joining us for the content resurrection. This is the 3rd session of 3. It's a live how to infuse your brand into AI generated content. So excited to share this content with you today. Such a hot topic, with AI being on the rise everywhere. So I'm gonna introduce our first guest to the stage. So I would love Ashley to come to the stage, join me, share a little bit about who you are, what you do, where you work. Hey, Ashley. Hi, Lindsay. Thanks for having me. Yes. Of course. So tell us a little bit about who you are and what you do. Sure. My name is Ashley Gross. I am the founder of AI Workforce Alliance. Really, the only thing you need to know about me is that prior to 2020, I had never had any technical experience with coding or any other fun little STEM experiences in general, I would say. I just didn't think I was cut out for it. But in 2020, I was heck bent, I should say, on condensing my work week from 40 hours to 10, and I started dabbling in generative AI tools to do just that. And from there, that really leapfrogged my career into rolling out AI to an enterprise marketing org of over a 100 marketers prior to GPT's release. What was really, really cool about that experience and, in my mind, just made me change Pathway as completely was 3 months into implementing those AI use cases, we were actually able to overachieve our annual pipeline target of 90,000,000 and hit a 115,000,000. And so, once I saw that ROI, I was like, this is what I'm doing for the indefinite future. I'm gonna go help other businesses solve problems that exist in every single organization using this technology that was literally created to help solve for that. So, that's a little bit about me. Well, you are definitely the right expert to have on the stage. So glad we could snag you for this session, Ashley. Thanks so much for being here. And we have one other guest I wanna welcome to the stage. So, Scott, if you wanna come join us and tell us a little bit about who you are and what you do, where you work. Yeah. Hey, everyone. I wish I had as cool photos as Lindsey. Let me just say that because those are epic. My name is Scott Towsley. I am the head of content at Clay. Clay is a creative go to market tool for growth, marketing, ops, more technical go to market functions typically. And prior to running content at Clay, my role is actually in growth marketing, doing a lot of experimentation, some growth product work for about the last 10 years. So I actually found out about Clay, which is, does have AI features, within the product, when I was running a growth team at a climate startup. And this was just fundamentally floored with not just not just Clay, but with GPT and with all these products that were coming out because 10 years ago, I started in content marketing that evolved into, like, growth marketing work. And what was content marketing in 2015, or 2014 relative to 2024 is the it's not you can't even call it the same thing. It's so different and it's so fascinating and it's so interesting because it's not just written content. It's audio. It's video. It's repurposing. It's actually a quite a technical function now, and it's very, mind blown. So I'm excited excited to be here. Yeah. I resonate with that intro because I was trying to think in my head like, where was I in 2015? I was, I think, running the content function of the largest credit union in Alabama, but you're so right. It was so just like blog, blog, blog, blog, blog. And I'm trying to think, like, did we even do any video? This is gonna age me a little bit. I mean, I guess, you know, like, what year it was. You can probably put it together. But the only video I can really remember us doing heavily is Facebook lives, and that was about it. So if anyone remembers when that was a hot, commodity in marketing Oh, that's hot. Anyways, not here to talk about Facebook live today. We are here to talk about AI, aligning AI with your your brand voice. Super excited to get started with that. Scott, I think you alluded a little bit to this, but I want to dive in a little deeper here. To start us off, can you share more about how you're using AI in your marketing workflows today? As a heads up, y'all, your first poll is coming in. Please answer that poll as Scott tell and how he uses AI in his workflows. Yeah. So we use it for a handful of different things. It's almost with nearly everything we do, to be honest. I mean, I guess I could start with tools specifically. We use, of course, GPT all the time. It's either, like, GPT or Anthropic. I feel like it's where people are typically at. We use GPT a lot. But if I can categorically put it into almost there's the net new original content, and that becomes really hard. You almost we tend to treat GPT as our friendly assistant to come up with ideas and bounce ideas back and forth, but we don't depend on GPD, Anthropic, or others to create the net new original content. We do, however, feed it to repurpose content more at scale, and that would be for the written content. But for, audio visual, we use Visual Electric. So any image that's on clay.com/blog, that's generated using, Visual Electric, and we have a custom style prompt for that. We started feeding some of those prompts into Runway to see if we can create videos from there. We've done some fun things with 11 Labs and YC startups that are built off 11 Labs, infrastructure. So it it really spans everything. We try and incorporate well, we don't try. We do incorporate it into almost everything we do whether but it's really on the broad spectrum of it's assisting us as we create net new content or it's helping us repurpose content really quickly and more efficiently. Yeah. Yeah. I think that is, a great bookend to the first session we had, right, talking about distribution and repurposing. And I love that you brought up visual AI tools. I think a lot of people are still sleeping on visual AI tools. And I've just started using the visual electric tool and testing and playing, and it's it's really crazy what you can get. You still need refinement. You still need a human mentor for it, but I love that you focus mostly on kind of the visual elements of that. And so, Ashley, what AI tools are in your marketing workflows? So many. But I would say as far as how I incorporate AI into my workflow, I look at automating a lot of the things that I need information on. Purely from a data standpoint, I like automating my audience segmentation, predicting how campaigns are gonna perform before they even launch. And if we wanna talk about, you know, specific tools and tracking behavioral patterns across multiple platforms, so I can better understand the customer's journey that I'm taking them on. You know, I'll just spitball a couple of my first tools. There's RB2B, which is the automatic segmentation, and then that will go to Clay, which is the enrichment, scraping data from multiple sources to create a more complete view of my audience. You know, inter sorry. Excuse me. Intercom with an automation to HubSpot so that, again, like, as my customers in real time are asking me questions, they're getting that real time interaction with me, but it's also benefiting me because it's sending that information back to my CRM. And then I would just say, like, 2 other tools that are in my workflows right now would be Zeek and AI, just for tracking behavioral patterns across multiple platforms, specifically video, and then Market News to evaluate my content for just SEO and relevance. Oh, I love it. Such a robust AI toolkit from both of you. And I wanna share this poll really quick just for context for both of you kind of where our audience is at. So we asked them where they would like to use AI more. So it looks like right now, the winner is editing video. Okay. Well, I'm glad y'all are here because if you haven't heard, GoldCast ContentLab is a great video editing tool, but I'll leave it at that because we we have we'll dive more into that here in a little bit. And then let's see. What's popping into second then? Oh, it's actually really close with automating workflows. Okay. And it looks like 3rd is creating video. So I think, Scott, you kind of gave the creating video aspect here. Scott, Ashley, any thoughts about how this is panning out right here? This is awesome. This, like, live poll. This is so cool. It's such a cool way to moderate the event. Yeah. I think, like, the editing video is really interesting because I know Adobe is coming out with or came out with basically a lot of these features that, with, Veed. If anyone's used Veed, I thought Veed is really fascinating product where you can kinda type a prompt and it will basically create an ad. I've used that a lot before. A workflow has been as we're thinking about product launches at Clay, this is something I was doing recently. I go and I go for a walk, and I open GPT 4 on my phone and I just start rambling about what this thing is. And I say, write a script. Give me a 2 minute script on this. I'll take that. I'll put it into 11 Labs and it will and then I can listen back to the script, and then I can edit it from there. Once it sounds good and the narrative sounds good, what I would love is to be able to plop that into a video tool like Veed. I think Veed can do it, where it pumps out, like, almost a wireframe as if you're wireframing a landing page, but it can do that for video. I haven't found anything that I feel like does a good job at just the wireframing, but then being able to take it just like if you're in Figma and you have a FigJam wireframe or you have it mocked up, you can then work with the designer to take it to the finish line and it looks great. I haven't figured out the right workflow there yet, but I very much agree with the audience that that is a huge one just because video is very differentiated and there's a higher barrier there. Therefore, you can stand out amongst the crowd a little bit more. So if anyone does have ideas, on how to do this, please let drop it in the chat too. But yeah. No. It's like a huge, thing that we've been talking about a lot on the team. Well, I do have a secret I can share with you, audience, that if anyone saw the promo video that we use for the content resurrection, it was kind of a man typing at a computer and then some files spookily on the screens and, that kind of vibe, I actually used AI to make that. So the actual video itself, was crafted from our video designer. But to get it's not quite probably what you're looking for, Scott, as far as, like, the play by play. Like, sometimes you can have, like, your layout, but I went into our custom GPT. I fed it my, registration page for this event, some other details, some email copy, things like that. And then I prompted it. Now turn this into a event promo video that is built the way a scary movie themed, promotional, you know, trailer. That's what a movie trailer would be. So that's kind of how I packed it together, and it actually gave me I ran it, I wanna say, 2 or 3 times, and then I kind of did the, stream boarding from there of pulling my favorite segments and and sections and plugging them together. So I don't know if that helps. I don't know if that's inspiring for anyone, but Yeah. That's awesome. For you. So, drinking the Kool Aid, I guess. So we talked about tools, but I wanna ask you both, like, what do you love about the AI tools you use, especially if you're thinking about, like, making your content more authentic, making things sound more like you, aligning with that kind of branded side of it. Anything specific that you love in those tools, that helps you get a little bit more back to, like, who you are and what you sound like. Ashley, what are your thoughts there? Yeah. I would say my biggest call out is, like, the adaptability. Right? Like, it has to be able to adapt in real time, and it also needs to be able to have contextual understanding because even though I'm one person and I have one unique brand voice, it shows up differently depending on what channel I'm using or what formatting I'm using. So having something that not only understands my brand voice and the contextual understanding of when I'm in this situation or on this application, I'm gonna talk like this and avoid this, But also having insight into, you know, being able to look at historical data for how something performed. And to both of your points earlier, maybe even breaking down and doing a retro on how it was performed and, like, how the talk track actually ran so that we can continue to adapt our brand voice because nothing's perfect. Everything needs to be iterated and it can always do better and be improved. Great points there. Scott, anything to add on that? Anything specific you're looking for as you're starting to implement and adopt AI tools in your tool deck? Yeah. What came to mind when you asked that is the ability to scale our brand and creative team. And what's so great about I'm just gonna cherry pick a tool and say, Visual Electric. With that product, you can set a style, like a style prompt. And so the style prompt actually gets set by the brand team. And they go in and they make it and refine it and really fine tune it until it feels good consistently across the board. Typically, prior to these types of tools, it took a long time to have the connectivity between what someone who's not as creative like, the output isn't creative. Me. I'm like an aspiring creative person trapped in an analytical brain. And I would love to do what they like, the there's somebody Puneet Atklay and Sarah Atklay, and they're so amazing. What they're what they can come up with is so fascinating. And these AI tools almost allow us to put their brain into a machine and then create our own versions, which outputs what their brain is. And so it's I think that is fascinating, and it's this way that we can scale the creative superpowers of people. And I feel like that's where design and video and, brand can really go is it it unlocks those that don't have that creative capacity but would like to, to be able to move faster. And I think that's a really fascinating, unlock with a lot of these tools. And I wanna point out something that you say that I don't think is said much in just, like, the general domain, but it highlights the superpowers of your human teammates. And so it's not competing. It's not replacing. It's not overtaking. It's highlighting and and pulling more of that out and making it accessible across other people in your team. So I think that is a really fabulous way to think about it. I wish it was a perspective more people would pick up. So thank you for saying that. Really great way to package that up. And so now I wanna ask you all, so when you're using these tools, how can you ensure that the AI generated content aligns with your brand voice and tone? Because we've all been in those situations where you either are generating yourself and you're like, this sounds nothing like me, or probably worse, you're online and someone is sharing something and you're like, this is complete AI garbage gobbledygook, like, stop it and end the madness and the nightmare. So, Ashley, any tips on how to be able to, better train and use AI to align so it's not just like, right? Yes. I'll zoom out before I zoom in. I think that the understanding that a lot of folks need to unlearn and reach themselves is don't generate content with AI. Use AI to enhance your own thoughts. Look at it as almost let's, like, rebrand it into a content reformatter instead of saying, like, generating, you know, content, a content reformatter because that's really what you wanted to do. And so, how I achieved this, well, I think, is I use templates. So I have templates for how I write my blogs, how I write my articles, how I write my PR pitches. And then I have research that I do by hand. And then I'll, you know, feed it to my computer or voice memos, just transcripts of some kind, I guess, is what I'm trying to say, where I'll have my voice memo just, like, listening to my monologue of what I'm thinking about and a talk track I have or something. And taking all those compilings, right, taking the research, the voice memos, the templates, the transcripts, and then reformatting them into what I specifically want. Right? Because it's the output is in the templateized version of what I want, what my audience likes to see. And it's me talking and my opinions and thoughts, but I think that we're all naive and we think our first ideas are always our best ideas. And I know for me especially, I need to talk it out and go through 4 or 5 ideas and conceptualizations before I hit a golden nugget. So, that's what I that's my process. That's what I found works for me as far as a workflow perspective goes. I know for you all watching, Kelly was on session 2 just now. She's the exact same way. She loves to talk to AI and have it learn her thinking and voice and tone and mannerisms through literally just verbal speaking into the tool. So, again, using AI as your friend, is that a little too creepy in that that way? I don't know, but I said it. So, Scott, any tips or things you found that work well with being able to better use AI to align to either your own voice and tone or Clay's voice and tone? Yeah. I love talking to AI because I don't have to talk to myself so much. And and I'm actually not even kidding. I feel like I do this a lot where, I will just open up, again, I use OpenAI, but I'll open it up and just talk and have it summarize it and repeat it back. And I feel like that's really, really helpful just to to better articulate the ideas. Just like sitting down and ripping through a first draft of something of what's on your mind, it's gonna be a complete mess. It's not always gonna make sense. And it's the same thing when you're just trying to talk through an idea. But if I'm gonna sit down with, a friend or a coworker or, god forbid, Angela, my partner who would never wanna hear me talk about any of this stuff, It's it's kind of a waste of the other person's time and because it's not refined thought. And with the AI tools, it really helps you find your right tone. And so I I almost bucket it as the 2 big pieces, which is get your message across that you want and having this AI buddy. Is that creepy? I don't know. But having this AI buddy, you know, by your side, I think is really, really helpful. Once you nail that and you feel like you have a good tone, then you can feed the examples and have a custom GPT and all of those great things or prompts with Claude. I think there's a a massive utility there as well, especially if you have a consistent format. Like Kieran Flanagan from HubSpot, he talks about this a lot on having, like, he built, like, some apps within Claude, and, he has, like, really fascinating use cases on repurposing and posting his LinkedIn content. But, yeah, I I feel like it that's where I've found the most value is on the ideation to get to the crisp point of view, which is sort of what you're saying, Ashley. And then the second is, just getting the repurposed content consistent so it sounds like your your voice and tone. But, again, I don't think we've cracked the second one. I mean, I think all of us marketers need more buddies. I mean, I won't talk for everyone watching this video right now, but I think we could all use another buddy. And I will also say, if anyone is interested in learning more about how Kieran from HubSpot uses AI, we actually featured him at our AI Marketing Alliance Summit, earlier in this year. And so if you head over to goal cast dot o at our on demand events, we actually have a whole session with Kieran talking about how he uses those AI tools, how he makes those apps, things like that. So a great plug there, Scott. Thanks for setting that one up. But Oh, yeah. Of course. Next question is, what do brand needs to have on hand before using AI? I think we all are well aware that the adoption of AI tools is going to continue to rise both from a personal standpoint as an organizational standpoint. So, Ash, I'd love for you to answer this one first of what do organizations need to be thinking about, especially when it comes to aligning AI generated content or AI formatted or templated or whatever terminology you wanna use. How do we set ourselves up for success? You know, are there style guides, voice profiles, video best practices? Like, what should we be thinking about and having prepared as we go into the next year? Absolutely. All the things that you just said, but I would also say from a foundational element, and this might be a hot take. I don't know. Let me know in the comments. Accurate data, because data is supposed to be the source of everything that's true. It's what you build your customer personas, the mapping, the segmentation, all the automations. So that's not accurate. And if the folks in your CRM are not appropriately, enriched and they're not even high intent leads to begin with, then nothing else that you're trying to add on top of that is going to work. And you actually have no real way of measuring how successful your brand voice is. And so that's actually what I would say you need to start with, is going into your CRM and making sure that the data in there is accurate, indeed, high intent leads that you're trying to target and appropriately enriched so that you can test your brand voice. Because when you're building a brand voice aside from AI tools, my process at least or whatever I have done in the past that has worked has been, you know, to to look at a company and look at the executives and take, you know, the brand messaging and apply it to each executive that makes sense as far as pillars go, and then work your way down from that and build out the brand voice so that everything bubbles up to one brand voice, but everybody has got a different consistency and a different style and conversational flow to how they approach it. Scott, what you got? Yeah. I've I feel like thinking about this is, you have to have your own point of view on the world. Like, AI is not gonna tell you what your point of view is. And it's literally like, LLMs are quite literally guessing the next word. That is the most logical to put. And, of course, the outputs are phenomenal. Like, they're they've gone so much better. They'll continue to get a lot better. But treating it as, more of to create the original content as brainstorm buddy, I feel like is is really important. But once you have it and once you nail it, then it's really just feeding examples. And even, Ashley, you mentioned Intercom earlier. We're using Intercom Fin. And so it's their AI product where you feed it educational documentation because that's one thing we're thinking about a lot is how do we improve activation and retention and giving very specific use cases. And so we want to basically have these really tight feedback loops where we learn what are the questions being asked in support, where is our educational documentation, and what are the coverage gaps right there. And then that gets prioritized because it's all messy, unstructured data, and that gets prioritized and says, go create content for this because these are the most frequent credit or questions we get. For example, at Clay, it's a credit based pricing model. And so people will come in, haven't used the product yet and are just poking around and experimenting, and they may have clicked the wrong button and they burn through some credits. They're like, oh, whoops. I didn't wanna do that. And so how can we have of course, like, always improving the product experience. But if that is the most common question that's asked in support, we recognize that, and then we create educational material that's good. And what we're always measuring is, like, support deflection rate to understand if they are rudimentary informational queries, we can answer them, right, automatically, versus if it's something that takes a little bit more like, if it's pulling from data from the Snowflake warehouse, those are totally different questions, and we would have to handle those differently. So I it just depends, right, like, what part of the flow you're leveraging AI for that's significantly a little different in content marketing and more on, like, the product side of things. But, yeah, just a different territory of another place that we're thinking about how do we leverage the AI content. Yep. And coming back to that day to day data. Right? Just all about data at the end of the day, so keep it clean, y'all. So the next question I wanna go into before we get into some little fun video surprises for both of you in the audience, is what steps can you take to train AI to better align to your brand voice? I think we've all been in situations where we're trying to use AI, and it just doesn't sound the way we want it to. So any tips, tricks, pointers, advice about how we can all better be best buddies with AI, because that seems to be the theme today, to train it to sound more like our brand or our leadership or ourselves. Ashley, you wanna take this one first? Sure. I would say the best thing that you can do is give it examples, feed it examples of what to say and what not to say so that it picks up on things. My BrandVoice document, just for context, is 40 pages long. It has, like, so many examples of different, you know, types of formats and different examples of, like, good and bad things to say. But, also, to Scott's point earlier, I have intercom on my website. Right? So every single 30 day period that goes by, there's an automation where it's ingested, you know, through a make automation, and my brand voice is updated. And the questions that are being asked get reprioritized based on the frequency and volume they're being asked. Because the worst thing that you could possibly do is not match your internal brand voice with your external brand voice. So, HubSpot's great. It's got so many properties that I don't think people take enough advantage of, so I'll just give them a quick shout out because that's what my CRM is. I like to have the integration set up so that whether you're on intercom asking questions about a product or service that I offer or you're on my LinkedIn DMs, wherever you are, those questions get to the way that I'm conversationally handling them, the flows that I'm putting people on. All of that gets analyzed with AI because I don't have time to do that manually. And then my brand voice is iterated to keep up with that so that it is, again, like, just constantly being validated that it is a uniform brand experience. And, As I'm filling in those knowledge gaps and those pieces that are missing, then I should see different patterns, which means my brand voice can evolve and become elevated over time. Ashley, I brought up your slide about your build your brand voice with AI. Can you talk a little bit about what this, document is that people can access? And by the way, this is accessible in that docs tab on the right. So please open that up and access that really quick. But, Ashley, tell a little bit about this, brand voice builder, piece you're giving, and then we can kind of transition into that next slide that makes it a little bit more actionable step by step too. Yeah. This document is super simple. It just kind of gives you a step by step way of how to grab all of those data points and how to actually work through a process where you're taking all of this information and then filtering it based on the certain criteria that you actually need, and then being able to look at it holistically and evaluate it and say, for instance, I say and like all the time. So if I'm having AI analyze a bunch of videos, unfortunately, that's going to come up a lot, but that doesn't mean that that's part of my brand voice and I don't want it to actually remain there. So it walks you through the ways to kind of go through this from a workflow perspective. And then, you know, if you're looking at this slide, quite literally, this is what I do. So I've got ReadAI as my notetaker. So it's on every prospect call, but also every internal call that I have with my staff. And I take those meeting notes. I transcribe them either with Read AI or Otter AI, depending on what I'm doing, along with, you know, voice memos and some other just brainstorming applications that I like to use just to talk to myself in other places and avenues. And then I use Jasper because I'm a marketer, and so Jasper's workflows are really great. And I have groups set up in Jasper so that depending on what type of department task there is, It's taking my brand voice and it's applying it to content. It's applying it to product because conversationally and contextually, my brand varies depending on what avenue you're trying to figure out more information about. And then how I iterate it over time is by looking and having automation set up between HubSpot, between Intercom, between ReadAI, between all of my social sites and enrichment tools, actually, so that every 30 days, there's automatically an audit being done on the data in my CRM and on my Brand Voice to make sure that the two pieces are integrating really well together from a brand perspective. Man, I love that. It's a great workflow, great tools. As a reminder, everyone, if you want that slide, the slides are also in the docs tab, so feel free to grab that. And if you're watching this on demand, they're accessible on the page as well. So win win depending on where you're watching this. So, I have a few videos to play, but, Scott, is there anything you wanna add into the conversation about how to best train AI to match your brand voice and tone? I think, Ashley, in summary, give it examples, is it's that's the 80 20 rule. Like, you just feed it as many examples as you possibly can. And how we think about it is you have brand examples, so, like, the Clay brand, LinkedIn, newsletter, blog, things like that, and like the written, and then, of course, the visual, which I covered with Visual Electric. And then you have the person. So that's actually Clay's primary go to market function is through the distribution channel of other accounts and other partners versus our own brand account. And that's where we put a ton of energy and making sure that those that whether you're a creator, you're an agency, you're an employee, you're an influencer, however, whatever category or integration partner we, put that person in, that we can ultimately reduce the friction and make it as easy as possible to talk about the thing that we ideally would like them to talk about. And so by doing that, we've just tried different different tools. The only tactical tool I'll throw out there is, there's one called Yarn, which Tommy from Clay rolled this out. It's fascinating, where we basically had a product launch, and we would have a creator just read a script, and then it would ultimately make the announcement with all the visuals on what this product launch was without them having to do anything. And it basically had a completed product with a written LinkedIn summary, based on their own tone. So we've experimented with different products like that, but, ultimately, give it examples. That's the key. Well, I love that tip because I have a little surprise for both of you and for the audience here as well. So I have 2 videos I've crafted using Goldcast's brand voice tool. So what this is is within ContentLab, which if nobody knows, ContentLab is our AI powered video editing tool in the back end. You can upload videos, pull videos, record, and gold cast into it. You can snip. You can clip. You can repurpose. You can craft copy. But one thing we recently launched is our voice profile tool. And so I had a little bit of fun, and I did exactly what both of you just told me to do, so I'm following directions. I took some LinkedIn posts of yours, and I fed it into that to see how accurate it was, how close we are, how good is it, what are your thoughts. And so I will tell you audience, they have not seen these videos yet, so we're all on this adventure together. So I am first going to share Scott's, voice profile. So let's take a look at this video. So first, I am going to create a voice profile for Scott for his LinkedIn. So you'll see on my screen right now, this is the back end of ContentLab. This is Goldcast AI powered content repurposing tool. We're going to go over to this brand kit tab. You're going to see there's a voice profile, right at the top. We're going to hit that, and you're going to see we have actually a lot of voice profiles in our Goldcast account. So, we've built them out for our leadership, for our marketing team, so there's a ton of different personas and voices in here. So, we're going to hit Add New Tone. You're going to see you can add your brand guidelines, you can add an example copy. So, that's what we're going to do today. We're going to put in example copy. And you have upwards of a 100000 characters you can upload here. There's a lot of space to get your brand voice right by putting in a ton of different examples. You can put in blog posts, LinkedIn posts, copy from your website, any kind of copy that would harness your brand voice and tone the best, you can copy and paste into here. So what I've done is I have created a Google Doc that has a bunch of different, LinkedIn posts from Scott. So we're gonna go down here and copy and paste. Pop back over here, put this in here, and all you need to do is just designate post 1, post 2, post 3, so the AI can understand where the endpoints are for each of those posts. And then we're gonna hit Save. And now we're just gonna have the magic happen. We're gonna hit Next. It's gonna analyze all the copy you just threw in there. It's gonna take about a minute to generate. So, it's going to read through everything you put in there and then it's going to populate your voice profile, with standards of your tone and voice, and then also writing guidelines. So, there we have it. So, Scott, would you say that your tone is enthusiastic, conversational, personal, and informal? We'll have to get your take once this video is done playing. So you'll see up here that it populates, voice and tone here, and you can add things, you can edit things, you can delete things. You can also add more copy to refine this over time as well. And then you'll see too that there are writing rules it'll populate pulled out of the copy you put in here. So, for instance, for Scott, for his LinkedIn, voice profile in ContentLab, it would use short paragraphs, it would ask rhetorical questions, use emojis frequently, use metaphors and analogies, and then it goes on and on. You can edit, you can delete, you can add. You can refine this over time. And then over here, it's going to tell you, or show you, I guess, show you how this is represented in Content. So it's going to have like a sample post. It shows, across any of these generated voice profiles and it's going to use the voice and tone, section, as well as the writing rules to craft this according to how it read the copy you fed it. And so, you can edit these things, edit these things, and then reflect that in the sample to see how it's updated over time. So, Scott, let us know what you think about this voice profile in ContentLab. That was epic. I I love that. That was cool. All about it is is, feeding all the examples too. I was taking notes as you were going. Use metaphors. Yes. Ask rhetorical questions. Yes. Use emojis. Yes. Definitely. Oh, did I skip ahead? Okay. Oh, no. You're good. You're good. I'm keeping up with you. So keep going. I love it. Yeah. The the you use casual language, I would say that's probably right. Yeah. The right the writing rules seem pretty spot on, to be on. I'm I'm trying to, like, be a critic and pull one out where I'm like, no. It's not right. I didn't actually see anything on there. Well, I'm so happy that it aligned with what you were expecting, and you think it sounds like you. So appreciate that feedback. I am so happy to hear that it it got your voice profile. And I will tell you all, you like I said in the video, you have a 100,000 characters you can play within this prompt. I used, I think, like, maybe 2,000. I don't even know. It was, like, 4 LinkedIn posts. So, imagine how much better it could even get with even more content. So love it. Also, while we're watching this next video, everyone take some time to upvote the questions and the question answer you want answered most, because we have one more video to play that's gonna share Ashley's voice profile. But we do have time left for q and a in this session. Hallelujah. So please upvote some of your questions so we can, kinda go in sort order and answer what y'all want answered first. So now, Ashley, are you ready to see your voice profile? I'm ready. Let's go. K. Ashley, now it is your turn. I have taken 5 posts from your LinkedIn and already copied and pasted them into the create a new voice profile box. So we're gonna go ahead and hit save, and we're gonna have the magic happen. So I recommend using at least 5 LinkedIn posts if you're trying to make one that's specifically for your LinkedIn. Of course, the more the merrier. The more you can put into the tool, the better it's gonna be at reading your voice, understanding your style, pulling out those things that make you you. So let's see. It has read Ashley as being confident, conversational, engaging, persuasive, and encouraging. Ashley, would you say that's right? I'm super excited to find out. And then let's see as far as some of the writing rules here. So use short paragraphs for readability, provide actionable advice and takeaways, and post with a call to action or thought provoking statement, avoid industry jargon to make content accessible, very sentence structure for rhythm. And then if you look over here on the right in the try out your voice, this is actually very different compared to what it did for Scott. So you can see that it definitely has adapted to the format and the style that Ashley has in her LinkedIn profiles. So this is a really powerful example to show how the AI reads through what you're feeding it and can adapt this example copy to not only, like, how you write your post, but the style as well and the presentation as well. You can notice that this is a very different style layout compared to what was in Scott's. So, Ashley, what do you think about this? How accurate is it? Wow. I I didn't get to see this beforehand, so I'm I don't know if you wanted to show me this live because I'm kind of speechless, and this is not gonna be very entertaining. But I'm speechless because I I thought that I was doing the most amazing job, so I actually coded my own, like, little agent to take all these things and all these pieces and parts of me on the Internet and analyze them really, really well. So I'm I'm speechless because it works really well, but, also, I'm really angry that I spent 10 hours on a Sunday coding this on my own when this is now available, and easily accessible and can be done in what looked like 10 minutes. So thank you. I love to hear it. Wow. Saving 10 hours too. Who doesn't wanna save 10 hours? So and I will say I was even really surprised as I was going through this exercise because, I mean, I feel like I know this tool very well because I'm in here all the time. I'm building voice profiles for the Goldcast team, all those kind of things. But even as I was going between creating Scott's and creating Ashley's, I was just, like, really fascinated when even looking at that sample copy text because it uses kind of the same messaging across all the voice profiles. There's consistency, but it'll adapt it to your voice profile. And I was even floored by just how it formatted it differently. I was like, oh, woah. This is like real creepy smart. So Verifyingly accurate. I'm glad you both enjoyed that exercise. And for anyone else who's interested in doing that, you can actually go do that right now. ContentLab does have a free account level, so you can go to goldcast.io/contentlab, start your free account, and you can play with voice profile and create your own. And then magically make all this content that sounds more like you and less like AI. But, anyways, I think it's time for some q and a. So I'm gonna hop over into the q and a, and I'm gonna kinda sort by most upvotes and see what we have. Let's see. So, everyone take some time. Go upvote which questions you want answered first. So let's see. I'm gonna go and pull oh, here's one. How do I know AI isn't creating the same content for someone else? This is an interesting question. Ashley or Scott, any response to this one? I would say 100% guaranteeing, I don't think it's fair to ever promise you. There's no way to not fully understand and articulate it. We're just learning these things as they're continuously being iterated. I would say my number one suggestion would be, 1, having a really good understanding of what data is and is not acceptable to put into an LLM according to you and your standards and your company's standards, not just general policies and and practices. And I would also say make sure that you know, you know, what best practices are for anonymizing prompts, making sure that you're toggling off settings in, you know, chat gpt and even LinkedIn, for instance, on how, you know, it uses your data to train its own models. And then I would also say, as far as creating the same content, this is why it's really important and why I kind of did my own little rebrand on the word, you know, instead of saying generating content, let's just start saying content reformatting. Because they should be your thoughts and your perspective. That's what makes it valuable, and that's what makes content rank really high. So use your own thoughts and take the time to do your own research and form your own perspective on things. And then use the AI to reformat that into, a best practice or or the best formatting for what you're trying to achieve. Scott, anything to add on to that? Yeah. The I I mean, how do I know that it's not creating the same content? We don't. Like, there's not really a a great way to figure out if the it definitely is. You kinda feel it. You know, that's not a great answer, but you can oftentimes. There's it's like a spectrum of, you know, this is very clearly if you're on LinkedIn, you're looking at comments, you're like, okay. This is very clearly generated completely with AI. But in terms of how can we actually tell, my follow-up question, I guess, would be what are what would we be trying to figure out? Even if we do find out that it was created, with AI, what's the underlying intent behind it? So that's just where my head goes. And I think I will add our our sense of goal cast is AI always needs a human editor, and that's the way you ensure what you're putting out is not the same as anyone else's. Because once you put in your human edits and your human perspective, that's what's gonna make it your own and gonna make it that valuable, you know, individualistic content versus, like, AI generated. So mic drop on that one. Paul asks, what kind of training tools or courses do you recommend for marketers, agencies, and orgs who are beginning with AI usage? I think it depends on what you're trying to accomplish. I actually try to guide people to think about their business and what problem it is that they need to solve and get alignment on all levels on what that problem is that they need solved. And then form your use cases and vet the AI tools that were built to use to solve those use cases based off of that. And then you can kind of go seek out training and and knowledge. But in my opinion, there's so much on the market, and there's not enough strategy, which is why I use AI so that I can spend more time doing strategy. So taking a little bit more time to benchmark, you know, what is the processes and what are some of the silos that exist in your organization today that need solved for, and then work your way up from there and apply AI. But AI is an amplifier and an enhancer. So if you don't figure out where the crap is and where the good is, don't add AI to it until you internally know that because it's just gonna enhance whatever is good, but also whatever is bad and not working. Mhmm. Yeah. That's that's a good point. I've I feel like identifying where are you spending the most time, and can you use a AI product to alleviate some of that time that you're spending there, if it's possible to do that. Right? Like, if you're doing operations or you're trying to find, data and research something at scale, you could use Clay. If you want to you have a consistent tone and voice, but you wanna be able to repurpose, at scale, on LinkedIn, you can use Goldcast link. If you want to, yeah, if you want to ultimately, like, figure out how to repurpose images, you use Visual Electric. But it just comes down to, like, where are the constraints that you have and where are the tools that you can plug in. And what our honestly, our honest approach has just been, what is it, Edgar Dale's cone of experience theory, where you learn more by just doing it. Going through a course, watching it, going through the full thing, yes. You'll it it it is helpful depending on the quality of the course, but we actually found we learn way faster by just saying I have no idea how to use this thing. I'm gonna go try and figure it out. I'm gonna read documentation, and then we'll just learn and start to make our own conclusions on, how we feel about that product after dedicating. You know, we try and just say we're gonna take 3 hours max to spend time on this product and see if it's gonna help alleviate, like what Ashley was saying, the very specific problem we're hoping to solve right now, which for us, one of those things is video editing. So we have been playing around with other products. As Nike would say, just do it, maybe. I think it's a a good way to synopsize that too, Scott. That's right. I love that too because I don't think a lot of people think that way, so I love that you're bringing a little bit of a fresh perspective into that conversation. And I will add one kind of area where you can learn. I mentioned it earlier about our AI Marketing Alliance, but we actually have our next summit coming up in November. I've dropped the reg link in the chat there. But this is a great virtual summit. I wanna say it's like a half day virtual summit, that we're hosting as part of the AI Marketing Alliance. It does go into a bunch of information about use cases and how to best use AI, and this one's gonna essentially be about ethics, legalities, transparency, usage. So if that sounds like interesting AI training to you, feel free to go register for that event. It's gonna be a great time. But next question that I have to go over is from Julie, and Julie asked, what are your best tips for building a good GPT? I'll answer this one first because we have gone through this exercise ourselves here at GoldCast, but I think, again, it's the more that you can feed it, the better. But make sure what you're feeding is accurate, up to date, timely. You know, don't just feed it the whole historical knowledge of your organization because a lot of it's gonna be outdated. A lot of it's gonna be void, especially if you're in tech or software. Probably a lot of your product's gonna change from, like, you know, blog post 3 years ago or video tutorials from 2 or 3 years ago. So feed as much as you can, but make sure it's relevant, timely, accurate. Doing a call before you just shove things into chat GPT because once it has it, it's not gonna let it go. So just think about that. That's my advice. Scott, any advice that you wanna share? Yeah. That was such a good point. Once once you have it or once it has it, it won't let it go. It is hard to remove information from there, once you're feeding it. I feel like that's very underrated advice is when you are building out the GPT to be, do a lot, do, like, a big QA before you do that because it can, of course, any LLM can hallucinate and come up with BS. But really making sure that what you're putting into the machine is good. And as the saying goes, garbage in, garbage out. Right? So just put in the highest quality that you possibly can when you're building a custom GPT and really spend the time because, you know, you'll get orders of magnitude at scale content out than what you put in, which means the input has to be of the utmost highest quality. Ashley, what are your tips? Yeah. My tips would actually be and and I'm saying this from experience because when the ability to create a GPT came out in, I believe, 2023, if not late 2022, I actually used the 1st GPT and built it to be for our sales organization. So in the enterprise, we had around 200 AEs, and we were onboarding so many that the more senior level AEs were having to train the newbies, which obviously took the, you know, attention off of the high intent leads and converting and sales cycles. And so I matched their workflow. And so, essentially, what I did was I interviewed the AEs, and I said, okay. What do you do first? What's the order? Where are you pulling the information? And they said, you know, competitive intelligence, account planning, and then outreach sequences. This is the way we're doing things. These are the places that we go to get the information during each of these steps. And so when I created the GPT, I used those 3 steps. And so when you go into that GPT, it says, what are you doing first? You should be here. Have you already done this? Well, if you have, you can move on to the next step. If not, we need to start from here. Because the only way, in my opinion, to remain the human in the loop is to understand what's happening when you're not there. And so if you mimic the workflow that a human is doing, then you understand what is working and what is not working, and how that translates digitally. And then my second piece of advice would just be as far as getting mass adoption, assuming if you're gonna go through the effort of building a GBT, you wanna roll it out to more than just yourself, I would say you really need to think about how you can, essentially, foolproof this to the best of your ability. So what is what are all the ways that you can make this process easy so that no matter who's using it and what their skill set with generative AI is, they're gonna get really accurate outputs. And so the way that I do that is creating shortcut prompts. So instead of anyone having to go in there and actually prompt it, they can, you know, type in command forward slash one. And it gives them that prompt and it just makes them fully fill it out before moving on to the next step. Just trying to find any place that could be tampered with and then figuring out what I'm gonna put in place to mitigate that risk because that is a risk. That's a good point too on it's not just what you feed into the GPT, but it's what's the GPTs created, and there's the prompts that are then asked there. Because if you have varying people who are using the GPT, what they ask, like, what their prompt is will vary quite a bit, which is yeah. That's very fascinating. Alright. So, May, I love this question, and I'm so excited to hear Ashley and Scott's opinions here. So, May says not to be a devil's advocate, but what challenges have you guys experienced with AI? Scott, do you wanna tackle this one first? Definitely, if well, for everyone on the call that's tried to create any text on an image, and you get that crazy, it's so frustrating. I'm like, there's literally three letters I want you to do. Just 3 letters, and it's just completely wildly off. It's like a very tactical thing. But as we're using, image creation tools, if there has if we wanna, like, have text on the on there, that is a challenge, and that is a massive one that, I've struggled with, to to figure that out. Oh my gosh. I so agree. I mean, you can I don't understand what happens in the loss in translation between, like, my text prompt and then it putting out the image, and then just, like, somehow it misspells the? And you're like, this this is not rocket science, and so frustrating. And I don't know how to fix it. I don't even know, like, how can I come back and tell it, like, no? That's not right because I already told you right the first time. So, like, oh, don't get me started. Okay. Ashley, what are some challenges that you have hit with AI? I would just say, you know, if I zoom out, honestly, patience has been a really big one for me. Like, just having the patience because we're looking at the worst iterations of generative AI that we're ever gonna see in our lifetime. And especially coming from a marketing background and being a marketer, there's so many tools that have already been created that were supposed to solve problems that have not been solved that still exist. And so I think just having the patience and not getting super excited and gung ho when there is an AI tool that was built to solve that problem, that doesn't necessarily mean it's going to solve the problem the way that you need it to or to the extent that you need it to. So, really, my biggest frustration has been just having, a certain level of criteria that needs to to hit and to make sense for me to onboard myself to this because it's just another tool. And if it takes me 30 minutes just to play around and figure out what the basic features are, that might be an hour or an hour and a half for somebody that I have to train. And so, really, just just being able to capture the best ROI and seeing the features in the best light while this is all running in data and having the patience to do so. That's my wholehearted answer. Well, I love that. We have just a few minutes left. We did have one last question I want to kind of close out with. So we had a question come in asking about for those who haven't really started but want to get started and stick their toes in, you know, where should they start with AI? How should they start with AI? I think the, you know, just do it was great advice, but any other kind of last words around, you know, how how should we get started in this big old world of AI if we still haven't yet? I'll just give you my secret sauce. This is it, ungated. So, anytime, and this is what works whenever I did this and I had no context if it would work and it has worked across 11 mid market and enterprise companies. So, I can successfully say it still works. But I would say, first and foremost, when you're building out, you know, the initial pilot project use cases, make sure that they're going to solve a business functionality or they're going to be tied to the business so that it's benefiting everybody and not just one off pilot projects. They're all connected. Everyone understands their individual contribution and how that impacts the whole organization. I would also say, though, have anywhere from 7 to 10 initial AI use cases, and then go back to your teams and actually talk, not as an avatar or, you know, anything else or a clone. Go back and say, hey. Here's all of the use cases that we can experiment with. Which ones do you hate doing? Because people's sense of job security is not going to be threatened nearly as bad if you're giving them the option to choose a task that they hate doing and then showing them and experimenting how to automate it in real time. They're not gonna hate you for doing that. That's how you buy them in. And then I would say go one step further. And if you're serious about AI implementation in your organization, add some resources, add a learning and development budget, a Slack channel, add it as a core competency if people wanna get a bonus that year so that everybody understands that this is a top down strategic approach to AI implementation. I love it. I mean, I think that was the the best way to put it. Right, Ashley? Well, close it out well, and then, you know, Scott, as you said, and I alluded to, you know, let's just just do it. Right? Let's just do it. Let's tackle it. Let's try. Ashley, Scott, thank you so much. This is a great conversation. You all shared some wonderful, wonderful advice today. So I really appreciate you taking time to talk about resurrecting content using AI, aligning AI for brand voice and tone. Let's have less some AI robots out there and more, you know, talking like us. Right? That that's the dream. And maybe we saw a little glimpse. Right? A little glimpse. Absolutely. Yeah. Thank you, Lindsey. This is awesome. Right. Oh, well, we will see you later, everyone. Thank you so much for watching the content resurrection. We hope you enjoyed all three sessions of the content resurrection. Really great time. We'll see you at the next event. Right? Bye, everyone.