← Marketing Against The Grain
Loading summary
Transcript1 lines
- [00:01]
A
Gemini3 has just blown our minds. But hot on the trails is Claude with its new Opus 4.5 model. And it is mind blowing. And I'm going to give you three marketing use cases you can use today to start to get real value from the model. I'm going to show you how you can create an interactive artifact to represent your ideal customer profiles. I'm going to show you how you can build an incredible interactive dashboard to analyze all of your marketing campaigns. And then I'm going to show you how Claude Opus 4.5 can build you a content hub to remix any content you want in multiple different formats. That's three use cases with Claude Opus 4.5 that you can start using today. All of that and more on this episode of Marketing against the Grain. Okay. Over the past number of weeks, my AI behavior has changed dramatically. Really? Ever since OpenAI launched ChatGPT 5, it was a little underwhelming. I still use ChatGPT for the most part because of its memory, had all of my memory in there, so the results were tailored to me. But. But then Gemini 3 comes out and it is just such a seismic improvement on what ChatGPT5 is that I've started to use Gemini 3. Not just that Gemini 3 is a better model. Their video model, Veo 3.1 is a better model than Sora. And then their image app, Nano Banana Pro is amazing and is way better than ChatGPT image gen. So they went across the board. But then Claude, which is really trying to play in a different space, it really wants to kind of be the backend for all agents that enterprise companies build, released their Opus 4.5 model this week. And so I got to play around with it, I got to use it for some marketing use cases. And I want to take you through those today to show you how powerful of a model it is, especially when you use it to build interactive artifacts that you can share with your marketing team. So let's go through three use cases. I've adapted all of these from this incredible use case library that Claude put together. So one of the ways that you really drive adoption of AI within your companies is to give people inspiration, which is just easy copy and paste prompts that they can go and start to use to see how they can use it in their day to day workflows. Now, luckily, the large language models agree and they're starting to put out these use case libraries where we can actually go and run powerful prompts to see how this could work in our day to day work. Now I've adapted three use cases. The first one is how you build customer Personas. Now, customer Personas are fictional representations of your customers. They allow you to do much more tailored marketing. It's a great foundational piece of content to make sure all of the marketing you do and even the selling you do is tailored to who your audience is. One of the best use cases of AI is to build those customer Personas for you using unstructured data. What do we mean by unstructured data? Things like how your customers have talked to you on sales calls, maybe in your chat logs, maybe based upon survey responses and any external information AI can get from review sites or forums where your customers hang out, really trying to get the language they use, and then build these ideal customer Personas for you. Now, what we're going to do is show you how easy this is so we can try Claude. We can get in and we kind of use, for the most part, the verbatim prompted to give. I changed a few things. One of the things I'll say is, at the start, it's telling you, hey, build this customer Persona based upon customer feedback from things like sales calls, support conversations, survey responses, that kind of internal knowledge. What I would say is, if you want, I would also pair that with external review sites and forms. For this example, I'm not doing that because I want it to be really simplistic, but I would just recommend you do that. If you follow along here and if you want an easy place to go copy and paste the prompts, I put them on my sub stack. I know a lot of you all reach out to me and say you watch the YouTube show and follow along on the sub stack to get the prompts. And. And so my substack that you can go to, which is marketing AI action, you can just go copy and paste them, but they're easy enough to get through Claude as well. And so this here is really using internal information. Now, you may say, I don't really have internal data at the moment. And so I don't know if I can, like try this out again. If you just want to say, hey, use external information you find on review sites and forms, that's a pretty good way to start. You don't need the internal information. If you want to see what it could look like with this internal content. Or what you can do is basically give Claude this prompt and say, I want to run this prompt. I want to see what it looks like. Can you give me synthetic data to basically create sales calls, notes, support conversations, survey responses, so you can see how it looked. Now that's what I've done because I don't want to expose internal data from HubSpot. So all of the things you're going to see from me are synthetic data. So we run this and you see here the output make the artifact professionally and elegantly designed as it will be shared with others. That's the main thing to take away from each of these prompts is how easy you can share these artifacts with your marketing team. And that turns cloud into a real valuable asset for you as you do all of this marketing. So here we go. It runs it on the synthetic data and this is really cool. So it gives me HubSpot CRM customer Personas. I had synthetic data around HubSpot's CRM, the people who would use our CRM platform. And we had the Rapid Scaler, we had the Pragmatic Operator, we had the Enterprise Architect and the Technical Maximizer. And then it puts it into this really cool interactive site. So you can basically go to the Rapid Scaler. It tells you who they are, tells you the typical company size, what their budget is, who the decision maker is, which is really useful, what industries. And then look at this, it has first of all a quick overview of their goals, their pain points, and it actually takes out some of the core language that they use verbatim. Then it takes out this journey. I'm going to run this prompt on real data, by the way, because it's so good. The Customer Journey. If you're listening along on the rss, what I'm showing here is a interactive app that Claude has built within Artifacts that shows you a kind of one pager for the Persona and then shows you the Customer Journey stages. What do they go through? They go through research and it basically talks about what they're going through in that stage. Then they go into evaluation. It talks about the core thing. They're trying to figure out, can we afford this as we scale. Then it goes into onboarding. Need to be live immediately. This is like pretty accurate for the Rapid Scaler, who's an early stage company. Then the daily use is the team actually using this. And then growth costs are climbing faster than expected. And so then we can get into like real quotes. It's taken out from the synthetic data. And so you have this really nice Persona that you can actually share with your team. And look at this. I can click through the different ones. I can get into the Pragmatic Operator again, gives me all of the same information, gives me a cool different user journey that they go through, and they think about different things. And the technical Maximizer, they're in 20 to 200 employees, they have a different budget, they have a different set of needs, they have a different journey that they go through. You know, they're thinking about what's actually possible with APIs and automation. So pretty great. And then the way you can use this is anytime you're doing marketing, you can basically say, claude, I'm going to do an email campaign for the rapid Scaler, or I'm going to do a webpage for the Enterprise Architect. Can you create me a one pager that I can use for my AI assistant to tailor that for that Persona so you can keep this artifact? And anytime you're using AI to market to these different Personas, you can basically ask Claude to give you a description of that Persona. And then you can just upload that to Claude and say, tailor everything to that Persona. This is a great first use case. And the other thing which is so good about it, as you can see, this is a shareable link. So I can just share this with my marketing team, I can share this with my sales team, I can share this with my CEO, and they have a really cool interactive artifact to say, wow, like, this really brings to life our core customers. And again, we, we can get much more sophisticated with the prompt because we compare the internal data, which is all of the things we're doing here, sales calls, support conversations, survey responses, with external data, which would be forms and user review sites. Use case number one. You should definitely do this. Every company listening should be using AI to build a customer Persona and then using that customer Persona to tailor both your marketing and your sales. All right, we put together a list with all our favorite cloud hacks. It includes more than 60 prompts across content creation, strategy, campaign analysis, all the workflows you need. These aren't experimental prompts. These are reliable prompts you can use right now to completely transform your marketing. Scan the QR code or click the link in the description. Now let's get back to the show. Let's get into use case number two. So use case number two is basically, I'm going to analyze campaign performance. Now. There has never been a better time to be a marketeer, because now you have a data analyst who can not only do all of the analysis for you, but can actually build interactive dashboards to represent that data in any way you want, make it shareable so you can share with different team members and then actually provide real recommendations. So in this One here. This is going to analyze campaign performance. So we have a bunch of different marketing campaigns and we really want to see how we're going to prioritize the next quarter of marketing work. And I'm going to give it a bunch of data to show it how these marketing campaigns have been performing. And then I'm going to ask it to basically show me like what are the most successful channels by LTV to cac. I'm going to ask it to recommend how I should distribute my spend and my priorities in Q4 based upon these numbers. Again, I've asked Claude to create synthetic data so you would use your own data from your own systems. But let's get into what it's done. So we run the prompt again, a tiny little tweaks I made, but nothing that you need to make. They're just tiny little tweaks that I prefer. This is again running on synthetic data. And so again it's built a great interactive dashboard because we ask it to do that. So we analyze it and we ask it to like which campaigns and channels are working, where should we reallocate budget for Q4? What patterns am I missing across segments? This is so good context here we give It Our target ROI is enterprise customers have 3 times higher LTV than SMB. Industry benchmark is 200 to 250% ROI. I can shift up to 30% of my budget based on performance. So we give it some real guardrails and then it creates this cool interactive dashboard. So this is my Q3 performance plus my Q4 budget strategy. So the first thing it does is it shows me my total spend, what revenue I got back from that, the average roi. Obviously this would be amazing. This is based on synthetic data. The AI is very optimistic on how well marketing teams are performing. Shows me my conversions, then breaks it out by roi. Like shows me my channels stock ranked by ROI and then has this like really cool trend analysis. Like over time is my ROI and revenue trending in the same direction. So like gives me a nice starting point. Here's how things look. And then I can get into efficiency analysis. And so what it's showing me here is what are the most efficient channels in terms of CAC to ltv. So CAC is what is the cost to acquire the customer and LTV is what is the lifetime value of that customer. So If I spend $3 to acquire a customer, maybe that customer immediately pays me a dollar, but then over the lifetime of the customer, they pay me in total $10. So it's worth spending the $3 to acquire that $10. And so what you want to do is you want to optimize for the best channels in terms of the help you acquire customers that have the best lifetime value. And so it builds this cool little bubble graph to kind of start to show me the distribution of my channels by roi. And then it gives me my top performers email and SMB mid market email. Wow, it really likes email. Again, based upon synthetic data, yours will be like different than this. So basically pretty cool, right? It shows me efficiency analysis every month you'd want to do this. What channels are actually yielding me the best return, gives me the top performers. And then why AI is so incredible, it can start to say, well, what should I actually do? And you can see here, it's like got everything mapped out in terms of how I should reallocate budget in Q4. It basically tells me I should increase an SMB email, exceptional ROI and low CAC relative to ltv. And then it says, what is my current spend, my new budget, and then what is the change? And then basically gives me a little bit of projections. If anyone could get 70 to CAC, that would be pretty amazing. 6900 return on investment. So basically we'll map out the exact changes you want to make and why you would make them, like the roi and then tells you where to maintain. Actually, interestingly enough, doesn't tell me to decrease anything. So that's probably one thing I would edit with it. I would probably say I can shift up to 30% of budget based on performance, but I can't increase overall budget because it's actually increased the overall budget because I'm increasing these and maintaining these, but I haven't reduced anywhere. So again, a little bit of refinements, but does show you how incredible this is. Like every month, every quarter, as long as I have all of my data, I can give it to Claude and I can ask it to build me this kind of interactive dashboard that allows me to see my performance in one place and gives me all of the recommendations I need. Now we saved the best to last. Everyone likes content, so this one has definitely been tweaked from the original. But again, you know, you get in there and you tailor these for your own needs. But what we're doing here is we're taking a blog post and then we're going to turn that blog post into a LinkedIn carousel, a Twitter thread, an email, nurture sequence, podcast, talking points, infographic outlines. So AI as a content remixing tool is really good. What I would stress here is it will give you a first draft and then you have to make that first draft much better so you will not copy and paste. Do not think of it as an automated workflow where I just go, copy, paste, copy, paste. You want to take these as first drafts and add a little bit of magic and then post them. But just look how good this is, right? We ask it to build a little content hub to represent this, to make this really easy. Because if I have someone doing this, I want to copy and paste this to my teams who own all those channels, so they can take that as first drafts and use that content. That's the way I can start to really accelerate my content production. Because I'm taking one thing and created many things. And so it builds me this little content repurpose hub. And so if you want to build on this, what you would ask Claude to do is, hey, build a intake box, or I can give you any URL. And then you will repurpose that into these different content assets and create a tile for each asset. So I could have multiple assets under each tab. At the moment, it's just giving me the one LinkedIn carousel post, the one Twitter thread, the one email series, the one podcast, one infograph. What you want to do is basically input a URL, a piece of content, and then creates a tile with the same name as that piece of content. So if I put in a piece of content to say Kieran is the most amazing AI person on the planet, it will then take that post, create a tile for LinkedIn saying, Kieran is the most amazing AI person on the planet. And then within that tile, you click in and it will have the LinkedIn post for you. Same with Twitter. So you can have multiple pieces of content stored in a backlog. And anytime you see a piece of content, you can just add it and it will repurpose it into these different channels. That is a little thing that I'm doing that you should do. And I'll showcase on a future episode, likely next week, where I have three apps that I've built that are really great internal apps for marketers. And so look what it's done, right? It's created the LinkedIn carousel slides from that one blog post, it's created the Twitter thread from that blog post, it's created my email series. Look how good this is. Look, look right. Look at the podcast now, again, this is a very basic prompt. What you can do is make this much more sophisticated because you can have formats that it adheres to. So you can have a podcast format that you like. And so anytime it repurposes into a podcast format, it will use a different format and style. Guess what? That episode is coming as well. This is really just kind of just the basic and that it gives me a pretty good outline for an infographic. Just look how good this is of one basic prompt and you have this whole interactive content hub built by Claude off a very basic prompt. Imagine how more sophisticated I will and can make this. But let's just show you how good it is. Right? The infographic. Let's take the infographic outline and use one of our favorite new tools, Nano Banana. Now I did use Claude for this. It's not a image model Claude. So it actually just built me kind of a website, right? Nanoban one pager. Let me show you what nanobanana built because it's actually pretty incredible. And so here's the infographic. So look how good this is for first draft, right? And that just came from Claude taking a blog post and creating the outline and then me just going to Nana Banana and putting in the exact outline. Right. What I basically said to Claude is can you turn the AI infographic outline into a prompt so I can have an AI assistant build it and then I just copy and paste that into Nana Banana. And here is the infographic. Holy smokes. It's never been a better time to be a marketeer who loves AI tools. Because Gemini 3, the model VO 3.1, Nano Banana Pro, Opus 4.5 and interactive artifacts is a tool set like no other. Like no other. And so this is an episode to get you started. If you really want to go try add some more use cases, go check out the use case site from Claude. Adapt some of these for your needs. I'm going to come back in a future episode to show you three internal tools for marketeers. That content hub being one of them. But two others that are mind blowing that have been built with Gemini 3 and Opus 4.5. What a time to be alive. I hope this was a great episode and you got a ton of value. And until next time, I'll see you all later. Sam.