AI Has Leveled the Marketing Playing Field  🔗 **Source:** [AI Has Leveled the Marketing Playing Field](https://share.snipd.com/episode/3e7616bc-71d2-45b5-9fcd-e5dd92e38c54) ## Highlights - **AI Democratized Neuromarketing Tools** - AI has democratized neuromarketing tools that used to require expensive labs and consultants. - Roger Dooley explains cheap AI simulations and LLMs let small businesses run eye-tracking, emotional analysis, and behavioral tests for the cost of a cup of coffee. Transcript: Roger Dooley Well, absolutely. In fact, you're correct, John. I have been writing about neuromarketing for 20 years, and I've actually come to the conclusion that now we have entered the era of neuromarketing 2.0. Right. And the big difference is that all of the tools have been democratized. When I started on the topic long ago, initially, I focused on the sort of hard science type things like measuring brain activity with EEG. But these tools were only available to the biggest brands. So I kind of pivoted to include a lot more behavioral science, psychology. I'm sure that many of your audience know about Robert Cialdini's six principles or seven principles now and try to apply them. But in any case, what's happened is not only have the tools for measuring customer reactions, things like eye tracking, for example, it used to be very expensive, had to be done in a lab, You had to recruit subjects. Now you can even use AI simulations that are very accurate that, you know, the cost as much as a cup of coffee. So you can use them not just on some really big campaign, but you want to see, gee, what, okay, I'm redesigning this webpage. What are people going to look at? You can get that answer instantly. So that's one area. But to me, the even more significant areas are, you know, I think that every smaller business has two big problems, you know, not enough time, not enough money. And they want to apply things like Cialdini's principles. They want to apply tools like cognitive biases, loss aversion, and all those tools. But just trying to keep all those in your head at one time is difficult. Just to read Cialdini's book, Influence, and Daniel Kahneman's book, Thinking Fast and Slow, is a thousand pages. So, you know, that's a lot of, that's a heavy lift for most small business owners. But what now you can do, you can use the very, these AI tools, large language models, Claude, ChatGPT, Gemini, that are available at such low cost to do this heavy lifting for you. ([Time 0:02:06](https://share.snipd.com/snip/b5cad319-0141-4d41-9bf3-565cf01b2042)) - **Amish Businesses Used AI To Repair Sensitive Customer Emails** - Dooley shares a story of Amish businesspeople using AI to write difficult customer emails and avoid damaging relationships. - The AI rewrote sensitive messages to communicate delays empathetically, helping preserve important customers. Transcript: Roger Dooley In fact, I just did a short video, an article about how even Amish business people who you do not really consider in the the vanguard of high tech used AI to write more difficult emails. Like you've got to communicate with a customer who's got a project behind schedule, but they're a big customer and he can't afford to irritate them and lose their business. ([Time 0:04:54](https://share.snipd.com/snip/9930e048-fa61-475f-b93c-240cd69bce11)) - **AI Often Scores Higher On Emotional Intelligence** - LLMs can outperform many humans on measured emotional intelligence because they've ingested vast literature and research on emotion. - Dooley cites a Swiss study where AI scored in the 80s while humans averaged around 60 on emotional tests. Transcript: Roger Dooley Exactly. And believe it or not, John, just last year, researchers in Switzerland ran an experiment. They gave emotional intelligence tests to measure people's emotional intelligence to both humans and to the major AI models of the time. And what they found was that the AI models scored far better on emotional intelligence than the average human. Humans were in like the 60% range. AI was into the 80s. ([Time 0:05:49](https://share.snipd.com/snip/4c7ebc2d-6ff8-4e66-b5c3-5aca87190f3d)) - **Create A Virtual Nudge Unit With LLMs** - Build a virtual expert team inside an LLM by naming authorities (Cialdini, B.J. Fogg, Dale Carnegie) and asking them to debate your marketing. - Use the model's multiple perspectives to get low-cost nudge-unit advice for your project. Transcript: Roger Dooley So you could put Robert Cialdini on there. You could put B.J. Fogg, who's one of the founders of the whole principles of behavioral design, behavior design. You could put Dale Carnegie on there. If you're in a direct sales situation, you really want to relate to people because he still has insights that are valuable. You could put a conversion expert like, oh, say a peplier or Brian Massey to name a few of my Austrian compatriots here who are conversion experts. And you can have them work together to advise you on your marketing, to look at your website, look at your landing page, look at the direct mail that you're going to send out. And you can get this advice. You can have them discuss it, debate it, and get into the details in the book. ([Time 0:07:49](https://share.snipd.com/snip/4fe5a3a6-ab9f-4b5f-ac15-5a30eaf6ab8c)) - **Run An AI Friction Audit On Existing Assets** - Critique existing assets with AI to find friction and suggest behavioral interventions like social proof or authority. - Do feed the webpage, email, or landing page to the model and ask for competitor analysis and specific fixes. Transcript: Roger Dooley Well, I've got a whole bunch of very specific prompts, but I think what I would start with depends on where you're at. If you've got something that you're trying to improve, you could show that to the AI and ask it to critique it from various perspectives. You could ask it to look for friction. Of course, you know, I'm the friction guy too, right? Talk about how making things easy is better. And we all put these barriers, you know, on websites and other processes. We don't really think of as being barriers or difficult, but to a new customer, they do look like a barrier of some kind. So you could have it do those kinds of analyses. Then you could have it suggest different behavioral principles for interventions that will work better. Do we need social proof or do we need authority? And you can also have it go out and analyze your competitors, see what your competitors are doing and use those ideas to improve yours and also see where they're falling flat. One of the stories I relate in the book is I did just an experiment for a local pool supply company. I created a hypothetical pool supply or pool service company and had to do an analysis for my little local market here on the west side of Austin. And it came out and identified all the competitors. It found their websites. It looked at their reviews on Google and Yelp and so on. And it analyzed their websites and found some really strange things. ([Time 0:09:02](https://share.snipd.com/snip/e705a156-6548-4e08-9cf3-ea8885376f6b)) - **Local Pool Company Analysis Revealed Odd Competitor Copy** - Dooley tested AI on a hypothetical local pool company, having it analyze competitors, reviews, and websites for positioning flaws. - The AI flagged a bizarre competitor comparison chart and recommended positioning changes tailored to the market. Transcript: Roger Dooley I created a hypothetical pool supply or pool service company and had to do an analysis for my little local market here on the west side of Austin. And it came out and identified all the competitors. It found their websites. It looked at their reviews on Google and Yelp and so on. And it analyzed their websites and found some really strange things. In one case, it found that this one competitor had this comparison chart. And this was like on a, not on their homepage, it was on a deeper page in the website that showed us and our competition. And it was really wild where, you know, we don't peek in your windows, we don't pee in your pool. And it's like all these crazy things that it said, okay, this is what this competitor is doing. So it pointed out that this is probably going to turn off more people than it's going to help, unless they've had those specific problems, but which aren't really typical of pool service Companies. But it can analyze what they're doing and then advise you on how to position your company to compete better. ([Time 0:09:58](https://share.snipd.com/snip/0c656417-d07c-4800-86be-39eb3c2cc2bf)) - **Give AI Context And Treat Outputs As Conversations** - Provide rich context to your AI: past communications, positioning, customer info, and goals so outputs match your brand voice. - Treat AI responses as a conversation; push back, ask why, and iterate until it aligns with your voice. Transcript: Roger Dooley Well, actually, what you're bringing up there, John, is a bigger picture, which falls into the context area, providing your AI model with context about your business. And so first of all, for your voice, you can give it a whole bunch of your past communications, writing and so on, and it can learn your voice. And there's techniques for doing that, but also you can talk about your positioning, your where you want to be if you're at the high end of your market and you're trying to create a specific Image then obviously the ai has to know that so it can give you better advice you can provide it information about customers and it can help you develop customer profiles too you know It could i think in general one piece of have, John, is people type a command into an AI, a prompt, and then say, okay, I got my results. It's far better to do that as a conversation where you probe and you push back on it. Because AI wants to please you. It'll give you answers you don't like, but push back on it and say, okay, well, why are you doing this? Would this other thing be better? Or what about this other concept I had? And when you push back and ask more questions, that's when you get your real insights. ([Time 0:12:00](https://share.snipd.com/snip/8c8ebd60-373a-432e-9f26-653d96e4952c)) - **Prevent Poor AI Results With Context And Multiple Models** - Avoid disappointing results by supplying context and probing deeply; ask the model to research competitors, market factors, and customer profiles first. - Try multiple models (Claude, Gemini, ChatGPT) because each can produce different useful ideas. Transcript: Roger Dooley For one about the company, the customers, the marketplace, not pushing the AI to develop part of that context. Because you could have it go out and do the surveys. What's the competition doing? What's my market like? What are the factors affecting my specific market at this time? And your AI can develop some of that knowledge for you. And you can create a whole document. Like one AI I like is perplexity for deep research. If I'm going to do a lot of research on a particular topic, maybe I'm going to introduce a new product. I really want to learn all I can about the product, the market and whatnot. You can use any of the models. Perplexity is one, but also the others do deep research. You can have to create a 20-page report, which you can then use for context in preparing your model. So I would say that not having enough context is one common mistake. And then the other is not really probing deeply enough, not pushing back, not asking for other answers. And there's something else that I highly recommend is always trying this with more than one model. You know, right now they're cheap enough that you don't have to just adopt one and stick with that one. If I'm doing something important, I'll ask Claude, which is my sort of number one go-to, but I'll ask Gemini and I'll ask ChatGPT as well. ([Time 0:13:40](https://share.snipd.com/snip/2faddcd8-72aa-4f43-977a-2d71cbbe6214)) - **AI Predicts Human Emotional Reaction To Messages** - AI predicts how humans will feel about a message by synthesizing novels, research, and common interpretations, despite not feeling itself. - Dooley ran a cruise line notice through Claude and got a far more empathetic rewrite than the original tone-deaf communication. Transcript: John Jantsch Talk a little bit about that from a practical standpoint, if I'm writing a sales page or a pricing page. Right. Roger Dooley Well, we already talked about or we mentioned the idea that AI scores better on emotional intelligence than humans because it has consumed so much knowledge about human emotions. It's read novels, it's read scientific papers. And so it understands emotions fairly well. And if you prompt it properly, you can get it to see those sort of emotional blind spots that you miss because we all know what we know and we think our customers know. We think we know how they're going to react. And sometimes we do. And certainly there are exceptionally skillful salespeople and leaders who have that intuition about how customers will react. Maybe Steve Jobs, you know, say, no, don't worry. You know, nobody's seen this thing before, but they're going to love it when it comes up with the iPhone. But, you know, most of us don't fall into that savant category. And we think we're sending some communication that's going to be easily understood and accepted the way we mean it. I mean, a couple of examples I've got are one when the Duolingo CEO sent out a letter, which is our, that was supposed to be very uplifting. It was that they were becoming an AI first company, you know, and it sounded really exciting and positive. Unfortunately, the way people interpret it was, okay, the boys are thinking, oh man, the robots are coming for our jobs now. And so it ended up being a demotivator and they had had to backtrack and say, no, that was what I meant. And these things happen all the time. And these aren't dumb people. It's just that sometimes the way we think our words are going to land doesn't work. I'm not going to go into details, but a great story about a communication to luxury cruise line passengers who got a letter in their cabin about a departure or an arrival change that totally Screwed up their flights for a corporate photo shoot. And so these luxury passengers were suddenly having to scramble for flights in the middle of Florida spring break when there's no flights, there's no hotel rooms. And the way that was communicated was in such a tone deaf way. There was not even an apology. There was a brief acknowledgement that if you incur extra costs, we'll reimburse them if they're reasonable. And, you know, it missed the whole point of empathy saying, okay, these are your people who are attached to your brand. They love your brand and you're doing this to them. You know, you've got to let them know that you care. And I ran it through Claude. It created a fire. First of all, it said, don't do this if it hasn't been done. This is a terrible idea. You're going to damage the brand. But it then created a much better letter that if you do have to create, do this thing, then here's how to communicate it. ([Time 0:16:56](https://share.snipd.com/snip/c84e2ae6-8f22-43a9-a2ed-ea26a03cef00)) - **Run A 30 Minute AI Behavioral Audit** - To run a quick behavioral audit, feed context and existing materials to the model, then ask it what it needs and for a prioritized plan. - Use Dooley's online prompt repository and ask the AI follow-up questions to refine results. Transcript: Roger Dooley A series of prompts to first, you know, add the context about the company, add the existing marketing materials that you're trying to audit. If you're trying to, you know, that's if that's your goal and then choose your experts or even have, you know, one thing that I always like to do with AI is if I'm not sure where to start, I will ask Claude or Gemini, okay, here's my objective. How should I start? What do you need for me? And the other thing, you know, in the book, I did not append this to every prompt, but in the, I did create a little mini website on my own website, rogerdooley.com. That is, there's a link in the back of each, at the end of each chapter where I've got copy and paste prompts so people don't have to type in, you know, lines and lines of text from the book. That'd be pretty brutal. And also some improved prompts that are go, sort of go beyond what I could really include in the book because nobody wants a book that's full of pages of prompts. But what I was going to, what I was making is that I do recommend adding, ask any questions that will help you respond every, anytime you prompt an AI, just add that little phrase. ([Time 0:20:05](https://share.snipd.com/snip/a1b4bceb-b5d5-437a-baf2-582750e04fad))