Human-First AI MarketingScalable StorytellingStrategy with Avenue9s Mike Montague-AIMG AI Insiders

(00:31) Hey everyone, I'm David Berkowitz back with another edition of AI Insiders from AI Marketers Guild, part of the architecture family. And we're here with another special guest today, Mike Montague, Avenue 9. I mean anyone who puts human first at the top of their LinkedIn bio, big fan of mine.

(01:02) I actually mentioned, by the way, just to give them a plug, I was at an event yesterday Techconomy New York and I got to hang out with one of my idols, Douglas Rushkov, and I told him, I said, "Now your book, Team Human, is literally the most prominent book in my Zoom background." And I do this for a reason because it's more important than ever. And he's like, "Yeah, it's kind of had a renaissance." He's like, there are Jenzers who are now talking about my book and they'd never seen it before.

(01:27) So, anyway, highly recommend that. But I also recommend getting to spend some time with Mike Monagu. Mike, how's it going? Yeah, so great to be here. And those of you who don't know me or my show, I host the human first AI marketing podcast and David was a guest. It recently came out with his episode a couple of weeks ago.

(01:52) So you can go check out our conversation there if you want more of what happens today. But my philosophy in that human first take goes all the way back to LinkedIn. My first book was picked up by LinkedIn sales solutions and international sales training company called Sandler and it did 80,000 copies or something, talking about how sales people can use LinkedIn without being pushy marketers, spammers and just adding a thousand people and hope somebody buys from them.

(02:26) So I've been on this kick for a long time and when AI came out I said, "Oh, this is even bigger. This is the same thing, but now those LinkedIn spammy salespeople that I was talking to 15 years ago can do this at scale with AI agents to thousands of people. And this is going to be a mess if we don't take a human first approach.

(02:52) So, I went all in on this a little over a year ago and have been doing a deep dive really on how do we understand how to use AI in a way that we do the thinking first as humans, we put the leadership first, we put our employees first, and we put our customers first in our execution. And that's what I'm excited to talk about. Pretty cool. Well, I'm curious to hear more.

(03:15) So, just keep going on that. And also there's one caveat. Whenever I talk to a group of people, I like to let you know that I started my career as a karaoke DJ and club DJ and I entertained 500 drunk people four or five times a week for several years. So, you can't throw me off my game.

(03:39) If you have any questions, you can unmute, you can shout out, you can type it in the chat. We'll pay attention to that. And I know what I'm going to talk about today, but I don't have a rehearsed speech or present. And sometimes I forget to mention, which I should since not everyone joins every week, is that I always call these community conversations and not webinars.

(03:57) I usually turn off the sound of a webinar five minutes after I join and then just maybe politely leave it in the background. But I like conversations a little bit more and this is a crew where you will get some questions and thoughts on this.

(04:22) So I'm excited to hear your take on what scalable stories are and even frankly I'm like what's changed since talking with you a few weeks ago? First Earl wants no scrubs. No he is a Spice Girls fan. That's a good go-to karaoke song. I think that the scalable stories part here is important with AI. I've noticed what I call the concentrated orange juice theory that if you take a weak prompt, if you take a one-sentence prompt like write a blog post for me and enter, it's going to take that and water it down and try and stretch your one little micro idea into a full 800 word

(05:05) blog post and it just doesn't work that well. It's amazing at what it does. I mean, it's better than any human could do if you just said, "Write me a blog post about storytelling with AI." It'll write something, right? But it's not going to be that great, mostly because we're stretching and watering it down.

(05:31) Now, you can add more context to that and the more concentrated you make your prompt, the better it will do. So, if I give it a really meaty subject, like I want to write a storytelling blog post about AI tools for how to write screenplays and I believe these five things are true and these are the best AI tools. Write that blog post. Now, it has enough context to write something more interesting.

(06:00) But there's an even better way that I wanted to share with you today, which is AI is even better at the opposite, which is distilling watered down stories, conversations, information, and data from your business and distilling it into the concentrated orange juice that you can use for your marketing and make powerful stuff.

(06:26) So I think if there was a misconception that I would share today is that most people are using AI to water down their ideas and their marketing when what they could do is take all of their marketing and make it stronger with AI if you go the opposite direction. Does that make sense David? I think so and I feel it too as someone who's like Yeah.

(06:57) Can you give us an example how would you do that? Right. So what I do and these are I'll just share out the five steps and we can talk more about them. But number one is you have to start with the story first. Karan. So if I interviewed the company's founder, the top salesperson and your best client.

(07:16) Let's take those three in a 30 minute to hour long interview. And then I put those transcripts into AI and I said, "Tell me what this business is really about. What's the problem it solves? Who's it for? How does this company solve it better or differently than anybody else on the internet?" AI will give us a really good answer for our marketing plan with all of that context.

(07:42) But I can't just take those transcripts and post them out into the world. I need to do that processing. I need to boil it down to the concentrated stuff first. Then if I take that scalable story and add some additional context of how am I going to tell this story. So step number two would be to identify the structure of the story after it comes out of AI and I did the who's it for, what's it for? What are the three to five main points? What are the stories and examples I can pull from our business that make the point? Step three is what do I need to do to change

(08:20) it into today I'm having a community conversation with 20 plus people that's different than me going on an hour long podcast with one person or writing a blog post about this topic. So I need to add this context layer or the structure of what I want the output to be as step number three. And then step number four is I can repurpose that into all of the other things.

(08:51) So I think most of you know now and it's not very interesting to me anyway. Maybe you don't know that you can't stop at that one piece in marketing. Just because I created this awesome white paper with my five steps to scalable stories doesn't mean anybody's going to go see it. I also need to make the other formats.

(09:10) I can use Google notebook to create an audio podcast version of it. I can create a video presentation. I can create a gamma slideshow presentation. I can create social graphics. I can create LinkedIn posts to share and write people about it. I can create the landing page from the story.

(09:30) And all of a sudden, I've taken that structure of the story and multiplied it across many formats to get the complete marketing plan that I want. And then step number five is integrating all of these tools, automating the process, and getting feedback from how those went to improve the story for next time we do it. So those are the five steps.

(09:58) I think the last one is the hardest one, and that's the last mile stuff that I know David and others in AI talk about. I don't think we're quite there yet where this magically happens. There's a lot of human work at the end to review the outputs, move between tools, set up systems and processes and standard operating procedures for your employees on what to do with all this information that we collected.

(10:23) Can I share a trick I do? Yeah. In this context, I have custom prompts. As you said, context is really important. I may go to a competitor's page and in my prompt I define my own tool and then say look at this and tell me the differences between our products. So it's always building using context of what is already there.

(10:53) As you said, the more specific you are and then the next step would be to take all of the different research on different competitions or articles, feed them all into a rag. So now you have a much larger context and what can come out of it is way more useful. Yeah. Awesome. So if we look at this in depth, that's part of the story collection stuff that I do too: you can collect competitor stories, you can collect client stories, you can collect all the data and marketing that you've ever used before and put those into the database for a custom GPT and then start

(11:31) creating those instructions around what is the story, who's it for, what's it for. problem-aware stories are different than solution-aware stories and I'm not familiar with this group so I don't know how techy or marketing savvy everybody on the line is but there's a lot of context that we can add into this and I think that's the best part of AI: it can handle more data and so again people keep asking it to stretch rather than asking it to consolidate. So take this huge and massive amount of context and give me the most relevant parts. I do this with my

(12:12) sales pitches that if I was going to pitch David this morning, I would go into my chat GBT and say, "Here's my standard sales pitch. Tell me everything you know about David Burkowitz, his company, who he solves, and customize my pitch for him." So add even more context and make my content stronger for this sales pitch.

(12:32) What are the top five questions I need to ask David when I meet with him later and start getting more information from that company. Another trick I think is really useful is to use synthetic audiences to verify what you're saying. Just yesterday, one of the members of this group sent me a website pitch that she had loved.

(13:05) She asked me for feedback and I fed it into the synthetic audience and instead of me giving her feedback, I used 110 people to give her exactly what she needed. And you can do that with your own content and that makes whatever you're presenting a lot. Yeah, some of those are cool. Well, I have found some dangers in that. Marketing has been using personas for a long time.

(13:34) And the problem is similar to Chad GBT and generic AI: it creates a regression to the mean. You start getting average content because you're marketing to the average person in this role, not this specific person in the role. One of the magic things that you can do with AI is market to the person and not the persona. So, I love what you're doing.

(13:57) If that's all the data you have, absolutely. It's better than a generic pitch. But if you can get information on the person, I think that's more powerful. Earl has his hand up and I saw that you had a question there about measurement, too. Was that the question, Earl? Yeah, that's correct, Mike.

(14:13) Yeah, the first question was you mentioned earlier about how you've been using chat GPT or other AI tools to improve your stories. My background is in analytics and measurement. So I'm curious how you are measuring performance. Yeah. You could probably tell me some better options. I am not the best at measurement because I found in marketing and sales which is my background it's pretty obvious when stuff hits.

(14:37) It's not a small micro difference of like, oh, I got 1% more conversion rates on the story. When I create a winner of a campaign, it's usually eight to ten times the responses that I got on other stuff. So, most of mine is subjective.

(14:54) Is it better than I could have done myself? Is it better than what we currently have and keep some easy measurements on it? Looking then at three different layers of stuff which is the vanity metrics. Did it get more views, more engagements, more clickthroughs to the website? Looking at conversions, and then obviously the sophisticated one for sales and marketing is did it get me the ideal clients? Because I can create a viral campaign that gets millions of views and send a bunch of wrong people to my website and end up with clients that I hate working with who don't have enough money or I could get one really good click that pays me

(15:32) $100,000 and is the best client I've ever had. Right. So that's how I think about measurement. How about you? Yeah, I think you're absolutely right. From my experience and from what you just said I think that the content we provide whether it's a story or any publications we're producing on any platform depends on the purpose of the content we're sharing.

(15:58) Some are upper funnel on the brand story getting people interested and aware of the products and services we offer to the bottom funnel like you mentioned in terms of conversion rate. So I wouldn't say it's one-size-fits-all. I would recommend measuring any content based on the purpose of the content.

(16:17) Is it to build brand awareness? Is it to drive engagement with the brand? Or is it to drive conversions so you can get those sales leads so you can get those $100,000 clients that you're looking for? For me, it depends. I'm a consultant. That's my default answer. Exactly. Well, that's everything in marketing for sure.

(16:38) I'll tell you one of my favorites that I included in my LinkedIn book as well is: does it start conversations? Because there's some educational content that's nice to read. I'm glad they did that and maybe I'll give it a like or something, but I'm not creating a human-to-human conversation. So I think especially in sales and marketing these days and when there are so many AI bots and other things out there, we'll talk about other measurement things here in a second, but to finish that thought, human-to-human conversations is my key

(17:11) metric. Is it remarkable? Is somebody leaving a comment? Are they asking a question? Is it making them think? Is it making them reach out? That to me is the gold standard. And then the thing I was going to mention is whenever I'm doing a new campaign, I find that there are waves of impact and feedback responses. The first thing that finds your content is robots.

(17:38) They're out actively scrolling directories and robots. Right. The second thing are employees and competitors which don't help you either but they're actively paying attention to the industry. They want to see what their competitors are doing and so it still doesn't help. The third group is salespeople and people trying to contact you for stuff and if you put a form on the internet you're going to get somebody to offer you office cleaning and stuff like that. And then finally, if you

(18:11) stick with it long enough, you get out towards clients that you know and people that are already paying attention to what you do. Your inner circle people. And then third is that outer orbit of people who didn't know you at all, but they went to a search, they found a piece, they saw something that caught their attention.

(18:29) And it's hard to work through that slog. And it can be frustrating to not get stuck and be like, "Oh my gosh, I got a thousand views, but 900 of them came from robot website hits." That totally makes sense from what you described.

(18:49) In what you described, I would filter that as engagement in terms of actual human engagement with the content, whether likes, comments, posts, shares, etc. But that makes total sense. Thank you. Awesome. Anybody else have a question or dive in? comments on NBA. So, Mike, one thing I'm curious about because I'll show you what I just found on you is if you have any favorite tools when you mentioned researching your audience and sometimes specific people you're dealing with: are you using major LLMs or any specific ones? I just put you

(19:23) into one of my all-time favorite AI tools that I'm on multiple times a day. And you can see how accurate this is from happen stance. I'm obsessed with this one. But anyone wants to connect there, I love it. But it does a good job at least surfacing some of your content, things like that.

(19:56) and gives me some. When I just put Mike Monagu it wasn't that helpful but when I put Avenue 9 then it found the rest. There happen to be hundreds of Mike Monigu es in the world so I have a darn thing but we talked about this before David names can be tricky. Yeah I've got no pity for anyone else on this call at least so maybe not in this room I know Jim's got so we'll answer that one and then we'll Oh yeah Okay, I'll answer that one, then we'll get to Jim's question.

(20:23) Most of the time it depends on how hard people are to find. My favorite tool is one called Humantic. I talk about it all the time. It is one of those like CrystalKnows, if you've heard about that, where it can predict their personality profile based on publicly available information on them.

(20:42) It is awesome where I can import that data and enrich the contacts in my CRM like HubSpot and then I've changed campaigns based on their personality style. If you're a high D personality, I use the DISC one. D personality is somebody that's dominant, direct, they want big picture stuff, they're going to move fast, they make quick decisions.

(21:12) I want to write an email that fits all of those criteria. If they're high detail oriented people and they have a different personality, they want stats, they want figures, they want documented case studies, they want all of the research in the email.

(21:28) And when I changed the campaigns to the four different personality styles, we saw 300% increase in conversions to 3,000% depending on the personality profile. It was incredible. Super fun to play with. And that's where I really got into telling these stories to the person, not my story. I have to understand my story, but I have to tell it in the way the other person wants to hear it. if that makes sense.

(21:56) Yeah. Love that. Jim, I think that's a great level of personalization. But my question was: as you scale your storytelling through AI, where do you need to put transparency of labeling AI involvement—supported by AI if it's in the content side—and when you have transparency of advertising?

(22:38) I don't think we've landed on the level of expectations people have of human communication, AI enabled communication, AI generated communication. You follow the question. What are your thoughts on that? I have strong thoughts on this one that may or may not be mainstream for everybody. So, take what I have to say and make your own decisions. I think AI is a tool.

(23:05) Every image I make, I wasn't labeling that I used Photoshop or Adobe Illustrator to make the image. If it was my idea and I created something, I used a tool for it, I don't feel like I need to label that I also use Chad GBT to do it. But I'm partially dyslexic, so I've been using Spellcheck and Grammarly for years.

(23:27) I'm not going to credit Grammarly for editing my book that I wrote, but it was heavily used in writing both books. Now, I do think in larger pieces of content like a book or something, it's important if you're in compliance-based industries or you're pulling stats. You have to be really careful where those come from and finding the actual attributions to them.

(23:53) But I don't feel like for me there's any copyright issues. I'm creating original content and I'm going to edit it and shape the prompt enough that I'm crafting something with a tool. I'm not just copying somebody else's work. So I'm 99.9% sure that my stuff is original when I post it. But do you have a concern or a use case? Well, no.

(24:17) That seems reasonable to me, but maybe I would take it the other way around. Where would I expect something to be transparent? As I'm consuming this. So the other thing I was going to say is that I am not a fan of deep fakes or misleading.

(24:38) My human first philosophy is very much customer first too, which is if they would feel tricked if they found out the truth, then I don't think that's fair. So I'm not a fan of AI automated agents for sales or marketing. I know probably some people on this call use them and that's fine for me. I don't want somebody to go, "Hey, Jimmy sent me a message on LinkedIn. Can I talk to Jimmy?" And be like, "Oh, no.

(24:57) Jimmy's our AI robot and we didn't disclose that." And now I have a weird thing, right? So now you're getting there. I think there's some rule of thumb that I'm just not clear on and you're getting into that space. For me, I'm not going to pretend to be human.

(25:14) I'm not going to replace a human. If I'm using robots, I want it to be clear they're chatting with a chatbot and not a human or they're chatting with a human. I want them to know. In content creation or marketing and storytelling, I don't think it matters as long as I'm not trying to make a pretend story true like, "I worked with ABC clients and they got these results" and that didn't happen. I think those are more just ethical decisions for me.

(25:39) What do you think? I think people can. There was somebody who spoke here a couple weeks ago who put side by side videos created by humans versus the same list of key frames created by some Google tool and the telltale signs are it's too perfect or there are weird mistakes that are glaringly obvious.

(26:09) I think younger people are more attuned to that than most of the people on this call because they're growing up with it. It's not that disclosure is the issue so much as that it will become increasingly harder to trick people when something is created by AI versus a human. That's a hot take, Lisa.

(26:35) I think it's going to become easier to trick. No, I think because people's brains are developing too along with the AI. My kids who are teenagers can spot an AI-created thing on Instagram from 10,000 miles away. I usually can too, but they get it 100% of the time. I think that's where I come from too, especially in marketing: we can all spot a spam email just from the first three letters of the subject line, right? Humans are lazy.

(27:01) They write dumb subject lines that are one or two words long. If you see a full sentence in a subject line, it's either your boss who hasn't learned how to write that in the text area or it's a spam marketer trying to get your attention in the subject line. So yeah, I agree all that stuff is very. The weird side effect for me is I've realized especially for stuff I'm putting on LinkedIn lately and elsewhere, I'm editing it less.

(27:32) Because that first draft feels so much more raw and if that's how people know it's really me and that it's like this is the kind of stuff I don't even have AI looking over then I'll take it. I'd rather it be like—I mean I think about this so much with the book where it's like AI could have written a better book technically.

(27:57) I'm just hoping it's not a book more people would have wanted to read, right? Yeah. The other thing I think is interesting: did everybody see that last week Chat GBT fixed the M dash thing. So now if you tell it not to put—and Hanley is going to be so disappointed in her campaigning.

(28:16) I wrote her back and she replied to me laughing because I put the post. Yeah. And I sent it to Ann. I was like you won the war. So yeah, Mike, I've got two questions for you.

(28:48) The first is when we follow your five steps all the way through and we've got to the end there, but my question is really twofold.

(29:24) How has it fundamentally changed how you would approach the task in the past versus now from a value perspective, not efficiency? I know it's made you faster, it's made it easier to do more, but outside of efficiency, where is the value that it has created? Has it given you the ability to get—where's the value? It could be the value is in the additional variance you're able to create, maybe it's fundamentally changed how you

(29:24) do segmentation and then how you go to market in terms of the end message that you take to the end customer. Help me understand the value play that isn't just an improvement in overall efficiency. Great.

(29:47) Let's do that and then get your second question because this is awesome and I haven't quite thought through this. Personalization, like you said, is definitely one. The variance for me is really fun because I like to make stuff and tell stories. That part is fun. The other part is it's allowed me to make things I couldn't make before and I think we're going to see a lot more creativity going forward. I can't imagine the storytellers, the people in Hollywood that now can make anything.

(30:15) It's an amazing opportunity to make cooler stuff that wasn't possible. My example for that is in my playful humans book. I had interviewed over 250 experts in play, positive psychology, people that play for a living. I interviewed everybody from Justin Guini who lost the first American Idol to Kelly Clarkson to jugglers that were on the Tonight Show and magicians that fooled Penn and Teller. Wow. When I was writing the book, I was trying to figure out: can I as

(30:47) a human remember all 250 of these interviews and which one is most appropriate for the point I'm trying to make? Sometimes those things pop into your head and you go, "You know what? DC Glenn from Tag Team told me this great story about 'Whoomp! (There It Is).'" This will fit this chapter. Great.

(31:08) But now I can use AI and say, "Analyze all 250 transcripts and tell me the top five moments that you think might be most appropriate to the point I'm making." I think that dramatically increases the value of the book and the story I'm able to tell.

(31:27) So those are the front-end stuff and the back end is my first two answer. And then the question from an execution standpoint now: what does this mean? Take it back to advertising. In the past, you'd create three bits of creative. What now? Create 100 bits of creative.

(31:46) Is that what this is now? So that you can microtarget different audiences with different messages because you've gotten greater levels of understanding of what the end customer really wants? Fundamentally maybe changed your segmentation. Have we gotten there yet or is that what's next?

(32:10) Are we still working through the first phase of it now? I would say we're still working through it and 90% of marketers are not there yet, but the tools are there. There are tools that will change your website based on the person viewing it. And it's not one version or 100 versions.

(32:34) It's a custom version for each person who visits your website and it's rewriting the text on the page as they get information about who's doing it and what they're viewing. That technology exists and is available now. Same with advertising. There are AI tools and I've interviewed the people on my podcast and I can pull up some of these names. I have them bookmarked later, David.

(32:55) They will change the ads based on who's viewing it and adding additional context or rewriting stuff. And I think we're going to see you can now make your own version of The Simpsons and stuff like that—if you don't like this episode, I think we're going to see a Black Mirror Netflix episode where you're going to be in it or your city and if you and I watch the same episode, we'll see different things because everything is a completely custom piece of content. That's not right for all

(33:25) solutions and all budgets. For me the interesting point and to go back to your first question: it makes strategy a lot more fun because now I must decide what's a human job, what's an AI job, how could humans and AI make something significantly different, and what's my marketing strategy. In a world where I could make anything, what should I make for whom and to what purpose becomes a much more interesting question than basically up until two years ago, our question in marketing was what can we afford to create, and what's the

(34:05) coolest way we can make this within our budget and time allotted. Suddenly those two things are eliminated. That changes the game for me. One of the things you mentioned highly, Mike, was using context for prompting for your story creation.

(34:22) I've seen everybody's an AI expert on LinkedIn apparently these days. Do you have any frameworks for your prompt engineering that include context or do you have your own approach based on your research and experience? The short answer is probably not.

(34:48) In my experience and marketing expertise, I think I ask better questions than most people. I don't have prompt frameworks. I have a few things saved: when I'm editing a podcast I know I'll need titles for the podcast, a podcast summary, a YouTube description,

(35:07) a blog post and tagged keywords, social media posts, ideas for video shorts. I have some of those things saved and outlined so I can walk through that kind of standard operating procedure. But I wouldn't say they're revolutionary. I generally just ask for what I want.

(35:31) I think about it like if I had a really smart intern, if I just got a 140 IQ intern out of Harvard to be my marketing assistant, how would I explain this task to them? What would I need to do? Then I look and the other thing I probably do differently is argue with AI.

(35:53) The story I always gave was the M dashes where I put in the custom instructions, don't use M dashes. It puts them in anyway. So I say again in the prompt, can you rewrite that without M dashes? It still had one in there. And I said, just curious, is that an M dash between this word and that word? And it says, you're right, M dash.

(36:14) I'll remove those going forward. And I said, did you just troll me with an M dash in your response? Yes. And it says, that's correct, M dash again. I swear I'm not letting this go. But I do that with marketing materials too. So, my prompts are probably more multileveled and I make them smaller and iterate more and I argue until I get what's right and I subtract context if it starts to drift or add context if it starts to water down. Sure. I'm looking forward to checking out your book and whatnot, but I think I

(36:50) have a very similar approach. I spoke earlier this summer regarding my approach to AI and I think of it not as a tool; for me it's the team. I'm training my team what they need as if they're an intern I need to coach on how to produce what I'm asking for. Same idea. I'm with you 100%: we build them, teach them, train them, and we get the results we're looking for; until we do that, they're going to be as dumb as an

(37:20) intern. Agree. Onboarding is huge for any AI tools, assets. Chad, Yep. Mike, I love the example you gave about organizing and extracting the most important information from a lot of content. My experience has been that the more I provide an AI with a lot of information to work with, the more complex the task, the more likely it is that the AI makes a mistake, and as a writer journalist book author you have to fact

(38:06) check. I'm sorry that isn't one of your five steps. I just did something. I wrote a LinkedIn post yesterday and I had AI do the research, said please cite your sources, and then I checked each source and some of them were wrong. Would that be your sixth step or what do you think? Yeah, I kind of put that as part of five, but I glossed over it.

(38:31) In the kind of B2B marketing that I do, it doesn't come up that often. I was on a podcast yesterday for financial advisors and they have legal compliance and stuff and if you're quoting numbers or company stock probably definitely check. I found numbers it messes up more than anything else.

(38:51) It's really hard to get context around numbers. So I check stats and things like that but the biggest part of my process is the human review. I never copy and paste. I never let it publish without a human intervention. I always joke that my philosophy is the human first philosophy, but it's also human last and human in the middle as well. It's me too. It's got to be reviewed.

(39:20) It's very iterative. I worked as an editor for a long time and that set of skills is very helpful when asking it to produce something. The metaphor I use is Tom Cruise in Minority Report. You're doing all these things and there is a human actively engaged in it and that's the person you lose your job to, not an AI.

(39:47) But Adam, do you think there will be pre-crime units within our lifetime? Use predictive analytics to know. Mike, hold on a sec. I've got my precog going here. And it does broach an awkward subject about your cousin Louie and your— Yeah. This is what I'm saying. You don't need to precog like some mythical superhero.

(40:07) AI is now predicting crime like the corner yard patio door. My friend, we got 9 minutes and 30 seconds. It's not fiction, it is happening right now. Organizations like Tegna in the local TV business are doing exactly what Mike spoke about in terms of real time monitoring. They have access to all the police callouts.

(40:46) They have access to the logs regarding whether the police car got there and left 2 minutes later. If they left 2 minutes later then it's a false alarm. All of those signals are brought in and then used to construct stories for journalists and combined with additional context like information from the local hospital. All of this is happening right now. But there's nothing surprising about that. We've had insurance people using actuarial tables for this

(41:24) hundred years now over the last decade plus. Exactly. None of this is new. We make it sound very scary, but it's just predictive modeling at the end of the day. What is new is insurance companies are not necessarily using it in real time to inform us and we're going into an age where it's 100% about storytelling and doing that in the most expedient way possible.

(41:56) I disagree with that, too. That's been happening since CNN went to the 24-hour news network. You have Chartbeat doing real time analytics since 2004. I think what's different—and I hope you'd both agree—is the way we can account for the usage of that data has dramatically changed.

(42:22) I'm being very self-serving by saying this, but I was cofounder of a company where we were commercializing Tim Mers le's research about accountability systems and we actually published a paper where we were monitoring Palunteer's classification of gang members and stuff. The problem with Snowden wasn't just the ramifications but the FISA court that wanted to review NSA compliance folks' issues. The judge's issue was you cannot account for the algorithmic and systematic inferences you're drawing out of it. So okay, all the stuff you're talking about: we don't have any way to account for the use of the data at scale and view the black box.

(43:17) The black box trail—if we don't know how their algorithm is being coded or what prompts they're using—the black box problem. It's going to get interesting, David. A lot of exciting developments are happening quickly. The unlock is substantial. I'm an optimist though. I'm team Robocop.

(43:47) That's how I picture you, Earl. Exactly. Dead or alive, you're coming with me—with AI—with an M dash. Jim, did you have something? I might have gotten lost in the crossfire there. That's okay. There was a lot of drift. I was going to come back to something Lisa said about verifying everything. I'm selling this company and I got this—an LLM delivered this raving review that, in context, was one company talking about another company and a company is going to buy this company. I check the source and

(44:26) it's an engineer at the company I'm selling. It was deep on Reddit and ChatGPT had transformed it into this real positive review, the best on the net. It was complete. I had to take it out. That's prompt engineering, meaning that you need to be explicit with whatever AI tool you're using:

(44:52) Do not create inferences or hallucinations. You have to train them like an AI. Otherwise, it's going to try the most thing. It's on us to be better managers and editors in my opinion. You can't blame the tech. That's lazy. Well, that was what was delivered to me and I had to edit it out.

(45:09) I know. Jim, you reminded me of one more golden nugget before we wrap because you triggered something important. In these storytelling things there is bias—it wants to give us what we want and we all know that.

(45:40) One of the things I did when I published my Playful Humans book last year, I uploaded the whole thing and before I put it up on Amazon, I said, "If I get a bad review, ChatGPT, what is it going to say?" I want you to roast my book for me because I have happy ears.

(46:00) I'm excited about this thing. I'm going to launch it. And the moment I get a first bad review, I know it's going to hurt my self-esteem. It started typing and I had to walk away. I stepped back and thought, I need to take a deep breath. Am I prepared to read what it just spit out? Not enough people are steel-manning their own arguments or red-teaming stuff. I put it through: "Rip this apart.

(46:25) Tell me what it is." When I came back and read it, I thought, "This is great. Anybody that said this didn't get the point of the book. It wasn't for them. That's not what I was going for. So, I'm fine." I hit publish. Having it break it apart, make it better, "Tell me why this deal sucks" is a great question

(46:48) if you're going to enter the contract. Have you ever explained to anyone the meaning of your company name, number nine—transformation and preparation of new beginnings? My daughter was telling me about 111. 11:11, the angel number. Okay.

(47:13) And Paris Hilton named her company that, but the Avenue 9—want to school us on that. Ninth letter of the alphabet, AI built in from start to finish. What does it mean? You found it. I was trying to come up with a company name that would be generic that I could pivot in AI enough that it would leave some open avenues for me.

(47:32) I was trying to find something with AI in it and avenue starts with A and I is the ninth letter of the alphabet and then I looked in numerology at the number nine and it stands for humanity and it's the end of an era or end of a sequence and the start of a new beginning cycle. Very futuristic and I was like this is cool.

(47:57) Also I feel like people have a plan A and sometimes a plan B but nobody has an avenue nine to get to success. So this is—we all have one. Numbers versus letters. In ancient wisdom they say the eighth reveals the ninth. The ninth is the ritual realm. Yeah. So, I thought that was fun. It's been really good for me. Also, I got an eight character domain name for a .com, so that was a win as well. That's impressive.

(48:26) That's fun. Okay. Thanks for calling on me, David. Of course. Always bringing something good. And Mike, you know, the challenging—I've mentioned my other favorite AI app now is Rosebud, this journaling app. What you didn't tell me.

(48:55) It's a journaling, a bit of self-help. It analyzes your thoughts a lot. And when I push it to go deeper and challenge me, then it gets really fun. On its first level, it's got kid gloves.

(49:16) But you also—it has to be the kind of thing where there are some things where I probably don't want to hear a counterargument or something I've written where I don't want it to be roasted and I'd rather not. But maybe something before it goes out to a client or things like that.

(49:44) I just want to do the best work and I want to anticipate their arguments. So Mike before we wrap: what are best ways for folks to stay in touch with you? I shared a podcast link earlier and your LinkedIn but how do we keep these conversations going because there have been a lot of them today all in this one chat. It's been a fun chat.

(50:06) I appreciate you inviting me on and hopefully I can come back to a future one. I am an internet marketer, so it's pretty easy to find. Type Mike Monagu or Evan9 into your favorite AI and see how I'm doing with AEO. Mike Monu, LinkedIn is my preferred platform. The website and everything is Avenue 9, but the podcast is really the fun one for me.

(50:30) The YouTube channel just got over 10,000 subscribers last week, so that was a big milestone. I've been getting killer guests like David Burkowitz and other AI experts from around the world. It's been cool and I would recommend checking that out if you're into learning more. And the dynamic content personalization tool or tools you mentioned.

(50:56) Yeah, I was looking that up as well. The ad one I'm going to put in the chat as we wrap up here. The ad one I know I can find. While you do that, Mike, I was going to say thanks for everything you've shared today. It's super helpful. I'm not a marketer.

(51:21) I'm a product person, so hearing these developments is really insightful. Thanks. We also like a range of voices here. And it's fun where we've had this week and last with Colin Jevans from Nomics where it's open-ended conversation along some themes and it's fun.

(51:49) Not everyone needs to do the PowerPoint overload—just have some real conversations. Next week we will have a fun little vibe coding competition if anyone's around the day before Thanksgiving. I'll have an assignment ready and whether it's something you've done before or haven't, I'll send out a couple tips on a platform or two to register for.

(52:19) If you haven't—most people in the room have probably tried something you've liked—you can bring your own, but we'll— we had this open slot and have been meaning to do this for a while.

(52:37) And Mike, if there are any other links you want to share after that pay you later, just let me know. I'll share them with the community. You're always welcome to do so in our Slack. Thank you. I seriously looked at the time 20 minutes ago and thought, wait, is it 12:40 already? Are we getting in the home stretch? Because this one flew by. Thank you. And thank you all.

(53:00) I love getting to sit back and see where the sparks go. Have a great rest of your week. If I miss some of you next week—and I understand if I do—have an amazing Thanksgiving. See you in the Slack and everywhere else. Mike, more to come. I look forward to learning from you. Thanks, brother. Thanks, Mike. Thanks, David.

(53:20) Thanks, Mike. I have questions. Oh, cheers. Thank you. Go for it, Earl. Ask your question. No, I don't. I'm very shy, Jim. I'm obviously very shy. You're working on it. You're the opposite shy. I reserved. I reserved. Go get some lunch. Look forward to the follow up. Take care, David. See you next week.

(53:54) All right. Fantastic. Thanks, everyone.