Welcome to Tony Martignetti Nonprofit Radio, big nonprofit ideas for the other 95%. I’m your aptly named host, and I’m the podfather of your favorite hebdominal podcast. Oh, I’m glad you’re with us. I’d suffer the embarrassment of Onicotrophia if you nailed me with the idea that you missed this week’s show. Here’s our associate producer Kate, to give you the highlights. Hey Tony, here’s what’s up. Your AI brand footprint. What is this thing? Why should you care? And what can you do to improve it? George Weiner returns to acquaint you with his company’s study of how artificial intelligence will influence giving in Q4. Then, he explains the implications of the research, including that last year’s content strategy is obsolete. He also brings tactics for you and your content to get the recognition you deserve from Google Gemini, Chat GBT and their colleagues. George is chief whaler at Whole Whale. On Tony’s take 2. Hails from the gym If she can do it. Here is your AI brand footprint. It’s a pleasure to welcome back George Weiner. In 2010, he founded Whole Whale, a top 100 nonprofit focused digital agency supporting analytics, advertising, AI capacity, and digital fundraising. George is chief whaler. He’s also the co-founder of Power Poetry, the largest teen poetry platform in the US, a safe, creative, free home to over 1 million poets, and CTOs for good. A group of tech leaders at nonprofits that delivers social impact primarily through technology and digital strategy. You’ll find whole whale at wholewhale.com. You’ll find George on LinkedIn, where he is very active. Welcome back to nonprofit Radio, George Weiner. I’m, uh, I feel like we haven’t learned a lesson. You keep having me on. I’m honored every time I get the invite. I was like, wow, I didn’t mess this up. Thank you. Yes, no, you, you’re, you’re, you’ve earned a repeat, repeat appearances, absolutely. Um, you know, I, so I was happy to, uh, read the bio that you provided, but I don’t, I don’t think it captures. I don’t, I don’t, it’s not the bio that I would write if I were you, because, you know, you have this enormous tech background that you do go on, the bio does go on, which I did not mention that you were chief technology officer, I believe it was for 7 years at dosomething.org, which is enormous, enormous, turned into the enormous data capturing. And uh uh assistance for activating young folks, that’s enormous, but I would, so that they’re not that that that is to be minimized, but I still don’t feel like this all captures. I mean, you’re, you’re the, you’re a tech guy who understands it, explains it, uh, simply talk about like, I mean, you, you get in the weeds of tech. I mean, you’re like a coder. You write, you write lines of code. I do, I have, I’ve come in and out of it, interestingly, I used to be very actually like in the technical writing code and then I hired people smarter than me to write much better code and then I came in and out a lot of data, analytics, advertising. I love learning, I love understanding and then. Helping nonprofits find the angle, right? Like how do we leverage this thing? I’ve I’ve studied the whole book, just read page 17 and do it this way is what I love kind of getting at and recently I’m very heavy into coding again, but frankly with AI assistance um and and building up calls writer.AI, which is a great customized platform for nonprofits creating, um, creating AI generated content and I do it in as safe a way as possible. Exactly. Uh, yeah, so that’s kind of what I wanted to, uh, I’m glad that you, you mentioned. I wanted to get to a little bit, Cowriter.AI, a proprietary. Whole well created. Is that right? right? Yes, I, yeah, that, that guy, that’s the coding that I was referring to. Maybe you hired smarter coders to do better at it than than your initial cuts, but, uh, which. Nonprofits can use and train. Cautiously on their own, have the, have the model trained cautiously using their own content. Exactly, yeah, for you, not on you in the sense that we build it so that there is a central source of truth that is stored and protected for you that can then be pointed at any model and then custom built prompts and guidance based on what your team needs to do today, and we delete every chat every quarter and pay for carbon offsets on all of the queries uh that are conducted to try to at least approach carbon neutral as. Uh, difficult as that calculation is. Well, I admire the attempt. Um, you’re also a whole well certified B Corp, so you’re, you’re, you’re committed not only to the environment, but also to the, uh, to the causes of, of, of social impact. We’re trying. I think I’m really excited though about the upside of AI while still like having my little Tony voice in the side of my ear being like, Well, but what about, what about like this, it’s gonna steal the content here, it’s gonna cause this, and I feel like this um AI study that caught your eye sort of like really walks down this tiny tightrope of I’m excited, but also I, there is caution to be had for some of the insights we found. OK, we’ll, we’ll get to some of those, but I’m glad that Tony voice, um, accompanies, I’m not gonna say haunts you, accompanies you. The Tony voice accompanies you because I do have my concerns which I’ve, I’ve shared with Beth Kantor and Amy Sample Ward and then you, um, so listeners are acquainted with my. Uh, skepticism, concerns about the, uh, the, the widespread adoption of artificial intelligence and, uh, large language models, so. I’m glad my voice is accompanying you. Yeah, and I like, I like the, the, it’s like respectful challenging too when you come at it. I think there’s some folks that shut down arguments and be like, you’re wrong, I’m right, you’re like, you have this very like clever way of getting in my head on, you know, the edge cases that come up and I’m like, damn it, how am I gonna answer this for uh for for this type of argument and it’s important. OK, thank you. I’m glad. right. Um, the AI brand footprint. Uh, you study? Because uh you’re concerned that uh AI is gonna have an enormous influence over giving decisions in the 4th quarter that we are now in, right? AI. AI’s influence on giving. See, my questions are so you can’t even, where’s the question? The guy talked so long. I, I, where’s the kernel of the question? Yeah, you’re concerned about the, the artificial intelligence’s influence on our millions of giving decisions this, this, these 3 months. It is going to be an unprecedented influence for AI on philanthropy, and that is maybe a hyperbolic way, an excitable way of saying that AI continues to grow in its answering of questions we have. And as that happens, one of the questions that comes up in Q4 is what are the best animal nonprofits I should give to? What are the best ways to support mental health and youth for charities I should donate to? Where should I give for the most effective cancer fundraising efficacy fill in the blank of your cause? Those questions are going to be answered. In the millions of numbers controlling billions of dollars. I thought it was worthwhile to take a quick look at. The data behind that. OK, so you looked at 6 different Large language models. I’ll let you name them. I have them listed here. Oh you want me to name them? No, no, no, I can go through it. It’s like a lot of technical, um, but I know there are some, some geeks among us and so here was our methodology. Um, we looked at 12 different cause areas and then for each of those areas, we chose 10 sort of iterations on the types of questions. A potential donor might make alongside those. OK, the next step is where should we test this? Should you just choose one model in the corner of the room? No, anything worth doing is worth overdoing, as I like to say. So we looked at 6 major models that kind of comprise the landscape as we see it, Gemini, OpenAI, and Tropic, meta Grok just in case, and um, you know. From there and also sorry, perplexity and from there we then literally send requests to all of them multiple times, get their responses back, and then analyze them. There’s a number of reasons for doing this, but on one level, we also have to understand this is a probability. It isn’t like Google search rank where I can definitively say you’re #3, you’re number 1 like. That’s not how this works. It’s like rolling dice each time, but when you roll dice enough times, you get that little nice little bump, the Gaussian distribution of a lot of things here and then out toward the sides less and less. I wanted to understand a bit more about the distribution of nonprofits being recommended in each of those areas across those models. How many flags of jargon am I gonna get thrown at me right now? No, that’s OK. No, you’re all right. Um, I do want to clarify, Gemini is the Google, uh, the Google product, uh, just to make that clear for folks, uh, and Anthropic, uh, is named Claude. I don’t know if folks might know that it’s Claude or Anthropic. OK, the other ones I think people are familiar with X is Grok, etc. um. OK, no, no, you’re OK, jargon jail. I’m, I’m listening. Uh, I’m, I just don’t wanna go to jail too. I’ll let you know. I’ll let you know. All right, so that’s a lot of questions because you 12 cause areas times 10 questions, different iterations of questions times 6. So 12 times 10 times 6 isn’t that something like 720 or so? Different over 700, but more importantly, like we’re talking about millions of words being analyzed by the end of the day because each of those prompts comes back with a bunch of texts that we then have to parse. Uh, and then from there what we did is we have to figure out like, all right, you know, what is the there there in this prompt and we began to go by counting the number of mentions of unique nonprofits as well as sources of influence. What does that mean? It means a lot because who did the AI point to as, who according to whom is this charity? Verified quality of donation. This is how we get to our first action step because there are some very clear, clearly prominently mentioned influencers that because see this is the value, uh, it’s more than, you know what, we’re not even gonna talk about who was the top named mental health cha or or the most common named animal welfare charity. We’re not gonna do that because I don’t think that’s where the value lies. The value lies in recognizing. Now see, I’m trainable, George. I’m trainable. Uh 63, but I’m still trainable. The, I think the value in this study is, is the, the threshold recognition that artificial intelligence is Capturing you, it cares about your work and, and what does, how does it learn about your work? What does it learn about your work? What does it say about your work when people ask either specifically about your charity or generally in your cause area or any other, any other reason you might show up in a, in a, in a search result? AI cares about. AI is paying attention to, and I don’t know, cares about is maybe a little overstatement, but AI is paying attention to you, to your work. OK, so let’s go to our, uh, our, our first, our first real, real takeaway, the influencers. I’ll let you name name the most common influencers that nonprofits have got to be, if they wanna be, if they wanna be thought of well by by these large language models, they need to be. Uh, um, approved by or have high standing with, who were these influencers? Yeah, I, I think this is an interesting way to answer this question is actually to give you a section of one of the responses from one of the prompts from one of the models, and I just want you to consider the implications and the prompt I gave this model, and this was a Gemini model was where to give for animal organizations like it’s kind of disfluent and sort of how I put it together, but where do I give is the central question I asked. And there’s one prompt that came back. It said before donating, always do a little research. Charity watchdog sites, these sites evaluate charities and efficiency and program spending. Charity Navigator, GuideStar, now Canada and BBB Wise Giving. Check their websites for their mission statements and how they use donations, local reputations, and finally, financial efficiency. A good rule of thumb is at least 70% to 80% of their budget should should go directly to program, not administrative costs or fundraising. You’re right. That’s instructive. All right, so that, that’s a pretty, that’s a, that’s quite a good answer. I like that answer. Yeah, I think unpacking it sometimes though is that what you’re hearing are, it is not going to primary source, it is going to evaluation platforms, your charity navigator, your guide star, your BBBY Giving Alliance, those public profiles matter more than ever. They mattered before, but they matter in your mind. You’re like, oh, you know, like a donor’s gonna like do some research and go check it out there, like, no, no, no, no, no, this is being baked in. Up front, before someone even finds you, they’re finding what other people think of you on these sites and others. It’s different. Huge takeaway, huge takeaway. um Give Well was another one that was named another influencer, yeah, no, but you were, you were quoting from 11 prompt out of that one response out of 1700. So, um, OK, Charity Navigator, uh, guide star, which is now Candid, Better Business Bureau, Wise Giving Alliance, Give well. Take away number one, you, you’ve, you’ve got to be thought of well, you’ve got to be well ranked. Not, not quite for the reasons. I mean, the, so, I mean, George, the reasons that we thought they were important still exist. There are still people who, uh, my dad before he died used to get the Better Business Bureau, Wise Giving Alliance printed guide. And he would check, check to see if uh a charity that sent him mail was listed in the printed guide. So I’m not sure people are using the printed guide too often anymore, but They, they are looking, the people who do go to your site are looking for those little, those little badges, the, uh, the, the high ranking badge, the platinum for Charity Navigator, etc. But now, even more so. The large language models are using the influencers, like you said, George, the 2nd order. Recommendation or, or, or evaluation sites, not recommendation, evaluation sites. In their, in their, in their responses. Yeah, and those types of giving guides, third party validation, there’s a lot to unpack there, but it’s not just, hey, let’s update our website, it’s go there. Another sort of nuance here were um mentions of GoFundMe. I was kind of curious when I was looking through the data of like, alright, how much of this is gonna be like. Find individuals and go that route and what we found was based on the mentions of GoFundMe and the request of like how do I help let’s say um youth mental health or poverty issues, right? Where should I go, how should I help? Like it’s an open ended question on purpose. We, we did want organizations we wanted what is it and how is it advising and for GoFundMe mentions actually, uh, 70% of those mentions came from Grok. So in the land of Grok, which is Twitter, which is X, which is Elon Musk, just to get all of the bingo cards there. It is disproportionately recommending to go to GoFundMe. It also recommended charities, mind you, but I found that interesting. Part of that is in the training set and part of that is in the sort of, you know, we do our own homework but to a bizarre degree in there. So, so Grok was looking at the, the frequency of GoFundMe campaigns for charities as a way of determining whether they were a good place to give. They first off, are meant we are meant, we are analyzing its responses across all of those causes, right? Every single cause, every single chance. So across all of those tested areas, we were seeing that it just surfaced that the user should go find and look. For someone in that cause area to donate to on GoFundMe as a point of helping that particular cause. Why that happens, you can speculate, but actually, you know, it’s training data set. There’s a lot of people that go on to X saying, hey, I need money for this, go, you know, fund me and GoFundMe, and that type of link and that type of cause might be overweighted, plus potentially an underlying, um, frankly baked in mistrust of institutional organizations. That’s a bias. Grok Grok had a, had a bias for GoFundMe campaigns. They all have a bias. Right, let’s talk about some of the biases, yeah, because, because this gets to what The large language models think of your nonprofit in the, in, in some of these biases, and then, of course, we’re gonna talk about what to do, how can you Enhance your AI brand footprint. In, in light of what we’re learning from the survey. So let’s talk about some of these biases like size, the, the, the ones that I, the ones that I saw, the, the, the results, these were. I don’t know, maybe not. 100%, but these were very large charities. They had large digital footprints. Yeah, in in most of the cases where we then do a sort of top 10 breakdown, we have a like a herding effect where it is the the larger, more well-knowns, uh, are at the top, you know, you look at the environment we end up with, uh, just using that as one example, the environmental mentions by model like on Gemini, the top two are environmental defense fund and the Nature Conservancy. And then on Anthropic, which is Claude, World Wildlife Fund comes up first, and the Nature Conservancy, Nature Conservancy wins on Open AI uh on Grok, World Wildlife Fund wins on Perplexity, the Nature Conservancy wins. So you see, it’s like the sort of jockeying, but those are massive organizations and you have to go down pretty far to you get to something like an Earth justice or Oceana or um. I’m trying to find like Arbor Bay Foundation, which is just not small, uh, defenders of wildlife, you know, coming in at the tail. Those aren’t the household names. Nature Nature Conservancy, World Wildlife, or well, yeah, um. All right, what other, so size, so, so our listeners here are in small and mid-size nonprofits. They are likely not in any of the environmental causes you just named specifically, and they are very likely not in any of the biggest names in any of the 12 cause areas that that you evaluated, but. I hasten to add there are things you can do. We’re gonna get there we’re gonna get there it’s coming. The fear not, as Grandpa Martin Eti used to say. Nunjawari In his, in his New Jersey Italian accent, Nunjawari, there are things you can do. It’s time for Tony’s take too. Thank you, Kate. I have a new tales from the gym. There’s a woman who’s been coming to the last. 3 classes, uh, that I take every Tuesday morning. It’s the only class I take each week, just, I just go to this one class. And she’s, uh, she’s quite active, she does, it’s an aerobics class, she’s does all the weights, she does most of the moves, you know, like we’re stepping back and forth or side to side, things like that. She’s, um, she moves pretty well. I, I, I have my eyes on her because uh she stands right in front of me all three times. She’s been right in front of me in class. And the thing is, you know, she, she’s got all this activity for the hour long class. But she’s on supplemental oxygen. She’s got a tank down next to her and the tube and the cannula in her nose, the whole class. So there’s some things she can’t do like she can’t step too far forward or back or side to side because the tube isn’t that long, but she moves enough and she keeps up. You know, like, yeah, I can’t help but see her. You know that just makes me think, I mean, if the woman with supplemental oxygen. Stays energetic through this hour-long aerobics class. Anybody, almost anybody could be working out. There, of course, there are people that are. More compromised than just supplemental oxygen, but. You know, anybody who’s not on, on oxygen, uh, we all could be working out to some degree if if this woman can do it. So she’s, she’s, um, pretty amazing, pretty, uh, uplifting and encouraging. That’s another tale from the gym. That stories take two. Kate I don’t know if you follow like gym, TikTok or anything on like Facebook or Instagram, but now I’m seeing a lot of fitness instructors coming up with different um workouts that are seated for people with maybe mobility issues or any sitting down um disabilities. Um, but that’s great that she can get up and do it and enjoy it and still be active with something that’s probably heavy, you know, to carry around. Yeah, well, the tank is, but it lays on the floor. Yeah, I, I haven’t seen any of that on TikTok or Facebook or online, but I, I have seen chair yoga. Which is, that’s for older folks who do have mobility issues. Balance, you know, balance could be an issue. Uh, there is chair yoga out there. We might, I don’t know if we do a chair yoga class in my little beach town, Emerald Isle, but I, I’ve seen chair yoga. But yeah, the woman is, um, is greatly uplifting. Yeah, she’s uh. She’s moving. It’s awesome. We’ve got Boku butt loads more time. Here’s the rest of your AI brand footprint with George Weiner. Other biases, what, what other biases did you find in our uh Our large language model friends. Our large language model friends, this kind of surprised me. I don’t know if you can classify it as a bias, but it kind of getting back to the herding effect, but in the number of charities mentioned, like the unique number of charities mentioned really surprised me between the cause areas that we go between. So something like the environment or poverty, there was a range like unique nonprofits mentioned across all of our sampling of 127. OK, I then go to something like cancer. And there’s a 30 organization spread that is massive. I was shocked. I was like, shouldn’t they all just probabilistically find a large number of things? There’s extreme herding in certain cause areas and action step that you might use here is actually do this type of research for your own backyard. Abstract, pull back and say what kind of question might somebody who’s caring about my cause area and my locality put in to find the most effective, best run, greatest places to give for X, and see who shows up, but remember you want to do this on a like incognito, not influenced by your own GPT if you are paying for it, I’ve trained it, you want to actually have it from a cold start. That isn’t biased by your bias and your information because then you’ll be like, obviously you’re the best mirror mirror on the wall, like, hold on. So does that mean you shouldn’t even do it from your your nonprofit, your your office browser? Is anything that’s passing in information to it, like geographic is fine, but I would say any other things that are passing information are essentially tainting it and tilting it toward something that’s more relevant to you when in fact, You want something that is better reflective of the underlying biases and approaches of that model to sort of explore it. We actually use direct to the API calls so we know exactly what information we gave it and exactly what we got back and could kind of like clean out the clutter, so to speak, of any customization going on or cache information or browser um influencer. Here’s a tip. Actually, no, here’s a very, very hard tip, and you’re like, I don’t know where to start. I want you to go to the site open router.AI. And that will let you just mess with whatever model you want and see what happens with you getting full control over the data that are being sent. Router. What does this do? Just look for router. Yeah, yeah, yeah. OK. All right, so that’s a, are you saying that’s like a safe place that that’s it’s less likely to be tainted. It will let you test different models side by side in a way that is safe in this land and gives you control over more of the variables and it’s really like kind of elegant for side by side comparisons. OK, OK, very good. All right. Um, All right, so Let’s move to, um, you know, what, what small and mid-size shops can do. Now, part of what they can do is what you are doing at whole well with AI brand footprint. Which I am helping you to do. I am propagating this for you. So explain what I, go ahead, you flesh out what I just said that I know you’re doing that we are helping you with. Uh, by the way, thank you for reinforcing AI brand footprint. Why we keep using this term is because it is a concept that we created at Wholeal to explain what’s going on when you’re discussing this ecosystem of information that AI is talking about and representing your brand on. You’re not getting a lot of data about it. But you know it’s happening. You’re starting to hear maybe somebody in development department be like, you know, they heard about us from, uh, they said from chat and it’s like, oh, that’s interesting. There’s a whole ecosystem out there and we’re just sort of, uh, scratching the surface of it with this study, but you should begin to care about what that footprint looks like, what’s influencing it, how big it is, if it is growing, or if it is shrinking. Why I am focusing on and why I love the fact that we’ve mentioned it like 17 times now is because we are trying to put this concept out there, imbue it with meaning and connect it to us so that when, not if, and it already happened. That Google overviews, that’s the little AI answering when we do Google searches, talks about this concept. It attributes it to us, it surrounds it with our lengths and our language rather than literally stealing it and just throwing it into the soup that is the overall. Uh, you know, word salad of the AI systems. So to pull back, how do you imbue with meaning, own concepts and encourage attribution? There are some number of tactics, but that is the game we are currently playing and laughing about as our inside joke here. Exactly. So we’re not gonna leave. Nonprofit radio listeners, you know, wondering what are some of these tactics? How can you Create something around your work that’s unique, will be attributed to you when, as you said, when not if the, the, the large language models find it. How do you How do you identify all this, bring this all back to you and your site, your work? Because you’re doing this is exactly what whole whale is doing with AI brand footprint. There I said it again for you. It’s gonna be in the, it’s gonna be in the transcript. It’s gonna probably be in the show note. This what, what, what tactics can we all learn from what you’re doing at whole well? So the content that we used to write for AI engine, for for SEO search engine optimization is dying. The idea that all I have to do is answer a question accurately and I’ll get credit is dead, because also if you realize that the AI could answer the same FAQ question, generic, hey, how many. Uh, how many ounces are in this amount of thing like that information has been commoditized and will no longer bring you traffic. Your 10 facts about this issue is not going to work anymore. What will work are first party data. What kind of information can you bring to bear on this topic? Have you surveyed your audience? Another way to think about it is according to whom? Is there a testimonial, a statement that can be attributed to the CEO, to the founder, to the stakeholder who received the service, because when AI comes and summarizes and takes that content, it actually is sensitive to uh trademarks, first party data attributes. and saying, oh, according to the local animal shelter or youth center leader that this is the thing you’re like, oh, that’s tied and anchored now to something that the AI will respect. I think there’s a lot to unpack in there, but hopefully you begin to see the nuance. This is all about optimizing content. For the, the AI tools, right? You’re, you’re, is that, or this is a subset of optimization. It’s a way to make sure AI respects the source attributable to the content it has scraped and taken from the site. More and more we’re going to see increased traffic from these AI bots that are coming to, you know, answer somebody else’s question and maybe they show where that information came from and maybe they didn’t. And in doing that, these are the initial phase of tactics when you create your content and also, frankly, it’s putting The human back in the content, like the stuff I think we may look back at writing of uh the how to like tie your shoe content like because it gets traffic was not that relevant to our organization like the, you know, 15 cutest cats to promote our thing really was only getting people on a very high level to maybe browse through our site. So this is hopefully return to real content. OK, but that had value. Um, the, you know, the 15 cutest cats. All right, well, well, let me take a look at this shelter. Maybe this is a, if it’s, if it’s a local place, maybe, uh, maybe I’m looking to adopt, maybe I’m looking for a place to volunteer or obviously maybe give, so that, I mean, as, as a, as the beginning of a pipeline. That had value. It probably still does have, it does still have value, but what, you know, what you and I are focused on is the, the AI. Evaluation of your work. And from that perspective, the 15 cutest cats. Not valuable. It won’t drive attention. I mean, we are still humans here. This is where Tony voice is like on your shoulder. No joke, like cats. OK, OK. However, the fact, the nuance, the difference here is that you won’t get the attention you used to for that article. It used to drive attention. Without attention, you’re not gonna get the 1% of those people sticking around and giving you their email, of which 1 out of 10 makes a donation. That flow of traffic has been severed and is in the process of being severed, and you can see this action step. Look at your organic traffic year over year. It is at best flat if not going down. Even the smartest folks playing the game are publicly saying we’re in trouble when it comes to organic traffic. So that game is like, we’re on the decline. Uh, to come back, does that make sense? Yeah, yeah, it does. Um, that is, that’s what I’m talking about, it’s a source of organic traffic. But I’ll play the cute cat game if you want. So let’s say I’m sitting there and I’ve got like a a cat shelter, like, OK, uh, how could I turn this into something quotable, something from, uh, AI won’t steal or if they do, they’ll attribute it to me. Well, maybe back to QAs, I could actually do a study of over the past quarter actually. 70% of the cutest cats, based on my cuteness score versus ugly, cute or ugly, were adopted, yet 90% of ugly cats don’t get adopted. And suddenly I’ve got first party data that you collected, you surveyed, you did it is wildly interesting because I’m like, what’s an ugly cat, right? And now you’re playing the game because when AI comes it’s like according to the Long Island local cat shelter, ugly cats don’t get adopted at a rate of 90%. That’s right, that’s a very great. OK, OK, that’s an excellent example. of, of this, the tactic that we’re talking about. Please, if you go out there and do an ugly cat survey on a rate of adoption, please send it to me cause I think I’d be curious. What’s an ugly cat is the greatest question. You could you could use AI to make the determination so that way it would save you from the whole like internet fallout like according to this AI and how cute is this cat from 1 to 10? You can blame it on AI. OK, you, you mentioned some other things, uh, testimonials, quotes from the CEO, you know, all again, all it’s first party attributable to you, so, so then in and I’m gonna use Google too because that’s the, the primary search engine that people use. So in that Google summary. Powered by Gemini, right? This is all Gemini results that we get when we see the Google doesn’t say AI summary or something like AI overview AI mode. They’ll change it next week, so fine. It’s Gemini. It’s, it’s Gemini, correct. That’s the underlying model you’re correct. So we, we want to get attribution. I mean we wanted to say from whole whale from the Long Island cat shelter. Correct. You wanna be the authority that pops up on the side so you do in fact get a potential click and brand impression associated with the topic of someone saying like uh why do ugly cats not get adopted? Where do we put these testimonials in quotes? They’re basically embedded on your site, right, writing your content, right in line, um, you can use quote tags, you can use uh what’s called schema markup to make sure that when the AI is reading it. Uh, it is more quickly attributable, and in those testimonials, like you can just also just make up quotes like I now have to come up with some sort of clever thing to say, and I think for the AI study, I literally put in there I was like I have to come up with something. So I was like, alright, this season, AI will determine giving more than any other in history, like it’s a little. Over the top and like a pretty safe statement but I said it, I quoted it, I shoved it in with a quote tag inside of there and now that’s one of those attributing factors. There’s another big thing I wanna touch on, but I wanna make sure that makes sense. Uh, it does, but all right, so remember your big thing, but you are in jargon jail for scheme, you are in jargon jail for schema markup. I’m gonna bend up there somehow. Uh, the schema markup is, uh, little HTML tags in there, so behind the scenes, like if I want to make something bold, I put a little B tag on it. Every P tag has that little spacing. H one tags make it at the top header. This is just another one of those types of tags in the system that the machines read and understand more quickly what’s about to come and what lives between the open and close of that tag. OK, very good. All right, you’re out of jargon jail and listeners, the, the person who does your website will know exactly what we’re talking about. This is HTML. AI will know what you’re talking about if you literally, this is an amazing thing. Like your expert is the say, hey, teach me what I need to know about Schema. What, what would I need for schema markup of this page? You can ask AI and it is you’re ready to go developer on that front. OK. The This is the thing that these are the things that concern me. This is where we get into creativity conversations. What was your other big point? The other big point is something that has dawned on me and it kind of, uh, it sucks, technical term, because I’ve realized a lot of my writing is being. Commoditized and is not going to be found by people, it’s going to be summarized. And what I realized is where do we go to be uniquely human, where do we go for trusted information and more and more it’s going to be, I think in audio and video, video that is harder to fake, not impossible, but much more unique, a little bit more messy and so for your tentpole content, your main focused content. You have to have a video associated with it for a number of reasons, including the fact that YouTube is the #2 search engine, including the fact that people are now going to say like, I get that all this text I’m looking at is all like AI driven. I want to hear a human say it cause at least that human had to read it before saying it and letting those words out of their mouths. And third, because it is also showing up very, very clearly in those AI overviews right below, they’re giving video answers to the textual answer that AI gives and then they are having the good old fashioned links which are going to go the way of the yellow pages, I think. So we’re talking about side by side content. Part 111 part is for the humans and the other is for, is for the AI models on our site, you’re talking about videos, videos for the humans, mostly, although AI I understand the, the AI summaries are they’re, they’re finding the video and, and promoting those for you if they’re on point. But then there’s the also, but there’s the, there’s the AI. Needed content like side by side. We’re not talking about having to, I mean. Yeah, I, I, I, isn’t that what we’re talking about? 22 different levels of content, one for, one for the machine and one for the, for the humans who do come to your site. I’m beginning to believe that we’re probably headed toward the, the text of your site is more of like a database for AI to reference and so what is it that you’re going to create that is uniquely human communicating in that way is going to be video. For now, uh, because that, you know, is, it is also getting the lift on these other platforms. It is showing your brand, your people, and your message in your words as opposed to AI summarization of its spit back out in text, sometimes it’s attributed to you, sometimes it’s not. There’s no way to sort of like just sort of like stealing video the same way you do text. So I’m really Emphasizing and have been for a while the library of content that rides alongside my written content, which you know I try a little hard out on but I, you know, make no illusion that AI isn’t literally copying it and shoving it into its system and answering the questions that used to drive attention for us. So how do I play this game? I look at the data, I look at, right, it seems that it’s still on YouTube searching. I can embed it on my page so I increase engagement on my page when people do in fact go there, which is a positive signal in the land of the Google and old SEO and current. So it is a way to think about when we’re talking about your AI brand footprint, it’s a sort of adjacent because of the way Google is showing it and the way that everyone else also copies Google, so like in perplexity, which is a. Kind of Google competitor, I think for AI based um discovery of information and searching. Um, they are all sort of playing this game of amending and appending the relevant videos they find to that topic. OK, this is um. Uh, to me, this is as revolutionary as when back when we were saying you need to have a website. Yeah, this is a big phase shift. This is the yellow pages to website shift. I I don’t make that out of just sort of a throwaway statement. It is a big change and we’re in the middle of it. And so that is also one of those reasons I try to uh rant as much as possible, but wake people up to the type of content you’re about to create. You got a content calendar. Alright, we’re gonna map out 2026 and here are the like. 36 articles we’re gonna write, I’m like, take a beat because I think if you continue to write the way you used to and create content the way you used to, you are um. You’re creating another listing on the Yellow Pages. I don’t want folks wasting their time in that way. And also this should not be a surprise, but if you are using AI to write all of your content, why do you think AI is going to surface what it already wrote or what it can already answer? On your website, it is, you know, you don’t have to think that hard to be like, oh wait a minute, the hundreds of millions of people, like 10% of the adult human population on this current planet we are on uses AI to answer these questions like they’re, you know, you’re disintermediating yourself, you’re removing yourself when you simply press copy and paste from AI and by the way, people can tell when you’ve written it in a lazy way. Oh, that’s huge. That see you I wanted to, I, I. You were, you were given like a perfect summary. I thought, oh, this is a good place to end. I’ll just say that’s George Weiner, Keith Whaler and I’m out. Then you opened up, yeah, but then you opened up and, and by the way, um, you’re, you know, people can tell, yes, yes, uh, we’re still human here. And I don’t know. I, I’m not, I’m not saying that I’ve spotted every bit of artificially and artificially developed content that’s ever come across my screens. I’m not saying that, but. There is a feel of fakeness. To artificially generated. Paragraphs Yeah, it, you know what’s interesting? The human’s ability to do pattern recognition is tremendous. It’s, it’s unbelievable what we’re able to tune our attention to and what our brain simply learns behind the scenes, like we can now smell an AI generated image of it. We can smell the AI generated text, maybe it’s the vivacity, maybe it’s the uh unusually accurate cadence of number of words per sentence. There is something that you pick up and we are all collectively building. This ability to see it and maybe I’ll take a step back if you don’t believe me in our amazing pattern recognition. I want you to think in your mind of a stock photo of the following successful business person shaking hands and you’re like, it’s like you could smell it. It’s just the way it’s image focused cropped and it’s just generic humans doing business thing. The same way you can spot stock, we can now and are building that muscle, that pattern recognition for AI slop, work slop, whatever category of a bunch of AI generated texts, which is why it’s also fun talking about this new phase of like putting the human back in, how do you do that with video? How do you do that with testimonials, how do you do that with actual first party data? Like, I used to waste tons of time writing like very, very long articles and thinking and researching. But like Google searching and pulling back information, instead, I threw all that away. I spent all of my time doing this AI study to be like, I, I get to like geek out on this topic, go way deep. And then put that out instead, so like you’re just channeling your energy of creation in a different vector. I think your uh ability to spot stock photos analogy is is is spot on too. All right. That’s I think that’s very valuable, George. Well, we were here to make sure your audience didn’t walk away with panic. I feel like a lot of conversations go to this like hand wringing, what do we do next? So you have some takeaways. You got a lot of tactics. Uh, let’s let’s give one more shout out AI. That’s the research that’s the proprietary first person data that you’ll find at wholewhale.com. Go, just give him a break. Give him, give him some organic traffic. Just go to wholewhale.com. Don’t, don’t not click through from anywhere, just, just type in Wholewhale.com and go. Do it for George. Let’s see, let’s see if he gets a burst in, uh, in, in the month of October. And you’ll find George at uh on LinkedIn. He and I are very active there together, uh, just, you know, valuable content, George, then valuable, valuable ideas here. Thanks very much for sharing all this. Uh, thanks for giving it a larger audience. Next week, HR for non-HR professionals. OK, that, that, uh, that was supposed to have been this week, but George Weiner came in and he’s related to 4th quarter. So that’s why I squeezed him in because it’s the first show of the 4th quarter. So that’s the explanation, uh, ordinarily I would blame the associate producer, but this time, just this time, uh, it was, it was my own, my own. Uh, I would say rather savvy decision, but it was my decision, however you might characterize the decision, I made it. The savvy decision. If you missed any part of this week’s show, I beseech you. Find it at Tony Martignetti.com. Our creative producer is Claire Meyerhoff. I’m your associate producer Kate Martignetti. The show social media is by Susan Chavez. Mark Silverman is our web guy, and this music is by Scott Stein. Thank you for that affirmation, Scotty. Be with us next week for nonprofit Radio, big nonprofit ideas for the other 95%. Go out and be great.