The GenAI Genie: Unbottling Versatile AI Use Cases

 
Drew Neisser: Hey, it’s Drew. I’m excited that you’re here to listen to another episode of Renegade Marketers Unite. And if this is your first time listening then welcome. This show is brought to you by CMO Huddles, the only marketing community dedicated to inspiring B2B greatness. And that has a logo featuring penguins. Wait, what? Yeah, well, a group of these curious, adaptable and problem-solving birds is called the Huddle. And the B2B marketers and CMO Huddles are all that and more, huddling together to heat up the coldest job in the C suite. And now that CMO Huddles has three membership tiers, we’re ready to inspire B2B Greatness at all levels. To learn more, check out CMOhuddles.com. Now before we get to the episode, here’s a shout out to the professionals at Share Your Genius. We started working with them over a year ago to make this show even better and have been blown away by their strategic and executional prowess. If you’re thinking about starting a podcast or want to turbocharge your current show, be sure to talk to Rachel Downey at shareyourgenius.com and tell her Drew sent you. Okay, let’s get on with today’s episode.

Narrator: Welcome to Renegade Marketers Unite, possibly the best weekly podcast for CMOs and everyone else looking for innovative ways to transform their brand, drive demand, and just plain cut through. Proving that B2B does not mean boring to business. Here’s your host and Chief Marketing Renegade Drew Neisser.

Drew Neisser: Hello, Renegade Marketers! Welcome to Renegade Marketers Unite the top-rated podcast for B2B CMOs and other marketing-obsessed individuals. Alrighty folks, you’re about to listen to a Bonus Huddle, a specially curated Huddle that we run once a month with experts sharing their insights into the topics that are most important to our CMO community. We call them Huddlers. This time you get an inside seat at a Bonus Huddle where Nicole Leffer, a true generative AI wizard and founder of A. Catalyst, LLC provides us with insights and inspiration on how to get the most out of generative AI tools like ChatGPT and a few of our other favorites. I know you will find amazing information in here and I guarantee you, you’re probably going to want to listen to this episode twice. So let’s get to it. So let’s get into this Bonus Huddle. I’m super excited to have Nicole Leffer back with us. She was here almost exactly six months ago. So much has happened since then. Nicole runs her own generative AI consulting company called A. Catalyst and has already been hired by at least three Huddlers to help enlighten, inspire, and train their marketing team. So Nicole, welcome back.  

Nicole: Thank you so much. My last time here was my favorite thing I’ve recorded, don’t tell anybody else.

Drew: Of course it was, well we have to top this. So I was thinking about our last time together, we really focused on just prompt crafting with ChatGPT and obviously, you keep making that an art form unto itself. But how is your approach to prompt crafting evolved with ChatGPT? And then of course, we’ll move on to other things.

Nicole: A lot of it has stayed very similar. The fundamentals are still the fundamentals, but I’m constantly learning new strategies and ways to start incorporating. So there are more advanced ways that you can start implementing within your prompting. So things like asking the AI system to think step by step and giving it the steps that you want it to take and how you want it to think through something can be really helpful. I’ve learned throughout conversations, asking it to evaluate its own response and then improve upon it can be really helpful. There are strategies called using delimiters, which is essentially putting characters at the beginning and end of sections of your prompt the same like special characters, I use three forward slashes in a row at the beginning and the end. And that gives containers of content so that the AI can understand the differences between the content. You want it to read something and use that as a reference that it knows that’s not the directions. That’s the reference. So lots of different things. I’m also telling it to take a deep breath before it starts doing something, you tell it, hey, take a deep breath and then do this. And a lot of that is keeping up with studies and research that’s coming out and starting to incorporate that into my prompting strategies.

Drew: Deep breath. Why would that even matter? 

Nicole: I don’t actually know. I’m just guessing other people have guessed about what it is. But for humans to perform better, like if you just pause and relax for a second, before you jump into something, take the time to think about it, you’ll do a better job. And the AI systems because they’ve been trained on so much human interaction have probably picked up on that but I don’t know for sure, maybe it really is just forcing them to pause and think it through before but it definitely works. That was not my idea, I would never have thought of that. So you’re suddenly seeing prompts everywhere in the last few weeks. But saying, take a deep breath. That’s because let’s research.

Drew: It’s funny because we have a Career Huddle with Jay Baer, who wrote this little book called ‘The Time to Win’, and it’s all about speed to response. But there is such a something as too fast. If you order your food, and it comes three seconds later, that’s a little too fast, if it was something that was worth preparing, and not fast food. So it’s funny that this is sort of this notion, speed is important but take a second. 

Nicole: If you’re nice to it, it does matter. You know, I’ve always done it. I’m from the south and that just feels really rude when you’re from Georgia, if you don’t say please, and thank you, that’s just the end of the world. So I just can’t not. But it turns out that Microsoft actually did a study on this and it actually improves the quality of the interactions and the results that you’ll get if you use common courtesy. 

Drew: Isn’t that true in everything else? So why not bring that here? That is also amazing and it’s funny because I find myself doing exactly the same thing, naturally just saying, “Oh, would you please do this?” or “Could you think about it a different way? That would be nice.” It’s fascinating. I want to make sure that folks understand the delimiter idea and I didn’t understand it until I saw your posts from about a week ago on LinkedIn. The post was you provided a long prompt, and then you said in the prompt, read this blog post and create an image, then you did the three slash lines, you added the blog post and at the end of it, you did the three slash line. So that’s the idea of delimiters, just to be clear, right?

Nicole: Yeah and that way, you can add different blocks of content. Say you have the prompt, you’ve got your instructions and everything but you want the AI to read your blog post, you want your AI to read your brand guidelines, you want the AI to read your persona information, and you want it to clearly understand which of these things is which of these things, the blog post, if it has instructions on it, that’s not instructions for the AI, that’s what’s in the blog post. By putting these characters on the beginning and the end to show that it is very unique blocks of the prompt and content reference and even title them so it knows what it is. It just helps it to really process and understand your prompt and be able to think through more complicated projects, easier.

Drew: Funny thing on this one is, I copied your prompt exactly, except I forgot to add the blog post. Then when I put that in there, and I clicked enter because I forgot that when you click return, that would just give me a new paragraph, it didn’t, it entered the information. It said “Happy to do this but you got to give me the blog post.” So I thought oh my god, brilliant. So even if you screw up because you said read the blog post, it knows to do it. So we’re staying with ChatGPT for a second and we’re staying with written content. What are you thinking about in terms of copywriting, in generating that, let’s talk about the tools that you’re using beyond ChatGPT and the common mistakes that you see people making.

Nicole: For copywriting specifically ChatGPT is still my number one go-to but I do also use Claude. Claude is awesome also for helping with the generative AI, especially if you need to summarize longer documents into copy, things like that. And then I absolutely love using Microsoft’s Edge browser. Edge browsers Bing functionality that’s built-in. If I have a document I’m working on and I need to generate copy to go inside of it, you can just have your document open in your browser, whether it’s Word Online or a Google Doc, and then use the little Bing button and have side by side to how that helped you with your copywriting.

Drew: I want to go back to Claude, love Claude, if you’re thinking about why do I need Claude. So let’s say you happen to write a book like I did and you wanted to just for fun, upload this book, which has 25,000 words, you can’t do that on ChatGPT. So you upload this long PDF, and then you give it instructions like you would ChatGPT, it’s funny, I did it on our generative AI report where we did man versus machine which I know you’re familiar with. And it gave me a great little summary that we could use as a blog post. And I didn’t have any concern because it was all original IP to begin with. So it was just summarizing original IP,  

Nicole: The number one reason I use Claude is for those longer documents that you need to be able to use or sometimes it can be better at thinking through certain analysis. I don’t think that’s a universal thing but it’s really good at anything around what skill set would be good for something. So it helps me to write like the beginning of my ChatGPT prompts of who I would want working on the project. I actually asked Claude to help me write that part. It’s really good at that if you want to do job descriptions, things like that Claude is really good at understanding who would be the ideal person to do this job. But for most things, I still don’t think that Claude is to the point where ChatGPT 4 is but for a couple of use cases, it is an awesome tool to have in your back pocket.

Drew: We do have a question from the audience, which is, “Is Claude any more accurate, or it hallucinates as well?  

Nicole: Every single one of these generative AI tools that has the problem with hallucination and making things up, you have to fact-check all of them, you cannot assume any of them are definitely going to be accurate at all.

Drew: You mentioned job description. Should I be worried that, yes, if you are a time traveler, that would be helpful. That would be a good skill, right? It doesn’t hallucinate that way. I’m imagining.  

Nicole: No, not quite that way. 

Drew: Okay, going back to and wrapping up this conversation. And there’s so much more to go into with copy but we did cover a lot last time. What are the common mistakes that you’re seeing when it comes to using Gen AI tools for copywriting?

Nicole: Well, there’s a lot, but the biggest ones I’m seeing are not providing the ideas. So just giving a very generic prompt. That’s like “write me a blog post on XYZ topic” and letting it just come up with all of the ideas versus “write me a blog posts on XYZ topic and incorporate these ideas into your blog posts make it about this, this is the key points, I want to make”, huge difference in what you’re going to get out of it, the thought leadership, the authenticity.  

Drew: Let me pause on that one for a second and just put it in the language that I know, we’re talking about intellectual property here and we’re talking about your ability as content creators to rise above and make it special. You’re not going to do that if those ideas are coming from the tools. However, if you have your ideas, and you start with those, and then you want to take those and craft them faster, then you got it. Basically, that’s what our content test showed, although if you just ChatGPT only, even if you’re a master at it is only going to give you baseline information. However, if you have great IP, like an outline, or you have quotes, for example, or you have research that you’ve done, anything like that, and then you can accelerate the speed of it. So it starts with good IP.

Nicole: You can do that really fast and easy. I absolutely love when I want to do thought leadership, I will go to the ChatGPT, I’ll use the little voice button and I’ll just thought dump. it doesn’t need to be coherently put together or organized or anything like that, ChatGPT is great at that. I’ll just thought dump all of those ideas into the voice part on ChatGPT, it transcribes them and then I’ll use that transcript it gives me back, you know, three or four-minute-long talking, I’ll use that and say now use this to turn into my blog post. You don’t necessarily even have to write anything, it doesn’t have to be a lot of work to give it that thought leadership ideas but not giving it any ideas to work from is the first thing that would be a mistake. The second thing that I’m seeing is not realizing you do have to edit it, iterate on it, go back and forth. Every now and then I’ll be shocked. It’ll be like, “Wow, this is in ready-to-use form.” but most of the time, there’s a little bit of back and forth, there’s human editing there so it’s not just ready to go.

Drew: This is a moment where if you don’t have good editors that understand good writing, because I have noticed the tendency of these tools to use words that add more words than they should, their choice of adjectives are pretty boring. Unless you encourage it to use a metaphor, they won’t. You really need to be an editor and really think like an editor, so that’s number two.

Nicole: Take what it gives you and put it into a new chat that’s an editing prompt and get it to improve and iterate on it too. You don’t necessarily have to do as much but that first out, not ready to go. 

Drew: Talk about that editing prompt because I don’t think a lot of people were doing, what does that look like?

Nicole: My editing prompt, it gives it all of the skill sets that the most ideal copy editor would have. I tell it the type of content it’s going to be editing and I tell it that I want it to tell me the before, the after, and the reason for a change that it’s recommending because I want to be able to decide, I don’t want it to just rewrite with the edits applied, I want you to tell me where it is, what the changes and why you are recommending this. Then I put in those delimiters here is the copy you’re going to edit and I delimit it so it knows the difference between the directions and the copy. Then it will spit back suggestions and it will tell you, this is what you said, this is what I would change it to, this is why. I would say I usually apply about 80 to 90% of its suggestions and I will use that with ChatGPT-generated content, I will use it with my own writing, so maybe I write a really quick rough draft, and it can turn it into much higher quality writing, I don’t have to think through all of that.

Drew: I’m imagining that you keep a running sort of Google doc of your prompts for this.

Nicole: Yeah, you know, it’s a two-second thing to copy it in and I just keep my Google Doc, that’s my prompt library, open. Whenever I’m working in ChatGPT, Claude, or any of these tools, it’s always there, so the things that I use a lot get copied and pasted in. I also know what’s in it, maybe I’m doing a novel project, maybe I’ve never done it before, but I know what prompts are in there that could just be quickly edited and then the second I have a new prompt for a new project to get added to that document, I always have it. It gets faster and faster to be able to create.

Drew: Just a note to the CMO Huddles team, we’re gonna create a library of prompts and we’re gonna basically start stealing all of the ones that Nicole shares and we’ll be including those. Not stealing, borrowing, because you’ve already put them out publicly, right? 

Nicole: I only put out a very small fragment. 

Drew: Right, so for the public ones will aggregate them in one place, and then other folks will share it and we’ll all get better at this, which is the goal. We do have another question. “I am now using MagaI, which lets me switch between ChatGPT 3.5 and 4.0, Claude, etc. It has a pertinent personality. Any thoughts on its usefulness of Magai?” Have you used it?

Nicole: I have no idea. 

Drew: Okay. 

Nicole: I have never used it, there are so many different tools out there!

Drew: So many tools, there you go. That’s a great segue to this because when you and I talk, there’s all these plugins now to ChatGPT and you would caution about adopting too many of these because it really is, a look squirrel kind of moment. What’s your story? What’s a rule of thumb for picking a new tool and deciding when to sort of invest energy into it?

Nicole: Yeah, so I have a few rules around this. My philosophy personally is less is a lot more with AI and you don’t need a million tools because the more tools you have, the less likely you are to be good at any of them, know how to use them and actually use them. Especially for your team. If your team has 20 different tools, they’re gonna use zero, if they have one or two different tools, they’re gonna use them all the time. It’s much better to actually build a skill set and tools that can actually do a lot for you, than to have 50 different tools, and you have to know which to use for one. That’s my starting point but when I am looking at new tools, I look at a few things. One, does this actually do something I can’t do with one of the other tools? Is there actually something about this that I cannot do with ChatGPT, that’s usually the first go-to because it can do 85% of what any tools out there can do. I’d rather just use ChatGPT for a couple of reasons. One, it’s got longevity, it’s not going anywhere, anytime soon, so building the skill sets there. It may not be forever, it may end up being MySpace, I have no idea but for the time being it is clearly not disappearing. If I know how to use it in there, the tools aren’t going anywhere, the tools that are being built by the kid in his mom’s basement with no funding, probably not going to be there in three months. That’s what so many of these tools popping up are, a lot of the people that are building these tools have no idea how much it costs to run AI. The first thing is can a tool I already use do this? The next thing is if the answer to that is no, who the heck is this company and what is their funding situation? If they do not have money, they are most likely not going to last, so you’re going to invest time, resources, energy and putting a tool into your workflow, and if they’re not going to make it, then that’s a lot of wasted time and energy. Also, do you trust their data security? How are they using your data? Who are they? Am I giving a foreign country my data or am I giving it to a company that I’ve locked down because that’s important with all of this. My own stack is a pretty small set of tools that cover the big things. ChatGPT is my overarching main tool I use, I’ll use Claude to supplement that where ChatGPT can’t do it and sometimes to just get secondary ideas. Claude and Bing Edge, I’m not paying for, those are inclusive, although if I had a business I would pay the five bucks a month or whatever for the enterprise. If I need voice stuff, ElevenLabs. If I need a couple of video tools, I’ll use Descript and Runway and I’ll use Midjourney, although who knows how long I’ll be using Midjourney with Dall-E. Voice, video, image, and text, those tools cover the vast majority. Now if I needed to do translation, I probably have a tool specific to doing translation but that’s it. I don’t think you need 10,000 tools. In all reality, a lot of the tools that can do the same thing as ChatGPT, ChatGPT actually does it better. 90% of the time, even if you play with it, ChatGPT’s version of it is so much better, so I don’t see the point.

Drew: We do have an ongoing list of a Google sheet with a bunch of utility apps and tools that CMOs in our community have used and shared but I think what we’re going to do is we’re going to add Nicole’s top six to that document. Okay, you mentioned Dall-E and I have to tell you, so I hadn’t used it until yesterday when I was preparing for this show. I am going to share my screen for a real quick second because this is amazing. Before I show the screen, I just want to see this premise, I was looking for Nicole’s prompt for creating an image based on a blog post. The blog post was about our generative AI test where we did the man versus machine, so in literally three seconds, Nicole said ask it to do four. This is the first one, there’s the man, there’s the machine. This is the second one. There were two others. This is a very elegant design for a blog post. Again, I didn’t do anything else with this. Then we have this one. Then I said for fun, alright, well, CMO Huddles is, you know, our mascot is a penguin. I then said create the same idea, but use penguins. So this is what it came up with. That’s really cool and then I said, okay, I need a square image, so create a square image and it created that. Oftentimes for our blog posts, shame on us renegade.com, we’ll just use an image from who knows, some image house. This is a lot better than the stock image. By the way, I had never used this tool until yesterday. Now, since we’re in Dall-E, and we’re looking at those images, what other things should we be doing when we’re using Dall-E within it,  to sort of iterate?

Nicole: A few different things. One is you can just tap back and forth to make those adjustments, which you did but you can keep going with that. You can also ask it to make you different variations of the exact same thing but give me four variations. That’s one thing but the other thing is that was an example and that prompt is asking it to come up with the idea itself. You could also give it the idea, you can give it the context, or like I ned an image that represents X, Y, and Z. You could give it an actual prompt like create me an image that has a man who is doing a podcast, he’s standing in front of a bookshelf he’s talking to a microphone, I want it to be a cartoon, I want him to have a thought bubble that says CMO Huddles, you could be that specific with it and then it will turn around and turn it into the ideal prompt for Dall-E to do it. There are a few limitations with Dall-E that we should all know about, if we’re going to talk about getting into it though, one being the spelling is not going to always be perfect. It’ll be hit or miss. Now those of us who have been using image generation for a minute are overjoyed with what Dall-E is capable of when it comes to text and spelling because it used to look like any of these tools until Dall-E 3 came out. It looks like Egyptian hieroglyphics meets ancient Greek meets complete gibberish, there was nothing that resembled text. Now at least it’s just misspellings. A lot of times you will have issues getting stuff spelled properly so use words with limits to it. The other thing you need to know is it’s not nearly as good at making realistic-looking photos. Midjourney still is significantly better at making real-looking humans, animals, whatever it is. So it’s not fantastic about that and you need to look out for, does the hand have six fingers, does it have an extra head, is the foot backwards, those little things do pop up, now that pops up in all of them. Just be aware of those couple of things, really just chatting back and forth and it’s so new, Dall-E 3 just really hit our account so we’re all still learning. The other thing is if you do use Midjourney, or image generation, when you click on the image, it will tell you the prompt it used and you can take the prompts because ChatGPT is going to write a prompt, you will copy that prompt and put it into Midjourney and it may create you an even better quality image, just forget the writing. The other thing I love doing is taking my Dall-E image, opening it in Photoshop, and editing what I don’t like in Photoshop because you have generative AI and Photoshop too. If it’s like 85% of what you would want but you want to change one little element, you want to remove the random word that’s misspelled, just go into Photoshop, select it, type in the generative AI that pops up what you want it to be, how you want it to change, it could take a little work.

Drew: This Is the big difference between Midjourney. Dall-E within ChatGPT was a revelation to me because Midjourney has been a journey and I do continue to believe that you want someone on your team who is truly an art director, going into that tool and taking advantage of it, we had Kenny Friedman, he’s amazing. He’s an artist, and he really knows how to do it but for our blog posts, and I’m not worried about copywriting or anything like that. It’s awesome and it’s so easy, and there’s no learning curve.

Nicole: What’s so cool is anybody can get out of the box and integrate and change, you don’t have to know how to write the image prompts to get good stuff. One thing I would caution because these are, what this group of people is, before you are putting any AI-generated images into your public-facing content for your company, just make sure your legal team is cool with it because they might be fine with your AI-generated writing and not be okay with AI-generated imagery. So just be aware of that.

Drew: The Supreme Court has already ruled you can’t trademark these images. So that in itself tells you that these images you will not own. If you care about owning the content that is on your website, then obviously, this puts these tools in a whole different way. Except you can go to other places to get these images that have the same tools where you’re paying a stock photo company for access. So it’s just not as easy.

Nicole: I’ve just seen a lot more like the legal side not okay with AI imagery. So it’s just something to be aware of as a marketing leader.

Drew: We talked about the tips for copywriting, any other tips when it relates to image creation, specifically with Dall-E?

Nicole: So you can give it like style. You don’t have to just say create an image of XYZ thing. You can tell it the exact style you want it to be in. You cannot tell it a specific artist, which I think is great that they built that in but you can tell it cartoons, specifics. If you want to know – you have a style that is on your own website, like you always do it, and you have no idea how to describe this to an AI system, because it’s just the art you’ve always created, you can go into the default model of ChatGPT 4, paste in the images that are examples of the art, and ask it to write you a description that you could give to Dall-E of the style of art because now you can put images as part of your prompt. So it can analyze the image and go, “This is neoclassical,” whatever it is that you want to replicate a concept, you could ask it to give you that description so that you can get closer to it with Dall-E.

Drew: Okay, I can’t wait to test that. I can think about it as in renegade.com we have various sections of images and you could say, look at this image, and now I want to replicate it in a different way for a different section.  

Nicole: Yeah, I was preparing a training for a client, they have very specific graphic art all over their website and one of the problems they’re having is they don’t have that artist anymore, they can’t replicate it. I put it into the Vision tool and asked it to describe it- how it came back, I was like this cannot be a real style of art, it was like, “futuristic work cartoon.” Lo and behold, I told Dall-E to create something in that same description, and it looked exactly like the stuff on the website.

Drew: I’m having a little bit of a pain moment in understanding the writer strike because this is a moment, from an artist’s standpoint, the fact that you could do that. Talk to your legal teams before you do any images or anything like what we just talked about. Okay, I know you’re excited about Vision. I don’t think I have access to it, it’s in beta, is that right? Or is…

Nicole: You should have access to it. To get to Vision, it’s actually showing up when you go to your paid ChatGPT account, select GPT 4, and you just select the default model. Now there is a little picture icon that shows up on the left side of where you enter your prompt – where you actually go to enter – there’s now just a little picture icon and if you click that, you can upload images, as long as that’s showing up on yours, you can put your cursor in there and you can just paste it out of your keyboard with a keyboard shortcut. If it’s like a screenshot, you can just paste it in, you don’t have to download the screenshot, upload the screenshot – much faster copy.

Drew: So the applications of this, you could drag in your landing page of your homepage and ask it to assess it and ask it to what? 

Nicole: SEO, Conversion rate optimization, design suggestions, come up with a new image that we could put on the banner? What do you want? It can see – like a person would be able to see and understand visually – see your screenshot of your homepage, or your landing page, or your PowerPoint slide. So anything that you may want assistance on, it can see and comprehend. It just becomes part of your prompt context.

Drew: I’m really curious about that. Let’s just go back to the example of: “Do an assessment of our landing page optimized for target audience engagement,” right? And we’re gonna say, “Our ideal customer profile…” Do an assessment for this and make sure that we can decrease our bounce rate and increase the path to say, “demos.” How do we trust its analysis?  

Nicole: Partially your own expertise comes in, “Well, you know, I didn’t even think about that in this context.” You know, you didn’t notice it, because it’s your own brainchild. I’ve seen it so many times – you, like, miss the things. So it’s not to replace your own expertise completely. Use your own logic, use your own analysis, use your own marketing expertise to go oh yeah, these suggestions actually make sense. But also, you shouldn’t just be randomly applying AI added – do the A/B testing. And so this is a really good way to go. “Okay, now we have something to A/B test.”  

Drew: But I’m also thinking about a CMO who’s interviewing for a new job, and they want to look at the top five competitors and they could say to this tool, analyze these five websites and compare and contrast which ones are doing better at this. It’ll do it in a few seconds… 

Nicole: Yeah, maybe two minutes.

Drew: Okay because we’re going to tell it to wait and we’re going to be polite.

Nicole: I do recommend you don’t put more than three images on a single prompt – if you need to do multiple prompts. I mean, it’s like if you need to do more than three, change the prompts. So like, “Here’s the first three… wait and tell me when you’re ready for the next three,” – it can prevent it from getting a little bit more confused about the images and it really helps to not do more than three at a time..  

Drew: Even just to compare and contrast these three. We have a lot of CMOs on the “Transition Team” right now and they’re preparing for interviews. People are asking for lots of insights more. This may be getting into our next area but it could take hours to prepare. These tools could change that.  

Nicole: Completely – I mean, makes it much easier.

Drew: So we’ve talked about how we could sort of use it in that way to look at and challenge your current website, or design of any visual thing. Let’s talk about other things like the strategic use of it. So again, I’m preparing for an interview and I need to look at this one brand, and then I want to look at say two others. I want to understand the really unique differences, if there are, between the three. How would you do that?

Nicole: Give it the screenshots and tell it what you want it to do.  

Drew: Are there any prompts from your experience that you’ve done to help you get strategic insights?

Nicole: For that exact use case, I’ve never done exactly that. But strategic in general is where it can really help to ask it to think about like this, and this in order. Follow these thought processes, explain it to me, now evaluate your response, and is there anything that you could do better? Like, tell it “You did a great job” – like always be, “You did a great job. This is a fantastic start. I believe in you. You could come up with even better suggestions. Evaluate this and tell me how we can improve upon it.” That can really help with a lot of the strategic stuff to get better by asking, “Is there anything you miss? Please review this. Is there anything else that you didn’t mention?” Ask it to evaluate its response and put it in order of importance – all of that kind of stuff. Asking it to really think through what it did is really helpful. Also, regenerating multiple answers at each step to what I did, and seeing what it said because it may think about it a little different each time. And so you may get even more insights by just taking from the best parts of each generation.  

Drew: Yeah and then I’m thinking in terms of a philosophical structure for how you approach brand, for example, and you’re all about Geoffrey Moore and “Crossing the Chasm.” You could say, “Let’s look at that from this standpoint.” Or you could be a big fan of “Renegade Marketing” and that book. Then you could say, “Okay, look at it from that standpoint…”  

Nicole: You can give it the exact framework you want it to work within. Ask it to give it to you in a specific output style you want. If you want to put it in a table if you want – and even just strategically – like one of the things I’ve done is totally different than that. But I took my LinkedIn posts on my own page. I just expanded them – like I went to the “Posts” feed on LinkedIn, I expanded all of them out and I started taking screenshots. If you had short posts you could do it much shorter – put it into vision and ask it to turn it into a table that had the topic in three words, a two-sentence summary of the post, and the number of comments, the number of reactions, the number of shares for each post. And so I got a table back that showed me all of my content that I had created because it could actually read it. And it could see the number of reactions to it. It could see the number of comments. I could see there – and it took a matter of seconds to get a table of all my LinkedIn stuff. If you wanted to do this for the social media of a company you’re interviewing with for example, you could potentially do that. And then ask it to analyze what type of posts performed the best. You could ask it for different things in that table and then to run that analysis.

Drew: I was just thinking again you’re a new CMO – a strong CEO has been out there and expressed a certain fondness for an author or a point of view. Again you can take that and almost create that character and use that to assess various things. So we’re now past copywriting. We’re past art direction. We’re into all sorts of other ways that you can use these tools. Let’s talk about a couple. We don’t have enough time to go through. But things that you’re finding that are just so helpful to you.  

Nicole: Everything at this point. It’s hard like when people are like “What are one or two things you use it for? And “What are one or two things don’t use it for?” I find it so helpful for literally everything I do and I operate from this philosophy of the first thing I ask myself before I jump into something is “How can AI assist me with this?” I try to apply that every single time. Sometimes it takes me a little bit longer because I use AI, I’m not gonna lie – sometimes it saves me 15 minutes and sometimes it saves me eight hours or 10 hours or 20 hours. I just kind of am in this mindset of “Let me try” unless I am consciously making the decision, I do not want to use AI for this – because either I internally just like really want to write this myself or I actually want to make sure that my writing abilities don’t atrophy. So like every now and then I’m like “You better write this yourself because you don’t want to forget how to do this.” The answer to that is I find it useful for everything and I think that that’s the mindset shift that a lot of people need is to start going “How can I apply this to all the things?” If you save five minutes 10 times a day you’ve just saved almost an hour.

Drew: And for everything that CMO does, obviously. I want a tool built into Microsoft Outlook that reads my email and says, “Okay, do I need to respond? What is there a generic response that I have?” There’s all sorts of things that I haven’t seen yet. I’d like it to look at my calendar and say “You have these 10 meetings coming up. Automatically prepare for them.” For example, I don’t have a prep call with any prospective Huddler or prospective podcast guests without doing a trial interview with them. “You’re Drew Neisser, you’re the host of the podcast. What kind of questions would I ask?” I don’t use those questions, typically. But they make it so much easier. I now know what’s generic and it’s so helpful. Whether it’s interview prep, I used them for travel planning, we just got back from Barcelona. I said, “Okay, think about the top 10 things to do in Barcelona, organize it by various areas, give me lunch recommendations, tell me what the time difference is between each place.” And it laid out a pretty interesting schedule, which we didn’t actually do exactly, but it was still helpful.  

Nicole: I’ve used it for a couple of trips now, and it was an amazing travel advisor because you can tell it exactly what you like to do, don’t like to do, how long you have. Both my trips were like road trips, and it was ready to get step-by-step turn-by-turn directions for little driving tours of areas and finding gems we would never have found ourselves, so I love it for that. Also one of my favorite things to do, this is like more personal life, although I think it could apply professionally too, if you have like an annoying situation you need to just send an email to deal with. So you could go in again to that voice app and ChatGPT on your phone, do the voice, explain the situation and just ask it to spin it into an email to deal with. You can literally just explain verbally “This is what’s happening. I need you to turn it into an email” and then do this, and then go to your computer and copy and paste that email out. So I did it the other day to deal with something in my apartment. Took three minutes and no energy.  

Drew: This is a big one, folks – annoying emails. These are the emails that make you sweat. These are the ones that you write the draft of and then you leave it there for half an hour and you go back to it. These are the ones where you’re worried you might upset someone or you have bad news that you have to share with someone, those emails which are a huge time suck, at least for me, because you know, I don’t want to hurt anybody’s feelings. Boy what a tip that is.

Nicole: It is so good if you need to communicate an email with somebody who has some type of a difference in their communication preferences. Like I’m Southern and I’m dealing with somebody who is Northern and super direct, and we like to talk around things, or I’m neurotypical and I’m dealing with somebody neurodivergent or the opposite, or I needed to send emails to people in Israel and the communication styles are different. “This is my natural communication style. This is their communication style or this is their culture that they’re coming out of. This is the context of our relationship dynamic and here’s what I would like to say but I need you to write this in a way that’s going to be well received by that person, even though they are located in this other country and they’re not American like I am,” or whatever it is and it can spin something instantaneously to be fitting for the actual environment of who the recipient is.  

Drew: And then you’re, again, you’re spending your time in this case editing and not drafting, and just painting yourself over and over again to try to get the words perfectly. Okay, we’re running out of time. There are so many other things that are coming up in terms of advanced data analysis, or just something as simple as looking at a website and say “Hey, create the color palette that we’re using,” all sorts of things that we’re going to have to do another show on, but obviously we will do that. You are one of our top go-to resources for generative AI. Who are some of yours?

Nicole: I’ll give you two that I love. One is a man named Ethan Mollick. He is a researcher, he’s a professor at the Wharton School and this is like his space, so he’s one of the people that does the actual research studies on this and shares a ton of data and lives and breathes this stuff.  

Drew: How do you spell Mollick? 

Nicole: M O L L I C K. Definitely like my number one social media follow. He’s on both LinkedIn and I believe he’s on X. The other person I absolutely love is a woman named Rachel Woods and she has a group called the AI Exchange, that’s like a paid community, but she posts a ton on LinkedIn and TikTok as well. Absolutely awesome. So those are my two go-to’s. Honestly though, the vast majority – I don’t follow a lot of people on social. I don’t trust most of what’s on social to be honest, it’s mostly people who have no idea what they’re talking about, talking like they do.  

Drew: Alright, well we’re running out of time but two do’s and one don’t for CMOs when training their teams on taking advantage of gen AI.  

Nicole: The first one would be before you start projects, ask them “How are you going to use AI for this?” So when you’re assigning the project, ask them that and point out ways that they can if they’re missing it in their blind spot and then when they’re finished, ask “How did you use AI for this?” The second one is create a culture that celebrates AI wins and failures. So just using the AI where you’re lifting up and praising publicly, especially publicly, because then everybody sees it and knows “Listen, appreciate it when somebody does something awesome with AI,” pointed out, celebrate it – “This is a win that everybody should aspire to.” And when somebody fails with AI, also celebrate. “Like they used it, they tried and that’s okay. And now you’re gonna learn and what did you learn from this? That was awesome.” So those two things for getting your team to adopt AI. The don’t is not going to be about adopting AI, it’s just a thing that is in my head – please stop doing this with AI, not necessarily this group at all. When you see somebody’s LinkedIn post and you want to make a comment on it, do not copy and paste that LinkedIn post into ChatGPT, ask her to read a comment, and then the comment without actually reading the comment. If that comment is the same words as the poster repackaged in a new metaphor, do not use it. Everybody knows that you’re commenting with AI. We got to stop.  

Drew: Yeah, exactly. So at that moment in time, if you’re going to actually comment as a human be a human.

Nicole: Rule of thumb – if you’re commenting on LinkedIn, be human. If you’re posting on LinkedIn, totally fine to use AI to produce faster, engage on some person.

Drew: Assuming it’s your IP, it’s your idea, produce faster.

Nicole: Engage as a person.  

Drew: Amen to that. Okay, first of all, thank you so much, Nicole and we’re very polite here. So maybe I’ll thank you twice because I want to be better than I would with the machine. But where can people find you? And what would they hire you for if you happen to have any capacity right now?

Nicole: I can always squeeze people in. So LinkedIn, my website is nicoleleffer.com – so it’s super easy. Those are the two best places and I’m also on TikTok although that’s not super marketing specific, that’s more of just the place that the random stream of consciousness about AI comes out. For work, I am working with CMOs and their teams to help them adopt AI. So the main thing I do is work with a CMO to understand pain points and what they want to do with AI, or maybe they don’t – they just know they need AI – and then design custom trainings to teach their teams the skills and the use cases to use those skills specific to the company. I also do some generic trainings as well.  

Drew: As I mentioned at the top of the hour, at least three Huddlers have hired Nicole. Insofar, at least one of them reported back extremely positive. So four gold stars out of four. Anyway, Nicole, again thank you so much for joining, can’t wait to follow what else you come up with in the next six months.

Nicole: Thank you so much. This was great. I really appreciate it.

If you’re a B2B CMO, and you want to hear more conversations like this one, find out if you qualify to join our community of sharing, caring, and daring CMOs at cmohuddles.com.

Show Credits

Renegade Marketers Unite is written and directed by Drew Neisser. Hey, that’s me! This show is produced by Melissa Caffrey, Laura Parkyn, and our B2B podcast partners Share Your Genius. The music is by the amazing Burns Twins and the intro Voice Over is Linda Cornelius. To find the transcripts of all episodes, suggest future guests, or learn more about B2B branding, CMO Huddles, or my CMO coaching service, check out renegade.com. I’m your host, Drew Neisser. And until next time, keep those Renegade thinking caps on and strong!