How to Harness the Power of AI for Your Book Marketing Success - BM404

Susan Friedmann [00:00:00]:

Welcome to Book Marketing Mentors, the weekly podcast where you learn proven strategies, tools, ideas, and tips from the masters. Every week, I introduce you to a marketing master who will share their expertise to help you market and sell more books. 

Today, my special guest is Michael Goldrich. He's the founder of Vivanda Advisors and a digital and AI strategy whiz. His book, “Too Many Hats, Too Little Time,” offers smart AI strategies for business efficiency. An educator at Northwestern University, Michael makes AI understandable. With a rich background in global marketing and a blend of literature and business education, he's an award-winning influencer at the intersection of technology and innovation. 

Michael, it's an absolute pleasure to welcome you to the show and thank you for being this week's guest expert and mentor.

Michael Goldrich [00:01:05]:

Thank you so much for having me. I am so excited to be here.

Susan Friedmann [00:01:09]:

Well, Michael, AI is your genius and I want to really tap into that because It's such a new technology, and I'm not sure how many authors are actually taking advantage of what AI can do for them. So we're going to have that wonderful conversation, tapping into your wisdom and genius. But before we get started, let's get very granular, which is what I like to do, as you know, and that is Let's start off by understanding what exactly AI is.

Michael Goldrich [00:01:47]:

Thank you for that, Susan. I'll keep it very, very high level. AI is just a branch of computer science, artificial intelligence, of which it has different components. The first sort of subset is called machine learning, which we've all been doing for a while. Netflix uses it. Amazon use it, and it's all about prediction. What they do is, machine learning does, is it looks at your past behavior And then based on those patterns it sees, it makes recommendations. And that's why Amazon knows what books you're gonna like.

Michael Goldrich [00:02:18]:

That's why Netflix knows what TV shows you're gonna like. Now a go a little bit step deeper, and what came out in last November is generative AI. And what generative AI is It's a form of this machine learning, which is all about patterns. But instead of making predictions, It's actually generating content. It's generating content based off of all the content that's already exists in the Internet, it looks at how words go together, how concepts go together, and it just generates it automatically. And it doesn't just do it for text, It does it for images, and it does it for videos as well. But it's all about generative. So when it come about AI, it's just a way That computer scientists are trying to use machines to help replicate human decisions and decision processes.

Susan Friedmann [00:03:07]:

There's a lot there. Wow. Yeah. But we don't necessarily need to know how it works. What we need to know, Obviously, and authors want to know it's like, hey, how can it help me? So how can it help authors market their books?

Michael Goldrich [00:03:25]:

That's interesting because I'm a new author. I write about leveraging this technology to help in decision making. What I've written about is when I first learned about generative AI, I was learning about how do you create The prompt. How do you actually ask the system what you want? And one of the things I learned is is that you're supposed to give it a role. You say act as a marketer, Act as an author. Act as a book publisher. Then you give it some context and questions. But I thought, well, if you ask it to act as something.

Michael Goldrich [00:03:57]:

And I said, I wonder if you can kinda give it a a persona. Instead of just act as a digital marketer, can it be acting as Michael, the CMO of this company with this set of expertise? And then I put that in the system, and it was able to do that. Then I thought, well, that's really interesting. I wonder if I can create not just a CMO, but I wonder if I can create a social media expert that kinda goes with the CMO, And I could. And I wonder if the 2 of them exist at the same time when the little chatbot in they can. And then I thought, can I have them ask each other questions? And they can. So then I thought, well, this is amazing because then now if I have a problem I'm facing at work, if I ask these personas and I can create 1, 2, and n number, I present the question to them and they'll give me back Answers based on their expertise.

Michael Goldrich [00:04:53]:

So then leveraging that, so that's what I write about in the book is how to create these personas. Now as a new book author, who's trying to get people to see my book, I was thinking, well, I really don't know. Like, I spent all my time writing this, but don't really know anything about marketing books, so I'm like, well, why don't I practice what I preach? And I went ahead and I created personas. I created An AI publishing PR expert, an AI content specialist, and an AI social media manager, Then I created an AI Amazon Ads expert. So now I have the 4 of them and I'm using them to try to help guide me in terms of how do I market my book. What is the best strategy for me to go ahead? And I asked them to prioritize and then have them challenge the other's list, and then come up with an approach, and then go from there. So actually, that's what I'm doing. And actually, that's how I ended up on this podcast with Susan today, Because my PR specialist told me, you know what? You need to get on a podcast, and here's a different approach to do this, blah blah blah. And then I reached out to Susan, And Susan's like, oh, yeah, Michael.

Michael Goldrich [00:05:59]:

I'd love to have you on.

Susan Friedmann [00:06:01]:

Wow. These personas are telling you what to do. Did they specifically ask you to be on or Tell you to come on to Book Marketing Mentors?

Michael Goldrich [00:06:10]:

They did not tell me that, but they said that I need to find Podcasts in which I can speak about this. So I then started to research various podcasts where I thought would be appropriate. I took this information and then went to another resource. However, had I taken these personas and I took the next level and kind of integrated them into these other sorts of data sources. They might have done that. But right now, they're stand-alone because this technology is free. What I present here is free. Anybody can do this, and you don't need to be technical to create these personas.

Michael Goldrich [00:06:47]:

You just need to Know that you can do it. One of the biggest things that I get from the book when people read this is the idea, like, I didn't know it could do that. Once people play with it and actually use it and practice, then it's like this really is your own personal advisory board. Can help you with anything. It helps me in terms of being a solopreneur. I can create my marketer, my CFO. I create even my general counsel that helps me Cut down on my legal bills in terms of looking at contracts and statements at work. In addition, it now is helping me on my book marketing, which is great.

Michael Goldrich [00:07:25]:

Obviously, you can't take everything at face value because it is all generative. Some of the information may not be exactly correct, But I think it's directional, and so it helps you. It gets you 65, 70% of the way there faster, which is amazing, in my opinion.

Susan Friedmann [00:07:44]:

Yes. It certainly is. And when you talk about all these different personas now I don't want to get too sort of in the weeds and Too overwhelming for our listeners. Where do you start? Which persona, for instance, would be the most important one for an author to start creating with AI?

Michael Goldrich [00:08:06]:

That's a great question. What I would do, and what I did, is I sat down and I went to chat GBT, and I said, I am a new author. I just wrote this book. This book is about this stuff, x, y, and z. What are the 4 personas that I should create that will help me Generate awareness from my book? Give them a name. What is their expertise? And a background and capabilities. And so that's actually how it spit out for me.

Michael Goldrich [00:08:37]:

This PR specialist, which was great, spit out for me the social media manager, The social media content person in terms of kinda gave me a plan. And then lastly, the Amazon ads persona tells me really What I need to do to play the Amazon game and try to get awareness and views there. One of the things I did try, I just thought, maybe I need like a book agent. So I created a book agent as well, independent, but the book agent was really Trying to get me to sell my book to a publisher, but I'm already published, so I realized I don't need the book agent. I think the publicist is the most important because this persona kind of knows a little bit about my situation and then You've fed enough context to, like, be able to give you some direction.

Susan Friedmann [00:09:29]:

One of the areas that I know many of our authors are challenged with, and that is finding their target audience. They think they know it. It's women, it's Men of a certain age, in a certain demographic, but really narrowing in maybe with regard to their message and the right more niche audience rather than just a target niching it down they find challenging. How would you create the right prompt to give you the answer that would be right for you in this situation? Is that possible even?

Michael Goldrich [00:10:08]:

Of course. So what I would do is I would create a marketing persona. I would create a marketing expert persona for your book, and I would basically prompt it and say after you create the entity, you'd say that it is an expert In finding perfect audiences for nonfiction books. It is perfect at kind of getting the awareness. So all the sort of stuff that you want, So you create this expert. Like, if you were to go out and hire somebody, what are the capabilities that somebody would hire to help you do that? That's what you just put in here in terms of this persona. And after it's all loaded in, you basically say, help me find my audience in terms of what you Who is my audience in terms of demographics, in terms of location, in terms of, socioeconomic level? Here is my book. Here's what it's about.

Michael Goldrich [00:11:01]:

Can you help me identify my audience? Ask me 10 or 15 questions that will help you identify the perfect audience for my book. Then you can load that up, And it'll spit out the 10, 15 questions. And based on how you respond, it'll give you a response, and then you can Take this, and then you feed it back in and say, can you please validate that this is the right audience? What would you say are other books out there that are similar, that resonate, so you can have sort of like a catch for there? But I would think that's how you can do it. You can have your audience kinda picked out for you. I would say in, like, 15, 20 minutes. Bang. Just like that.

Michael Goldrich [00:11:43]:

Of course. That's my opinion.

Susan Friedmann [00:11:47]:

But, yeah, it gets right to the meat of it. It's really not only asking for that, but then going deeper with it, because I think that's often With AI, what I've found when I do it is not necessarily going deep enough and just you ask it a question, it comes up with an answer, but not saying, hey, can you go deeper into that answer and give me more information? So 10, 15 minutes, that sounds cool.

Michael Goldrich [00:12:18]:

Yes. Yes. And actually, and I talked, Susan, to your point, there are different ways to do prompting. So there's and I think a lot of reasons why people have challenges with, like, this chatbot. It's because for the past, like, 10 years, Google has trained us to basically put into their little search field, like, 3 or 4 words, and we come up with a result in the answer, Which is basically a search of book links, whatever. Now when people go to this chatbot, it looks very similar to what Google has, and a lot of people Put 4 or 5 words in there and not very complete thoughts, and they expect to get something very different. And the thing is, how this chatbot works is the more thoughtful you are, the more context you put in there, The more sort of organized you are, you're gonna get amazing content. Now it isn't just the 1 and done.

Michael Goldrich [00:13:18]:

Remember, it's a chatbot. So you chat, so it'll give you something, and then you can react to it. Say, you know what? That was kind of okay, but can you be a little bit more detail, can you go with descriptive? You know, point number 2 is interesting. Can you elaborate on point number 2, please? And it's the other thing that's interesting as well as I'm reading some research About this is if you're polite to the chatbot, you actually get better content. If you use please and thank you, it actually is more responsive And gives you better quality feedback, which is weird.

Susan Friedmann [00:13:50]:

It is. Because I've heard both sides of that is, like, you know, just be nice to it. Treat it like you would somebody who's really important to you. So you're right. Using your manners, your please, your thank yous help. Why is that? What is it? Has it been programmed to just react differently if you say please and thank you?

Michael Goldrich [00:14:11]:

I am not sure. I'm not even sure if the people who designed the note, because another way that is similar but a little bit different It says when you ask the question, say, take a moment and think about it. And they said that just putting that kind of Input into a prompt as well, saying to take a moment and think about it, also gets better feedback. So I am not sure. Remember, it is trained not just books, but it's also trained off of interactions across people. It has some social media that it's been trained on, so it's possible that it's picked up on something on in terms of how people interact.

Susan Friedmann [00:14:51]:

One of the things that our listeners love are mistakes. I mean, this is like opening Pandora's box with this new technology. Well, it's not necessarily new technology, be it new to us, let's say, in terms of being able to use it effectively. What are some of the sort of common mistakes that you see that users are making? I mean, whether we're an author or whether we're not, but just generally.

Michael Goldrich [00:15:20]:

Alright. So generally, you ask it a question, and it's gonna give you What you perceive as some facts. Those facts are probably 50% of the time wrong. You cannot Take the content of when anything it says that is factual as a truth. This is called hallucinations. That's what it's called when they mix this up. Because remember, it doesn't think. It doesn't know.

Michael Goldrich [00:15:44]:

What it's doing is it's doing, statistical analysis as to what words should come up, and which phrases should come up next. But it's not validating that, and so it's pulling from the entire Internet. It's possible that this information was wrong somewhere, And it's just sort of replicating that as well. That is why it's so, so, so important, as they say in the, the space is to keep the human in the loop. People need to validate. Anything that is remotely factual, you need to validate before you pass that on. There is a lawyer who submitted a brief to a judge, and he just kind of put this all through the chat engine, and he said he had a case, and he realized that he put it in, and he came out with perfectly cited cases with legal reasoning, and A 100% proved everything that he wanted to prove and handed to the judge, and the judge is like, all this is made up. None of this is real.

Michael Goldrich [00:16:42]:

So you have to have to have to make sure that you check that. The other thing is you should never ever put any personal identifiable information into this. So I shouldn't put in people's names, their addresses, or anything to do with money about them because this goes up to the people at the companies, whether it's using Google Bard or using Chat He goes up to OpenAI. He sits out there for 30 days, and those people who are looking at this, could look at this to see As they evaluate information that are good answers or bad answers, but other people might see this. And it's also possible That it can get incorporated into the overall language model. This happened to Samsung where they were trying to validate some of their proprietary Terry technology, and they put it in there to ask some questions. And then the next thing I know, some of this proprietary technology Samsung is now out there for everybody because it just got incorporated into language models. So you have to be very careful as to what you put in, and you have to be very careful at what comes out.

Michael Goldrich [00:17:49]:

You have to Trust but verify.

Susan Friedmann [00:17:52]:

So what's interesting here with regard to validation is how do you go about validating it because It's coming from something that's already out there, potentially, and that may be correct or incorrect. How would you go about validating the facts that come out?

Michael Goldrich [00:18:12]:

What I would say is, for any facts that come out, I would say, please give me the source of where you got this and the link. And it's Possible the source and the link might also be made up, but it gets you at a starting point. Then you can also what I sometimes do is I jump back and forth Between different chatbots. So when I take home ChatGbt, and then I'll take that output, and then I'll run it through Google Bard So look through this. Can you validate any of this information for me? And then can you put source things with all the facts? And then I would take from there. Then I might even go to a third one, which is Anthropics Claude and also validated. So there are different language models out there that do this chat generative AI. You could check it across multiple versions, and then you'll pick up probably about 70% of it doing across like that.

Michael Goldrich [00:19:06]:

But then you just have to kinda do your own Google search after that.

Susan Friedmann [00:19:10]:

Yeah. Do your due diligence before we start putting stuff out there. Indeed. Let's talk a little bit about writing content, and creating content whether whether you use chat or bot or Claude, or yes. You want to create, let's say, social media content or an article. How do you go about that?

Michael Goldrich [00:19:32]:

Well, say, do, and say 1 new social media post about, say, my book, Too Many Hats, Too Little Time. I would first work with my social media AI to create the calendar for me. So it's okay. You know, it says, okay. It's December 22nd. You're supposed to create a Facebook post. I would then trigger off the AI social media manager That is a specialist in creating content. Say, okay, I need to create my December 22nd post based on the calendar.

Michael Goldrich [00:20:01]:

It's supposed to have a Holiday theme, but I also want to get people to buy the book before Christmas, have them download it before they go on their vacation, and create 4 or 5 different potential posts with image recommendations that could really create urgency. And so that's what I would do, and then I would create those 5 posts. And then I would look at it and say, you know what? I like the first one, and I like the third one. Can you elaborate on both of those? I like this part, but I'd like to tweak this other part. And so it can do that until you can get it refined. And then from there, you can kind of say, copy and paste it and go from there.

Susan Friedmann [00:20:44]:

Now I've read many articles where they start talking about languaging, the words that AI uses. And I've seen that they use the word that a lot. They use adverbs. They love the "lys", you know, lovingly, beautifully. You know, they love that kind of thing. So those are giveaways that might be AI-generated. How can you overcome those?

Michael Goldrich [00:21:12]:

That's perfect. So that's a really good question. I mean, there are also words, like, there is a word like realm, loves the word realm, loves the word delves. What you can do is, in ChatGpt, there is an area called custom instructions. And you can put in the custom instructions, Essentially, never use words like that. And you can actually say use sparing use of adverbs. Use sparing Uses of analogies. Structure the sentences in a varied way so it's not very robotic.

Michael Goldrich [00:21:48]:

So you can address that in the custom instructions. In addition, you can add training. Say, you know what? You say, I want you to write in a particular style, so you can also train it on, say, your writing style. So say, here's how I have written in the past. Please take a look at these 4 or 5 articles I've written. How would you define my style? How would you define my word choice? How would you define my whatever? So you can come across 4 or 5 ways to describe it. So then now when you go ahead and write something new, Say, I want you to write it with this different approach, but never using any of those adverbs or whatever.

Michael Goldrich [00:22:31]:

And so that way, it'll be a cleaner, more succinct approach. Now what I do is after I do that, I have created an AI persona editor that also kinda goes through gives it another one through To kind of clean it up even further, and so I run it through that editor that I have, and then it kind of spits out a nice clean version.

Susan Friedmann [00:22:55]:

Wow. I've got work to do. Michael, this is so fascinating, and I know our listeners would like to find out more about the book, or getting in touch with you, or both. So how can they do that?

Michael Goldrich [00:23:09]:

Thank you. They can actually go to The book is on Amazon, "Too Many Hats, Too Little Time," where they can look at it there. They can also go to my company site, vivanderadvisors.com. And on there, I have a contact page. So if any questions, they can go there and just submit a question, and that'll come straight to me. And I also have a telephone number on there that they can call as well. Perfect.

Susan Friedmann [00:23:33]:

Well, we'll put those in the show notes, and I'll do a link directly to the book. So it's a good starting point for people as well. Is it available as an ebook?

Michael Goldrich [00:23:44]:

It is available as an ebook, As a paperback, and now, recently, as a hardcover, which I think only my mom has.

Susan Friedmann [00:23:54]:

Card covers are good when you're looking at corporate. I know authors often ask me, should they do hardcover, softcover? And if their audience is corporate, then definitely some hardcover copies. But softcover is also great. I love the ebook. I'm going online and picking that up as soon as we are done with this interview because I want to get right in and look at those prompts and those different ideas of how to do things.

This has been absolutely fascinating, Michael. As you know, I mean, we're running out of time. We'd probably have you back again because there's so much more that we could look at here. Leaving our listeners with a golden nugget, What would your golden nugget be, Michael?

Michael Goldrich [00:24:42]:

I think the golden nugget is and you were right, Susan, that it would come to me as I was going through the conversation here. I think It is just go ahead. You gotta try it. I think a lot of people are intimidated because they think it might be very technically challenging or they need to have some technical expertise, but the thing is it's just a different way of doing work. And just try, and you need to try and use it every day. Because the thing is, It can really help, and you'll never know how it can help you until you try it. And don't give up. That's what I'd say.

Susan Friedmann [00:25:21]:

I think those are wise words, really wise words because I think often people are intimidated by this. It's something so new and people are talking about it and there's so many different components and it can be very overwhelming. Just even I go in there and I'm looking at creating images and there are, you know, all these different platforms and Forms, and how do you go about writing the right prompt and everything. So that's a whole other conversation, but you're so right. Just go in there, try it, play with it. Yeah. You can't make a mistake. And if you do, you will rectify it.

Susan Friedmann [00:25:57]:

You'll just ask it more questions.

Michael Goldrich [00:26:00]:

It's actually better to make mistakes because that way you know what didn't work. It's a learning point. If it was always smooth sailing, then nobody's learning anything. It's really when you're challenged when you're learning.

Susan Friedmann [00:26:11]:

Yes. I so agree. You've been amazing, Michael. So thank you so much for sharing your wisdom. 

And listeners, if your book isn't selling the way you want it or expect it to, let's you and I jump on a quick call together to brainstorm ways to ramp up those sales, because you've invested a whole lot of time, money, and energy, and it's time you got the return you were hoping for. Go to BrainstormWithSusan.com to schedule your free call. In the meantime, I hope this powerful interview sparks some ideas you can use to sell more books.